From ce89f5c9d12b2ba76189690a02249407686892ed Mon Sep 17 00:00:00 2001 From: gryf Date: Fri, 29 May 2020 17:04:53 +0200 Subject: [PATCH] Use the real constants module. This is progressing refactor of the calibre code to make it more readable, and transform it to something more coherent. In this patch, there are changes regarding imports for some modules, instead of polluting namespace of each module with some other modules symbols, which often were imported from other modules. Yuck. --- ebook_converter/ebooks/__init__.py | 65 +- .../ebooks/conversion/plugins/comic_input.py | 9 +- .../ebooks/conversion/plugins/docx_output.py | 128 ++-- .../ebooks/conversion/plugins/epub_input.py | 146 +++-- .../ebooks/conversion/plugins/epub_output.py | 108 ++-- .../ebooks/conversion/plugins/fb2_input.py | 90 +-- .../ebooks/conversion/plugins/lit_input.py | 39 +- .../ebooks/conversion/plugins/lrf_input.py | 56 +- .../ebooks/conversion/plugins/rtf_input.py | 154 ++--- .../ebooks/conversion/plugins/snb_input.py | 74 ++- ebook_converter/ebooks/docx/container.py | 95 +-- ebook_converter/ebooks/docx/to_html.py | 15 +- .../ebooks/docx/writer/from_html.py | 27 +- ebook_converter/ebooks/fb2/fb2ml.py | 176 ++++-- ebook_converter/ebooks/htmlz/oeb2html.py | 41 +- ebook_converter/ebooks/metadata/fb2.py | 119 ++-- ebook_converter/ebooks/metadata/opf2.py | 580 ++++++++++-------- ebook_converter/ebooks/metadata/opf3.py | 56 +- ebook_converter/ebooks/metadata/toc.py | 84 +-- ebook_converter/ebooks/metadata/utils.py | 23 +- ebook_converter/ebooks/metadata/xmp.py | 204 +++--- ebook_converter/ebooks/mobi/mobiml.py | 72 +-- ebook_converter/ebooks/mobi/reader/mobi8.py | 7 +- .../ebooks/mobi/writer2/serializer.py | 57 +- ebook_converter/ebooks/mobi/writer8/exth.py | 4 +- ebook_converter/ebooks/odt/input.py | 13 +- ebook_converter/ebooks/oeb/base.py | 550 +++++++++-------- ebook_converter/ebooks/oeb/parse_utils.py | 46 +- .../ebooks/oeb/polish/container.py | 23 +- ebook_converter/ebooks/oeb/polish/css.py | 7 +- ebook_converter/ebooks/oeb/polish/opf.py | 22 +- ebook_converter/ebooks/oeb/polish/parsing.py | 59 +- ebook_converter/ebooks/oeb/polish/pretty.py | 34 +- ebook_converter/ebooks/oeb/polish/split.py | 42 +- ebook_converter/ebooks/oeb/polish/toc.py | 321 +++++----- ebook_converter/ebooks/oeb/reader.py | 302 ++++----- ebook_converter/ebooks/oeb/stylizer.py | 28 +- .../ebooks/oeb/transforms/cover.py | 108 ++-- .../ebooks/oeb/transforms/flatcss.py | 86 +-- .../ebooks/oeb/transforms/htmltoc.py | 83 +-- .../ebooks/oeb/transforms/jacket.py | 11 +- .../ebooks/oeb/transforms/manglecase.py | 20 +- .../ebooks/oeb/transforms/metadata.py | 50 +- .../ebooks/oeb/transforms/page_margin.py | 5 +- .../ebooks/oeb/transforms/rasterize.py | 19 +- .../ebooks/oeb/transforms/split.py | 19 +- .../ebooks/oeb/transforms/structure.py | 29 +- ebook_converter/ebooks/pdf/pdftohtml.py | 63 +- ebook_converter/ebooks/txt/markdownml.py | 7 +- ebook_converter/ebooks/txt/textileml.py | 8 +- ebook_converter/ebooks/txt/txtml.py | 44 +- ebook_converter/startup.py | 20 +- ebook_converter/utils/xml_parse.py | 14 +- ebook_converter/utils/zipfile.py | 2 +- 54 files changed, 2383 insertions(+), 2081 deletions(-) diff --git a/ebook_converter/ebooks/__init__.py b/ebook_converter/ebooks/__init__.py index 66c1756..b4dfce0 100644 --- a/ebook_converter/ebooks/__init__.py +++ b/ebook_converter/ebooks/__init__.py @@ -1,12 +1,14 @@ -__license__ = 'GPL v3' -__copyright__ = '2008, Kovid Goyal ' - -''' +""" Code for the conversion of ebook formats and the reading of metadata from various formats. -''' +""" +import numbers +import os +import re +import sys + +from lxml import etree -import os, re, numbers, sys from ebook_converter import prints from ebook_converter.ebooks.chardet import xml_to_unicode @@ -30,12 +32,15 @@ class ParserError(ValueError): pass -BOOK_EXTENSIONS = ['lrf', 'rar', 'zip', 'rtf', 'lit', 'txt', 'txtz', 'text', 'htm', 'xhtm', - 'html', 'htmlz', 'xhtml', 'pdf', 'pdb', 'updb', 'pdr', 'prc', 'mobi', 'azw', 'doc', - 'epub', 'fb2', 'fbz', 'djv', 'djvu', 'lrx', 'cbr', 'cbz', 'cbc', 'oebzip', - 'rb', 'imp', 'odt', 'chm', 'tpz', 'azw1', 'pml', 'pmlz', 'mbp', 'tan', 'snb', - 'xps', 'oxps', 'azw4', 'book', 'zbf', 'pobi', 'docx', 'docm', 'md', - 'textile', 'markdown', 'ibook', 'ibooks', 'iba', 'azw3', 'ps', 'kepub', 'kfx', 'kpf'] +BOOK_EXTENSIONS = ['lrf', 'rar', 'zip', 'rtf', 'lit', 'txt', 'txtz', 'text', + 'htm', 'xhtm', 'html', 'htmlz', 'xhtml', 'pdf', 'pdb', + 'updb', 'pdr', 'prc', 'mobi', 'azw', 'doc', 'epub', 'fb2', + 'fbz', 'djv', 'djvu', 'lrx', 'cbr', 'cbz', 'cbc', 'oebzip', + 'rb', 'imp', 'odt', 'chm', 'tpz', 'azw1', 'pml', 'pmlz', + 'mbp', 'tan', 'snb', 'xps', 'oxps', 'azw4', 'book', 'zbf', + 'pobi', 'docx', 'docm', 'md', 'textile', 'markdown', + 'ibook', 'ibooks', 'iba', 'azw3', 'ps', 'kepub', 'kfx', + 'kpf'] def return_raster_image(path): @@ -49,8 +54,7 @@ def return_raster_image(path): def extract_cover_from_embedded_svg(html, base, log): from ebook_converter.ebooks.oeb.base import XPath, SVG, XLINK - from ebook_converter.utils.xml_parse import safe_xml_fromstring - root = safe_xml_fromstring(html) + root = etree.fromstring(html) svg = XPath('//svg:svg')(root) if len(svg) == 1 and len(svg[0]) == 1 and svg[0][0].tag == SVG('image'): @@ -65,10 +69,10 @@ def extract_calibre_cover(raw, base, log): from ebook_converter.ebooks.BeautifulSoup import BeautifulSoup soup = BeautifulSoup(raw) matches = soup.find(name=['h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'p', 'span', - 'font', 'br']) + 'font', 'br']) images = soup.findAll('img', src=True) - if matches is None and len(images) == 1 and \ - images[0].get('alt', '').lower()=='cover': + if (matches is None and len(images) == 1 and + images[0].get('alt', '').lower() == 'cover'): img = images[0] img = os.path.join(base, *img['src'].split('/')) q = return_raster_image(img) @@ -97,13 +101,14 @@ def render_html_svg_workaround(path_to_html, log, width=590, height=750): data = None if SVG_NS in raw: try: - data = extract_cover_from_embedded_svg(raw, - os.path.dirname(path_to_html), log) + data = extract_cover_from_embedded_svg( + raw, os.path.dirname(path_to_html), log) except Exception: pass if data is None: try: - data = extract_calibre_cover(raw, os.path.dirname(path_to_html), log) + data = extract_calibre_cover(raw, os.path.dirname(path_to_html), + log) except Exception: pass @@ -118,7 +123,8 @@ def render_html_data(path_to_html, width, height): result = {} def report_error(text=''): - prints('Failed to render', path_to_html, 'with errors:', file=sys.stderr) + prints('Failed to render', path_to_html, 'with errors:', + file=sys.stderr) if text: prints(text, file=sys.stderr) if result and result['stdout_stderr']: @@ -127,7 +133,8 @@ def render_html_data(path_to_html, width, height): with TemporaryDirectory('-render-html') as tdir: try: - result = fork_job('ebook_converter.ebooks.render_html', 'main', args=(path_to_html, tdir, 'jpeg')) + result = fork_job('ebook_converter.ebooks.render_html', 'main', + args=(path_to_html, tdir, 'jpeg')) except WorkerError as e: report_error(e.orig_tb) else: @@ -156,17 +163,20 @@ def normalize(x): def calibre_cover(title, author_string, series_string=None, - output_format='jpg', title_size=46, author_size=36, logo_path=None): + output_format='jpg', title_size=46, author_size=36, + logo_path=None): title = normalize(title) author_string = normalize(author_string) series_string = normalize(series_string) from ebook_converter.ebooks.covers import calibre_cover2 from ebook_converter.utils.img import image_to_data - ans = calibre_cover2(title, author_string or '', series_string or '', logo_path=logo_path, as_qimage=True) + ans = calibre_cover2(title, author_string or '', series_string or '', + logo_path=logo_path, as_qimage=True) return image_to_data(ans, fmt=output_format) -UNIT_RE = re.compile(r'^(-*[0-9]*[.]?[0-9]*)\s*(%|em|ex|en|px|mm|cm|in|pt|pc|rem|q)$') +UNIT_RE = re.compile(r'^(-*[0-9]*[.]?[0-9]*)\s*(%|em|ex|en|px|mm|cm|in|pt|pc' + r'|rem|q)$') def unit_convert(value, base, font, dpi, body_font_size=12): @@ -175,7 +185,7 @@ def unit_convert(value, base, font, dpi, body_font_size=12): return value try: return float(value) * 72.0 / dpi - except: + except Exception: pass result = value m = UNIT_RE.match(value) @@ -227,7 +237,8 @@ def generate_masthead(title, output_path=None, width=600, height=60): recs = load_defaults('mobi_output') masthead_font_family = recs.get('masthead_font', None) from ebook_converter.ebooks.covers import generate_masthead - return generate_masthead(title, output_path=output_path, width=width, height=height, font_family=masthead_font_family) + return generate_masthead(title, output_path=output_path, width=width, + height=height, font_family=masthead_font_family) def escape_xpath_attr(value): diff --git a/ebook_converter/ebooks/conversion/plugins/comic_input.py b/ebook_converter/ebooks/conversion/plugins/comic_input.py index c95e02d..5a81305 100644 --- a/ebook_converter/ebooks/conversion/plugins/comic_input.py +++ b/ebook_converter/ebooks/conversion/plugins/comic_input.py @@ -3,6 +3,7 @@ Based on ideas from comiclrf created by FangornUK. """ import shutil, textwrap, codecs, os +from ebook_converter import constants as const from ebook_converter.customize.conversion import InputFormatPlugin, OptionRecommendation from ebook_converter import CurrentDir from ebook_converter.ptempfile import PersistentTemporaryDirectory @@ -245,7 +246,6 @@ class ComicInput(InputFormatPlugin): return os.path.abspath('metadata.opf') def create_wrappers(self, pages): - from ebook_converter.ebooks.oeb.base import XHTML_NS wrappers = [] WRAPPER = textwrap.dedent('''\ @@ -267,7 +267,8 @@ class ComicInput(InputFormatPlugin): ''') dir = os.path.dirname(pages[0]) for i, page in enumerate(pages): - wrapper = WRAPPER%(XHTML_NS, i+1, os.path.basename(page), i+1) + wrapper = WRAPPER%(const.XHTML_NS, i+1, os.path.basename(page), + i+1) page = os.path.join(dir, 'page_%d.xhtml'%(i+1)) with open(page, 'wb') as f: f.write(wrapper.encode('utf-8')) @@ -275,8 +276,6 @@ class ComicInput(InputFormatPlugin): return wrappers def create_viewer_wrapper(self, pages): - from ebook_converter.ebooks.oeb.base import XHTML_NS - def page(src): return ''.format(os.path.basename(src)) @@ -303,7 +302,7 @@ class ComicInput(InputFormatPlugin): %s - ''' % (XHTML_NS, pages) + ''' % (const.XHTML_NS, pages) path = os.path.join(base, 'wrapper.xhtml') with open(path, 'wb') as f: f.write(wrapper.encode('utf-8')) diff --git a/ebook_converter/ebooks/conversion/plugins/docx_output.py b/ebook_converter/ebooks/conversion/plugins/docx_output.py index f1ea4eb..cf9e51f 100644 --- a/ebook_converter/ebooks/conversion/plugins/docx_output.py +++ b/ebook_converter/ebooks/conversion/plugins/docx_output.py @@ -1,14 +1,22 @@ -from ebook_converter.customize.conversion import OutputFormatPlugin, OptionRecommendation +import io +from lxml import etree + +from ebook_converter import constants as const +from ebook_converter.customize import conversion +from ebook_converter.ebooks.docx.dump import do_dump +from ebook_converter.ebooks.docx.writer.container import DOCX +from ebook_converter.ebooks.docx.writer.from_html import Convert +from ebook_converter.ebooks.metadata import opf2 as opf_meta +from ebook_converter.ebooks.oeb import base -__license__ = 'GPL v3' -__copyright__ = '2013, Kovid Goyal ' PAGE_SIZES = ['a0', 'a1', 'a2', 'a3', 'a4', 'a5', 'a6', 'b0', 'b1', 'b2', 'b3', 'b4', 'b5', 'b6', 'legal', 'letter'] +_OPT = conversion.OptionRecommendation -class DOCXOutput(OutputFormatPlugin): +class DOCXOutput(conversion.OutputFormatPlugin): name = 'DOCX Output' author = 'Kovid Goyal' @@ -16,75 +24,63 @@ class DOCXOutput(OutputFormatPlugin): commit_name = 'docx_output' ui_data = {'page_sizes': PAGE_SIZES} - options = { - OptionRecommendation(name='docx_page_size', recommended_value='letter', - level=OptionRecommendation.LOW, choices=PAGE_SIZES, - help='The size of the page. Default is letter. Choices ' - 'are %s' % PAGE_SIZES), - - OptionRecommendation(name='docx_custom_page_size', recommended_value=None, - help='Custom size of the document. Use the form widthxheight ' - 'EG. `123x321` to specify the width and height (in pts). ' - 'This overrides any specified page-size.'), - - OptionRecommendation(name='docx_no_cover', recommended_value=False, - help='Do not insert the book cover as an image at the start of the document.' - ' If you use this option, the book cover will be discarded.'), - - OptionRecommendation(name='preserve_cover_aspect_ratio', recommended_value=False, - help='Preserve the aspect ratio of the cover image instead of stretching' - ' it out to cover the entire page.'), - - OptionRecommendation(name='docx_no_toc', recommended_value=False, - help='Do not insert the table of contents as a page at the start of the document.'), - - OptionRecommendation(name='extract_to', - help='Extract the contents of the generated %s file to the ' - 'specified directory. The contents of the directory are first ' - 'deleted, so be careful.' % 'DOCX'), - - OptionRecommendation(name='docx_page_margin_left', recommended_value=72.0, - level=OptionRecommendation.LOW, - help='The size of the left page margin, in pts. Default is 72pt.' - ' Overrides the common left page margin setting.' - ), - - OptionRecommendation(name='docx_page_margin_top', recommended_value=72.0, - level=OptionRecommendation.LOW, - help='The size of the top page margin, in pts. Default is 72pt.' - ' Overrides the common top page margin setting, unless set to zero.' - ), - - OptionRecommendation(name='docx_page_margin_right', recommended_value=72.0, - level=OptionRecommendation.LOW, - help='The size of the right page margin, in pts. Default is 72pt.' - ' Overrides the common right page margin setting, unless set to zero.' - ), - - OptionRecommendation(name='docx_page_margin_bottom', recommended_value=72.0, - level=OptionRecommendation.LOW, - help='The size of the bottom page margin, in pts. Default is 72pt.' - ' Overrides the common bottom page margin setting, unless set to zero.' - ), - - } + options = {_OPT(name='docx_page_size', recommended_value='letter', + level=_OPT.LOW, choices=PAGE_SIZES, + help='The size of the page. Default is letter. Choices ' + 'are %s' % PAGE_SIZES), + _OPT(name='docx_custom_page_size', recommended_value=None, + help='Custom size of the document. Use the form ' + 'widthxheight EG. `123x321` to specify the width and ' + 'height (in pts). This overrides any specified ' + 'page-size.'), + _OPT(name='docx_no_cover', recommended_value=False, + help='Do not insert the book cover as an image at the ' + 'start of the document. If you use this option, the book ' + 'cover will be discarded.'), + _OPT(name='preserve_cover_aspect_ratio', + recommended_value=False, help='Preserve the aspect ratio ' + 'of the cover image instead of stretching it out to cover ' + 'the entire page.'), + _OPT(name='docx_no_toc', recommended_value=False, + help='Do not insert the table of contents as a page at ' + 'the start of the document.'), + _OPT(name='extract_to', help='Extract the contents of the ' + 'generated DOCX file to the specified directory. The ' + 'contents of the directory are first deleted, so be ' + 'careful.'), + _OPT(name='docx_page_margin_left', recommended_value=72.0, + level=_OPT.LOW, help='The size of the left page margin, ' + 'in pts. Default is 72pt. Overrides the common left page ' + 'margin setting.'), + _OPT(name='docx_page_margin_top', recommended_value=72.0, + level=_OPT.LOW, help='The size of the top page margin, ' + 'in pts. Default is 72pt. Overrides the common top page ' + 'margin setting, unless set to zero.'), + _OPT(name='docx_page_margin_right', recommended_value=72.0, + level=_OPT.LOW, help='The size of the right page margin, ' + 'in pts. Default is 72pt. Overrides the common right page ' + 'margin setting, unless set to zero.'), + _OPT(name='docx_page_margin_bottom', recommended_value=72.0, + level=_OPT.LOW, help='The size of the bottom page margin, ' + 'in pts. Default is 72pt. Overrides the common bottom ' + 'page margin setting, unless set to zero.')} def convert_metadata(self, oeb): - from lxml import etree - from ebook_converter.ebooks.oeb.base import OPF, OPF2_NS - from ebook_converter.ebooks.metadata.opf2 import OPF as ReadOPF - from io import BytesIO - package = etree.Element(OPF('package'), attrib={'version': '2.0'}, nsmap={None: OPF2_NS}) + + package = etree.Element(base.tag('opf', 'package'), + attrib={'version': '2.0'}, + nsmap={None: const.OPF2_NS}) oeb.metadata.to_opf2(package) - self.mi = ReadOPF(BytesIO(etree.tostring(package, encoding='utf-8')), populate_spine=False, try_to_guess_cover=False).to_book_metadata() + self.mi = opf_meta.OPF(io.BytesIO(etree.tostring(package, + encoding='utf-8')), + populate_spine=False, + try_to_guess_cover=False).to_book_metadata() def convert(self, oeb, output_path, input_plugin, opts, log): - from ebook_converter.ebooks.docx.writer.container import DOCX - from ebook_converter.ebooks.docx.writer.from_html import Convert docx = DOCX(opts, log) self.convert_metadata(oeb) - Convert(oeb, docx, self.mi, not opts.docx_no_cover, not opts.docx_no_toc)() + Convert(oeb, docx, self.mi, not opts.docx_no_cover, + not opts.docx_no_toc)() docx.write(output_path, self.mi) if opts.extract_to: - from ebook_converter.ebooks.docx.dump import do_dump do_dump(output_path, opts.extract_to) diff --git a/ebook_converter/ebooks/conversion/plugins/epub_input.py b/ebook_converter/ebooks/conversion/plugins/epub_input.py index 4316aa0..9056465 100644 --- a/ebook_converter/ebooks/conversion/plugins/epub_input.py +++ b/ebook_converter/ebooks/conversion/plugins/epub_input.py @@ -1,14 +1,19 @@ -import os, re, posixpath -from itertools import cycle +import hashlib +import itertools +import os +import re +import traceback +import uuid -from ebook_converter.customize.conversion import InputFormatPlugin, OptionRecommendation +from lxml import etree + +from ebook_converter.ebooks.metadata import opf2 as opf_meta +from ebook_converter.ebooks.oeb import base +from ebook_converter.customize.conversion import InputFormatPlugin +from ebook_converter.customize.conversion import OptionRecommendation -__license__ = 'GPL 3' -__copyright__ = '2009, Kovid Goyal ' -__docformat__ = 'restructuredtext en' - -ADOBE_OBFUSCATION = 'http://ns.adobe.com/pdf/enc#RC' +ADOBE_OBFUSCATION = 'http://ns.adobe.com/pdf/enc#RC' IDPF_OBFUSCATION = 'http://www.idpf.org/2008/embedding' @@ -16,8 +21,8 @@ def decrypt_font_data(key, data, algorithm): is_adobe = algorithm == ADOBE_OBFUSCATION crypt_len = 1024 if is_adobe else 1040 crypt = bytearray(data[:crypt_len]) - key = cycle(iter(bytearray(key))) - decrypt = bytes(bytearray(x^next(key) for x in crypt)) + key = itertools.cycle(iter(bytearray(key))) + decrypt = bytes(bytearray(x ^ next(key) for x in crypt)) return decrypt + data[crypt_len:] @@ -29,18 +34,16 @@ def decrypt_font(key, path, algorithm): class EPUBInput(InputFormatPlugin): - name = 'EPUB Input' - author = 'Kovid Goyal' + name = 'EPUB Input' + author = 'Kovid Goyal' description = 'Convert EPUB files (.epub) to HTML' - file_types = {'epub'} + file_types = {'epub'} output_encoding = None commit_name = 'epub_input' recommendations = {('page_breaks_before', '/', OptionRecommendation.MED)} def process_encryption(self, encfile, opf, log): - from lxml import etree - import uuid, hashlib idpf_key = opf.raw_unique_identifier if idpf_key: idpf_key = re.sub('[\u0020\u0009\u000d\u000a]', '', idpf_key) @@ -56,27 +59,28 @@ class EPUBInput(InputFormatPlugin): try: key = item.text.rpartition(':')[-1] key = uuid.UUID(key).bytes - except: - import traceback + except Exception: traceback.print_exc() key = None try: root = etree.parse(encfile) - for em in root.xpath('descendant::*[contains(name(), "EncryptionMethod")]'): + for em in root.xpath('descendant::*[contains(name(), ' + '"EncryptionMethod")]'): algorithm = em.get('Algorithm', '') if algorithm not in {ADOBE_OBFUSCATION, IDPF_OBFUSCATION}: return False - cr = em.getparent().xpath('descendant::*[contains(name(), "CipherReference")]')[0] + cr = em.getparent().xpath('descendant::*[contains(name(), ' + '"CipherReference")]')[0] uri = cr.get('URI') - path = os.path.abspath(os.path.join(os.path.dirname(encfile), '..', *uri.split('/'))) + path = os.path.abspath(os.path.join(os.path.dirname(encfile), + '..', *uri.split('/'))) tkey = (key if algorithm == ADOBE_OBFUSCATION else idpf_key) if (tkey and os.path.exists(path)): self._encrypted_font_uris.append(uri) decrypt_font(tkey, path, algorithm) return True - except: - import traceback + except Exception: traceback.print_exc() return False @@ -97,8 +101,11 @@ class EPUBInput(InputFormatPlugin): return t def rationalize_cover3(self, opf, log): - ''' If there is a reference to the cover/titlepage via manifest properties, convert to - entries in the so that the rest of the pipeline picks it up. ''' + """ + If there is a reference to the cover/titlepage via manifest + properties, convert to entries in the so that the rest of the + pipeline picks it up. + """ from ebook_converter.ebooks.metadata.opf3 import items_with_property removed = guide_titlepage_href = guide_titlepage_id = None @@ -128,7 +135,8 @@ class EPUBInput(InputFormatPlugin): titlepage_id, titlepage_href = tid, href.partition('#')[0] break if titlepage_href is None: - titlepage_href, titlepage_id = guide_titlepage_href, guide_titlepage_id + titlepage_href = guide_titlepage_href + titlepage_id = guide_titlepage_id if titlepage_href is not None: self.set_guide_type(opf, 'titlepage', titlepage_href, 'Title Page') spine = list(opf.iterspine()) @@ -148,7 +156,6 @@ class EPUBInput(InputFormatPlugin): means, at most one entry with type="cover" that points to a raster cover and at most one entry with type="titlepage" that points to an HTML titlepage. ''' - from ebook_converter.ebooks.oeb.base import OPF removed = None from lxml import etree guide_cover, guide_elem = None, None @@ -160,12 +167,14 @@ class EPUBInput(InputFormatPlugin): raster_cover = opf.raster_cover if raster_cover: if guide_elem is None: - g = opf.root.makeelement(OPF('guide')) + g = opf.root.makeelement(base.tag('opf', 'guide')) opf.root.append(g) else: g = guide_elem.getparent() guide_cover = raster_cover - guide_elem = g.makeelement(OPF('reference'), attrib={'href':raster_cover, 'type':'cover'}) + guide_elem = g.makeelement(base.tag('opf', 'reference'), + attrib={'href': raster_cover, + 'type': 'cover'}) g.append(guide_elem) return spine = list(opf.iterspine()) @@ -186,7 +195,8 @@ class EPUBInput(InputFormatPlugin): # specially if not self.for_viewer: if len(spine) == 1: - log.warn('There is only a single spine item and it is marked as the cover. Removing cover marking.') + log.warn('There is only a single spine item and it is marked ' + 'as the cover. Removing cover marking.') for guide_elem in tuple(opf.iterguide()): if guide_elem.get('type', '').lower() == 'cover': guide_elem.getparent().remove(guide_elem) @@ -215,8 +225,9 @@ class EPUBInput(InputFormatPlugin): # Render the titlepage to create a raster cover from ebook_converter.ebooks import render_html_svg_workaround guide_elem.set('href', 'calibre_raster_cover.jpg') - t = etree.SubElement( - elem[0].getparent(), OPF('item'), href=guide_elem.get('href'), id='calibre_raster_cover') + t = etree.SubElement(elem[0].getparent(), base.tag('opf', 'item'), + href=guide_elem.get('href'), + id='calibre_raster_cover') t.set('media-type', 'image/jpeg') if os.path.exists(guide_cover): renderer = render_html_svg_workaround(guide_cover, log) @@ -229,17 +240,16 @@ class EPUBInput(InputFormatPlugin): return removed def find_opf(self): - from ebook_converter.utils.xml_parse import safe_xml_fromstring - def attr(n, attr): for k, v in n.attrib.items(): if k.endswith(attr): return v try: with open('META-INF/container.xml', 'rb') as f: - root = safe_xml_fromstring(f.read()) + root = etree.fromstring(f.read()) for r in root.xpath('//*[local-name()="rootfile"]'): - if attr(r, 'media-type') != "application/oebps-package+xml": + if (attr(r, 'media-type') != + "application/oebps-package+xml"): continue path = attr(r, 'full-path') if not path: @@ -248,20 +258,18 @@ class EPUBInput(InputFormatPlugin): if os.path.exists(path): return path except Exception: - import traceback traceback.print_exc() def convert(self, stream, options, file_ext, log, accelerators): from ebook_converter.utils.zipfile import ZipFile from ebook_converter import walk from ebook_converter.ebooks import DRMError - from ebook_converter.ebooks.metadata.opf2 import OPF try: zf = ZipFile(stream) zf.extractall(os.getcwd()) - except: + except Exception: log.exception('EPUB appears to be invalid ZIP file, trying a' - ' more forgiving ZIP parser') + ' more forgiving ZIP parser') from ebook_converter.utils.localunzip import extractall stream.seek(0) extractall(stream) @@ -276,11 +284,12 @@ class EPUBInput(InputFormatPlugin): path = getattr(stream, 'name', 'stream') if opf is None: - raise ValueError('%s is not a valid EPUB file (could not find opf)'%path) + raise ValueError('%s is not a valid EPUB file (could not find ' + 'opf)' % path) opf = os.path.relpath(opf, os.getcwd()) - parts = os.path.split(opf) - opf = OPF(opf, os.path.dirname(os.path.abspath(opf))) + # parts = os.path.split(opf) + opf = opf_meta.OPF(opf, os.path.dirname(os.path.abspath(opf))) self._encrypted_font_uris = [] if os.path.exists(encfile): @@ -288,18 +297,23 @@ class EPUBInput(InputFormatPlugin): raise DRMError(os.path.basename(path)) self.encrypted_fonts = self._encrypted_font_uris - if len(parts) > 1 and parts[0]: - delta = '/'.join(parts[:-1])+'/' + # XXX(gryf): this code would fail pretty ugly, thus, this part was + # never used. + # if len(parts) > 1 and parts[0]: + # delta = '/'.join(parts[:-1])+'/' - def normpath(x): - return posixpath.normpath(delta + elem.get('href')) + # def normpath(x): + # return posixpath.normpath(delta + elem.get('href')) - for elem in opf.itermanifest(): - elem.set('href', normpath(elem.get('href'))) - for elem in opf.iterguide(): - elem.set('href', normpath(elem.get('href'))) + # for elem in opf.itermanifest(): + # elem.set('href', normpath(elem.get('href'))) + # for elem in opf.iterguide(): + # elem.set('href', normpath(elem.get('href'))) - f = self.rationalize_cover3 if opf.package_version >= 3.0 else self.rationalize_cover2 + if opf.package_version >= 3.0: + f = self.rationalize_cover3 + else: + f = self.rationalize_cover2 self.removed_cover = f(opf, log) if self.removed_cover: self.removed_items_to_ignore = (self.removed_cover,) @@ -352,15 +366,18 @@ class EPUBInput(InputFormatPlugin): from lxml import etree from ebook_converter.ebooks.chardet import xml_to_unicode from ebook_converter.ebooks.oeb.polish.parsing import parse - from ebook_converter.ebooks.oeb.base import EPUB_NS, XHTML, NCX_MIME, NCX, urlnormalize, urlunquote, serialize + from ebook_converter.ebooks.oeb.base import EPUB_NS, XHTML, NCX_MIME, \ + NCX, urlnormalize, urlunquote, serialize from ebook_converter.ebooks.oeb.polish.toc import first_child - from ebook_converter.utils.xml_parse import safe_xml_fromstring from tempfile import NamedTemporaryFile with open(nav_path, 'rb') as f: raw = f.read() - raw = xml_to_unicode(raw, strip_encoding_pats=True, assume_utf8=True)[0] + raw = xml_to_unicode(raw, strip_encoding_pats=True, + assume_utf8=True)[0] root = parse(raw, log=log) - ncx = safe_xml_fromstring('') + ncx = etree.fromstring('' + '') navmap = ncx[0] et = '{%s}type' % EPUB_NS bn = os.path.basename(nav_path) @@ -368,8 +385,8 @@ class EPUBInput(InputFormatPlugin): def add_from_li(li, parent): href = text = None for x in li.iterchildren(XHTML('a'), XHTML('span')): - text = etree.tostring( - x, method='text', encoding='unicode', with_tail=False).strip() or ' '.join( + text = etree.tostring(x, method='text', encoding='unicode', + with_tail=False).strip() or ' '.join( x.xpath('descendant-or-self::*/@title')).strip() href = x.get('href') if href: @@ -382,7 +399,7 @@ class EPUBInput(InputFormatPlugin): np[0].append(np.makeelement(NCX('text'))) np[0][0].text = text if href: - np.append(np.makeelement(NCX('content'), attrib={'src':href})) + np.append(np.makeelement(NCX('content'), attrib={'src': href})) return np def process_nav_node(node, toc_parent): @@ -401,20 +418,25 @@ class EPUBInput(InputFormatPlugin): else: return - with NamedTemporaryFile(suffix='.ncx', dir=os.path.dirname(nav_path), delete=False) as f: + with NamedTemporaryFile(suffix='.ncx', dir=os.path.dirname(nav_path), + delete=False) as f: f.write(etree.tostring(ncx, encoding='utf-8')) ncx_href = os.path.relpath(f.name, os.getcwd()).replace(os.sep, '/') - ncx_id = opf.create_manifest_item(ncx_href, NCX_MIME, append=True).get('id') + ncx_id = opf.create_manifest_item(ncx_href, NCX_MIME, + append=True).get('id') for spine in opf.root.xpath('//*[local-name()="spine"]'): spine.set('toc', ncx_id) - opts.epub3_nav_href = urlnormalize(os.path.relpath(nav_path).replace(os.sep, '/')) + url = os.path.relpath(nav_path).replace(os.sep, '/') + opts.epub3_nav_href = urlnormalize(url) opts.epub3_nav_parsed = root if getattr(self, 'removed_cover', None): changed = False base_path = os.path.dirname(nav_path) for elem in root.xpath('//*[@href]'): href, frag = elem.get('href').partition('#')[::2] - link_path = os.path.relpath(os.path.join(base_path, urlunquote(href)), base_path) + link_path = os.path.relpath(os.path.join(base_path, + urlunquote(href)), + base_path) abs_href = urlnormalize(link_path) if abs_href == self.removed_cover: changed = True diff --git a/ebook_converter/ebooks/conversion/plugins/epub_output.py b/ebook_converter/ebooks/conversion/plugins/epub_output.py index 9f522d7..df7c917 100644 --- a/ebook_converter/ebooks/conversion/plugins/epub_output.py +++ b/ebook_converter/ebooks/conversion/plugins/epub_output.py @@ -2,7 +2,11 @@ import os import re import shutil import urllib.parse +import uuid +from ebook_converter import constants as const +from ebook_converter.ebooks.oeb import base +from ebook_converter.ebooks.oeb import parse_utils from ebook_converter.customize.conversion import OutputFormatPlugin from ebook_converter.customize.conversion import OptionRecommendation @@ -132,39 +136,37 @@ class EPUBOutput(OutputFormatPlugin): recommendations = {('pretty_print', True, OptionRecommendation.HIGH)} def workaround_webkit_quirks(self): # {{{ - from ebook_converter.ebooks.oeb.base import XPath for x in self.oeb.spine: root = x.data - body = XPath('//h:body')(root) + body = base.XPath('//h:body')(root) if body: body = body[0] if not hasattr(body, 'xpath'): continue - for pre in XPath('//h:pre')(body): + for pre in base.XPath('//h:pre')(body): if not pre.text and len(pre) == 0: pre.tag = 'div' # }}} def upshift_markup(self): # {{{ 'Upgrade markup to comply with XHTML 1.1 where possible' - from ebook_converter.ebooks.oeb.base import XPath, XML for x in self.oeb.spine: root = x.data - if (not root.get(XML('lang'))) and (root.get('lang')): - root.set(XML('lang'), root.get('lang')) - body = XPath('//h:body')(root) + if (not root.get(base.tag('xml', 'lang'))) and (root.get('lang')): + root.set(base.tag('xml', 'lang'), root.get('lang')) + body = base.XPath('//h:body')(root) if body: body = body[0] if not hasattr(body, 'xpath'): continue - for u in XPath('//h:u')(root): + for u in base.XPath('//h:u')(root): u.tag = 'span' seen_ids, seen_names = set(), set() - for x in XPath('//*[@id or @name]')(root): + for x in base.XPath('//*[@id or @name]')(root): eid, name = x.get('id', None), x.get('name', None) if eid: if eid in seen_ids: @@ -223,28 +225,27 @@ class EPUBOutput(OutputFormatPlugin): first = next(iter(self.oeb.spine)) self.oeb.toc.add('Start', first.href) - from ebook_converter.ebooks.oeb.base import OPF identifiers = oeb.metadata['identifier'] - uuid = None + _uuid = None for x in identifiers: - if x.get(OPF('scheme'), None).lower() == 'uuid' or str(x).startswith('urn:uuid:'): - uuid = str(x).split(':')[-1] + if (x.get(base.tag('opf', 'scheme'), None).lower() == 'uuid' or + str(x).startswith('urn:uuid:')): + _uuid = str(x).split(':')[-1] break encrypted_fonts = getattr(input_plugin, 'encrypted_fonts', []) - if uuid is None: + if _uuid is None: self.log.warn('No UUID identifier found') - from uuid import uuid4 - uuid = str(uuid4()) - oeb.metadata.add('identifier', uuid, scheme='uuid', id=uuid) + _uuid = str(uuid.uuid4()) + oeb.metadata.add('identifier', _uuid, scheme='uuid', id=_uuid) - if encrypted_fonts and not uuid.startswith('urn:uuid:'): + if encrypted_fonts and not _uuid.startswith('urn:uuid:'): # Apparently ADE requires this value to start with urn:uuid: # for some absurd reason, or it will throw a hissy fit and refuse # to use the obfuscated fonts. for x in identifiers: - if str(x) == uuid: - x.content = 'urn:uuid:'+uuid + if str(x) == _uuid: + x.content = 'urn:uuid:' + _uuid with TemporaryDirectory('_epub_output') as tdir: from ebook_converter.customize.ui import plugin_for_output_format @@ -264,7 +265,7 @@ class EPUBOutput(OutputFormatPlugin): self.upgrade_to_epub3(tdir, opf) encryption = None if encrypted_fonts: - encryption = self.encrypt_fonts(encrypted_fonts, tdir, uuid) + encryption = self.encrypt_fonts(encrypted_fonts, tdir, _uuid) from ebook_converter.ebooks.epub import initialize_container with initialize_container(output_path, os.path.basename(opf), @@ -312,12 +313,12 @@ class EPUBOutput(OutputFormatPlugin): except EnvironmentError: pass - def encrypt_fonts(self, uris, tdir, uuid): # {{{ + def encrypt_fonts(self, uris, tdir, _uuid): # {{{ from ebook_converter.polyglot.binary import from_hex_bytes - key = re.sub(r'[^a-fA-F0-9]', '', uuid) + key = re.sub(r'[^a-fA-F0-9]', '', _uuid) if len(key) < 16: - raise ValueError('UUID identifier %r is invalid'%uuid) + raise ValueError('UUID identifier %r is invalid'% _uuid) key = bytearray(from_hex_bytes((key + key)[:32])) paths = [] with CurrentDir(tdir): @@ -335,7 +336,8 @@ class EPUBOutput(OutputFormatPlugin): if len(data) >= 1024: data = bytearray(data) f.seek(0) - f.write(bytes(bytearray(data[i] ^ key[i%16] for i in range(1024)))) + f.write(bytes(bytearray(data[i] ^ key[i%16] + for i in range(1024)))) else: self.log.warn('Font', path, 'is invalid, ignoring') if not isinstance(uri, str): @@ -374,11 +376,10 @@ class EPUBOutput(OutputFormatPlugin): # }}} def workaround_ade_quirks(self): # {{{ - ''' + """ Perform various markup transforms to get the output to render correctly in the quirky ADE. - ''' - from ebook_converter.ebooks.oeb.base import XPath, XHTML, barename, urlunquote + """ stylesheet = self.oeb.manifest.main_stylesheet @@ -388,23 +389,23 @@ class EPUBOutput(OutputFormatPlugin): for node in self.oeb.toc.iter(): href = getattr(node, 'href', None) if hasattr(href, 'partition'): - base, _, frag = href.partition('#') - frag = urlunquote(frag) + _base, _, frag = href.partition('#') + frag = base.urlunquote(frag) if frag and frag_pat.match(frag) is None: self.log.warn( 'Removing fragment identifier %r from TOC as Adobe Digital Editions cannot handle it'%frag) - node.href = base + node.href = _base for x in self.oeb.spine: root = x.data - body = XPath('//h:body')(root) + body = base.XPath('//h:body')(root) if body: body = body[0] if hasattr(body, 'xpath'): # remove tags with empty src elements bad = [] - for x in XPath('//h:img')(body): + for x in base.XPath('//h:img')(body): src = x.get('src', '').strip() if src in ('', '#') or src.startswith('http:'): bad.append(x) @@ -412,7 +413,7 @@ class EPUBOutput(OutputFormatPlugin): img.getparent().remove(img) # Add id attribute to tags that have name - for x in XPath('//h:a[@name]')(body): + for x in base.XPath('//h:a[@name]')(body): if not x.get('id', False): x.set('id', x.get('name')) # The delightful epubcheck has started complaining about tags that @@ -420,19 +421,19 @@ class EPUBOutput(OutputFormatPlugin): x.attrib.pop('name') # Replace
that are children of as ADE doesn't handle them - for br in XPath('./h:br')(body): + for br in base.XPath('./h:br')(body): if br.getparent() is None: continue try: prior = next(br.itersiblings(preceding=True)) - priortag = barename(prior.tag) + priortag = parse_utils.barename(prior.tag) priortext = prior.tail except: priortag = 'body' priortext = body.text if priortext: priortext = priortext.strip() - br.tag = XHTML('p') + br.tag = base.tag('xhtml', 'p') br.text = '\u00a0' style = br.get('style', '').split(';') style = list(filter(None, map(lambda x: x.strip(), style))) @@ -446,44 +447,44 @@ class EPUBOutput(OutputFormatPlugin): style.append('height:0pt') br.set('style', '; '.join(style)) - for tag in XPath('//h:embed')(root): + for tag in base.XPath('//h:embed')(root): tag.getparent().remove(tag) - for tag in XPath('//h:object')(root): + for tag in base.XPath('//h:object')(root): if tag.get('type', '').lower().strip() in {'image/svg+xml', 'application/svg+xml'}: continue tag.getparent().remove(tag) - for tag in XPath('//h:title|//h:style')(root): + for tag in base.XPath('//h:title|//h:style')(root): if not tag.text: tag.getparent().remove(tag) - for tag in XPath('//h:script')(root): + for tag in base.XPath('//h:script')(root): if (not tag.text and not tag.get('src', False) and tag.get('type', None) != 'text/x-mathjax-config'): tag.getparent().remove(tag) - for tag in XPath('//h:body/descendant::h:script')(root): + for tag in base.XPath('//h:body/descendant::h:script')(root): tag.getparent().remove(tag) - formchildren = XPath('./h:input|./h:button|./h:textarea|' + formchildren = base.XPath('./h:input|./h:button|./h:textarea|' './h:label|./h:fieldset|./h:legend') - for tag in XPath('//h:form')(root): + for tag in base.XPath('//h:form')(root): if formchildren(tag): tag.getparent().remove(tag) else: # Not a real form - tag.tag = XHTML('div') + tag.tag = base.tag('xhtml', 'div') - for tag in XPath('//h:center')(root): - tag.tag = XHTML('div') + for tag in base.XPath('//h:center')(root): + tag.tag = base.tag('xhtml', 'div') tag.set('style', 'text-align:center') # ADE can't handle & in an img url - for tag in XPath('//h:img[@src]')(root): + for tag in base.XPath('//h:img[@src]')(root): tag.set('src', tag.get('src', '').replace('&', '')) # ADE whimpers in fright when it encounters a outside a # - in_table = XPath('ancestor::h:table') - for tag in XPath('//h:td|//h:tr|//h:th')(root): + in_table = base.XPath('ancestor::h:table') + for tag in base.XPath('//h:td|//h:tr|//h:th')(root): if not in_table(tag): - tag.tag = XHTML('div') + tag.tag = base.tag('xhtml', 'div') # ADE fails to render non breaking hyphens/soft hyphens/zero width spaces special_chars = re.compile('[\u200b\u00ad]') @@ -498,7 +499,7 @@ class EPUBOutput(OutputFormatPlugin): if stylesheet is not None: # ADE doesn't render lists correctly if they have left margins from css_parser.css import CSSRule - for lb in XPath('//h:ul[@class]|//h:ol[@class]')(root): + for lb in base.XPath('//h:ul[@class]|//h:ol[@class]')(root): sel = '.'+lb.get('class') for rule in stylesheet.data.cssRules.rulesOfType(CSSRule.STYLE_RULE): if sel == rule.selectorList.selectorText: @@ -519,11 +520,10 @@ class EPUBOutput(OutputFormatPlugin): ''' Perform toc link transforms to alleviate slow loading. ''' - from ebook_converter.ebooks.oeb.base import XPath from ebook_converter.ebooks.oeb.polish.toc import item_at_top def frag_is_at_top(root, frag): - elem = XPath('//*[@id="%s" or @name="%s"]'%(frag, frag))(root) + elem = base.XPath('//*[@id="%s" or @name="%s"]'%(frag, frag))(root) if elem: elem = elem[0] else: diff --git a/ebook_converter/ebooks/conversion/plugins/fb2_input.py b/ebook_converter/ebooks/conversion/plugins/fb2_input.py index c84ea79..62f7b86 100644 --- a/ebook_converter/ebooks/conversion/plugins/fb2_input.py +++ b/ebook_converter/ebooks/conversion/plugins/fb2_input.py @@ -1,59 +1,57 @@ """ Convert .fb2 files to .lrf """ -import os, re +import os import pkg_resources +import re -from ebook_converter.customize.conversion import InputFormatPlugin, OptionRecommendation +from lxml import etree + +from ebook_converter import constants as const +from ebook_converter.customize.conversion import InputFormatPlugin +from ebook_converter.customize.conversion import OptionRecommendation from ebook_converter import guess_type -__license__ = 'GPL v3' -__copyright__ = '2008, Anatoly Shipitsin ' - -FB2NS = 'http://www.gribuser.ru/xml/fictionbook/2.0' +FB2NS = 'http://www.gribuser.ru/xml/fictionbook/2.0' FB21NS = 'http://www.gribuser.ru/xml/fictionbook/2.1' class FB2Input(InputFormatPlugin): - name = 'FB2 Input' - author = 'Anatoly Shipitsin' + name = 'FB2 Input' + author = 'Anatoly Shipitsin' description = 'Convert FB2 and FBZ files to HTML' - file_types = {'fb2', 'fbz'} + file_types = {'fb2', 'fbz'} commit_name = 'fb2_input' - recommendations = { - ('level1_toc', '//h:h1', OptionRecommendation.MED), - ('level2_toc', '//h:h2', OptionRecommendation.MED), - ('level3_toc', '//h:h3', OptionRecommendation.MED), - } + recommendations = {('level1_toc', '//h:h1', OptionRecommendation.MED), + ('level2_toc', '//h:h2', OptionRecommendation.MED), + ('level3_toc', '//h:h3', OptionRecommendation.MED)} - options = { - OptionRecommendation(name='no_inline_fb2_toc', - recommended_value=False, level=OptionRecommendation.LOW, - help='Do not insert a Table of Contents at the beginning of the book.' - )} + options = {OptionRecommendation(name='no_inline_fb2_toc', + recommended_value=False, + level=OptionRecommendation.LOW, + help='Do not insert a Table of Contents ' + 'at the beginning of the book.')} def convert(self, stream, options, file_ext, log, accelerators): - from lxml import etree - from ebook_converter.utils.xml_parse import safe_xml_fromstring - from ebook_converter.ebooks.metadata.fb2 import ensure_namespace, get_fb2_data + from ebook_converter.ebooks.metadata.fb2 import ensure_namespace + from ebook_converter.ebooks.metadata.fb2 import get_fb2_data from ebook_converter.ebooks.metadata.opf2 import OPFCreator from ebook_converter.ebooks.metadata.meta import get_metadata - from ebook_converter.ebooks.oeb.base import XLINK_NS, XHTML_NS from ebook_converter.ebooks.chardet import xml_to_unicode self.log = log log.debug('Parsing XML...') raw = get_fb2_data(stream)[0] raw = raw.replace(b'\0', b'') raw = xml_to_unicode(raw, strip_encoding_pats=True, - assume_utf8=True, resolve_entities=True)[0] + assume_utf8=True, resolve_entities=True)[0] try: - doc = safe_xml_fromstring(raw) + doc = etree.fromstring(raw) except etree.XMLSyntaxError: - doc = safe_xml_fromstring(raw.replace('& ', '&')) + doc = etree.fromstring(raw.replace('& ', '&')) if doc is None: raise ValueError('The FB2 file is not valid XML') doc = ensure_namespace(doc) @@ -62,22 +60,24 @@ class FB2Input(InputFormatPlugin): except Exception: fb_ns = FB2NS - NAMESPACES = {'f':fb_ns, 'l':XLINK_NS} - stylesheets = doc.xpath('//*[local-name() = "stylesheet" and @type="text/css"]') + NAMESPACES = {'f': fb_ns, 'l': const.XLINK_NS} + stylesheets = doc.xpath('//*[local-name() = "stylesheet" and ' + '@type="text/css"]') css = '' for s in stylesheets: css += etree.tostring(s, encoding='unicode', method='text', - with_tail=False) + '\n\n' + with_tail=False) + '\n\n' if css: - import css_parser, logging + import css_parser + import logging parser = css_parser.CSSParser(fetcher=None, - log=logging.getLogger('calibre.css')) + log=logging.getLogger('calibre.css')) - XHTML_CSS_NAMESPACE = '@namespace "%s";\n' % XHTML_NS + XHTML_CSS_NAMESPACE = '@namespace "%s";\n' % const.XHTML_NS text = XHTML_CSS_NAMESPACE + css log.debug('Parsing stylesheet...') stylesheet = parser.parseString(text) - stylesheet.namespaces['h'] = XHTML_NS + stylesheet.namespaces['h'] = const.XHTML_NS css = stylesheet.cssText if isinstance(css, bytes): css = css.decode('utf-8', 'replace') @@ -92,16 +92,20 @@ class FB2Input(InputFormatPlugin): if options.no_inline_fb2_toc: log('Disabling generation of inline FB2 TOC') ss = re.compile(r'.*', - re.DOTALL).sub('', ss) + re.DOTALL).sub('', ss) - styledoc = safe_xml_fromstring(ss) + styledoc = etree.fromstring(ss) transform = etree.XSLT(styledoc) result = transform(doc) # Handle links of type note and cite - notes = {a.get('href')[1:]: a for a in result.xpath('//a[@link_note and @href]') if a.get('href').startswith('#')} - cites = {a.get('link_cite'): a for a in result.xpath('//a[@link_cite]') if not a.get('href', '')} + notes = {a.get('href')[1:]: a + for a in result.xpath('//a[@link_note and @href]') + if a.get('href').startswith('#')} + cites = {a.get('link_cite'): a + for a in result.xpath('//a[@link_cite]') + if not a.get('href', '')} all_ids = {x for x in result.xpath('//*/@id')} for cite, a in cites.items(): note = notes.get(cite, None) @@ -137,8 +141,10 @@ class FB2Input(InputFormatPlugin): f.write(mi.cover_data[1]) cpath = os.path.abspath('fb2_cover_calibre_mi.jpg') else: - for img in doc.xpath('//f:coverpage/f:image', namespaces=NAMESPACES): - href = img.get('{%s}href'%XLINK_NS, img.get('href', None)) + for img in doc.xpath('//f:coverpage/f:image', + namespaces=NAMESPACES): + href = img.get('{%s}href' % const.XLINK_NS, + img.get('href', None)) if href is not None: if href.startswith('#'): href = href[1:] @@ -165,15 +171,15 @@ class FB2Input(InputFormatPlugin): ext = ct.rpartition('/')[-1].lower() if ext in ('png', 'jpeg', 'jpg'): if fname.lower().rpartition('.')[-1] not in {'jpg', 'jpeg', - 'png'}: + 'png'}: fname += '.' + ext self.binary_map[elem.get('id')] = fname raw = elem.text.strip() try: data = base64_decode(raw) except TypeError: - self.log.exception('Binary data with id=%s is corrupted, ignoring'%( - elem.get('id'))) + self.log.exception('Binary data with id=%s is corrupted, ' + 'ignoring' % elem.get('id')) else: with open(fname, 'wb') as f: f.write(data) diff --git a/ebook_converter/ebooks/conversion/plugins/lit_input.py b/ebook_converter/ebooks/conversion/plugins/lit_input.py index 8fbe8ce..446af67 100644 --- a/ebook_converter/ebooks/conversion/plugins/lit_input.py +++ b/ebook_converter/ebooks/conversion/plugins/lit_input.py @@ -1,17 +1,17 @@ +import copy + +from lxml import etree + +from ebook_converter import constants as const from ebook_converter.customize.conversion import InputFormatPlugin -__license__ = 'GPL v3' -__copyright__ = '2009, Kovid Goyal ' -__docformat__ = 'restructuredtext en' - - class LITInput(InputFormatPlugin): - name = 'LIT Input' - author = 'Marshall T. Vandegrift' + name = 'LIT Input' + author = 'Marshall T. Vandegrift' description = 'Convert LIT files to HTML' - file_types = {'lit'} + file_types = {'lit'} commit_name = 'lit_input' def convert(self, stream, options, file_ext, log, @@ -22,7 +22,7 @@ class LITInput(InputFormatPlugin): return create_oebbook(log, stream, options, reader=LitReader) def postprocess_book(self, oeb, opts, log): - from ebook_converter.ebooks.oeb.base import XHTML_NS, XPath, XHTML + from ebook_converter.ebooks.oeb.base import XPath, XHTML for item in oeb.spine: root = item.data if not hasattr(root, 'xpath'): @@ -37,22 +37,23 @@ class LITInput(InputFormatPlugin): body = body[0] if len(body) == 1 and body[0].tag == XHTML('pre'): pre = body[0] - from ebook_converter.ebooks.txt.processor import convert_basic, \ - separate_paragraphs_single_line + from ebook_converter.ebooks.txt.processor import \ + convert_basic, separate_paragraphs_single_line from ebook_converter.ebooks.chardet import xml_to_unicode - from ebook_converter.utils.xml_parse import safe_xml_fromstring - import copy - self.log('LIT file with all text in singe 
 tag detected')
+                    self.log('LIT file with all text in singe 
 tag '
+                             'detected')
                     html = separate_paragraphs_single_line(pre.text)
                     html = convert_basic(html).replace('',
-                            ''%XHTML_NS)
+                                                       '' %
+                                                       const.XHTML_NS)
                     html = xml_to_unicode(html, strip_encoding_pats=True,
-                            resolve_entities=True)[0]
+                                          resolve_entities=True)[0]
                     if opts.smarten_punctuation:
                         # SmartyPants skips text inside 
 tags
-                        from ebook_converter.ebooks.conversion.preprocess import smarten_punctuation
-                        html = smarten_punctuation(html, self.log)
-                    root = safe_xml_fromstring(html)
+                        from ebook_converter.ebooks.conversion import \
+                                preprocess
+                        html = preprocess.smarten_punctuation(html, self.log)
+                    root = etree.fromstring(html)
                     body = XPath('//h:body')(root)
                     pre.tag = XHTML('div')
                     pre.text = ''
diff --git a/ebook_converter/ebooks/conversion/plugins/lrf_input.py b/ebook_converter/ebooks/conversion/plugins/lrf_input.py
index 729c3b8..6f51f98 100644
--- a/ebook_converter/ebooks/conversion/plugins/lrf_input.py
+++ b/ebook_converter/ebooks/conversion/plugins/lrf_input.py
@@ -1,54 +1,52 @@
-import os, sys
+import os
+import sys
 import pkg_resources
 
+from lxml import etree
+
 from ebook_converter.customize.conversion import InputFormatPlugin
 
 
-__license__ = 'GPL v3'
-__copyright__ = '2009, Kovid Goyal '
-__docformat__ = 'restructuredtext en'
-
-
 class LRFInput(InputFormatPlugin):
 
-    name        = 'LRF Input'
-    author      = 'Kovid Goyal'
+    name = 'LRF Input'
+    author = 'Kovid Goyal'
     description = 'Convert LRF files to HTML'
-    file_types  = {'lrf'}
+    file_types = {'lrf'}
     commit_name = 'lrf_input'
 
     def convert(self, stream, options, file_ext, log,
                 accelerators):
-        from ebook_converter.ebooks.lrf.input import (MediaType, Styles, TextBlock,
-                Canvas, ImageBlock, RuledLine)
+        from ebook_converter.ebooks.lrf.input import MediaType, Styles, \
+                TextBlock, Canvas, ImageBlock, RuledLine
         self.log = log
         self.log('Generating XML')
         from ebook_converter.ebooks.lrf.lrfparser import LRFDocument
-        from ebook_converter.utils.xml_parse import safe_xml_fromstring
-        from lxml import etree
         d = LRFDocument(stream)
         d.parse()
         xml = d.to_xml(write_files=True)
         if options.verbose > 2:
             open(u'lrs.xml', 'wb').write(xml.encode('utf-8'))
-        doc = safe_xml_fromstring(xml)
+        doc = etree.fromstring(xml)
 
         char_button_map = {}
         for x in doc.xpath('//CharButton[@refobj]'):
             ro = x.get('refobj')
-            jump_button = doc.xpath('//*[@objid="%s"]'%ro)
+            jump_button = doc.xpath('//*[@objid="%s"]' % ro)
             if jump_button:
-                jump_to = jump_button[0].xpath('descendant::JumpTo[@refpage and @refobj]')
+                jump_to = jump_button[0].xpath('descendant::JumpTo[@refpage '
+                                               'and @refobj]')
                 if jump_to:
-                    char_button_map[ro] = '%s.xhtml#%s'%(jump_to[0].get('refpage'),
-                            jump_to[0].get('refobj'))
+                    char_button_map[ro] = ('%s.xhtml#%s' %
+                                           (jump_to[0].get('refpage'),
+                                            jump_to[0].get('refobj')))
         plot_map = {}
         for x in doc.xpath('//Plot[@refobj]'):
             ro = x.get('refobj')
-            image = doc.xpath('//Image[@objid="%s" and @refstream]'%ro)
+            image = doc.xpath('//Image[@objid="%s" and @refstream]' % ro)
             if image:
-                imgstr = doc.xpath('//ImageStream[@objid="%s" and @file]'%
-                    image[0].get('refstream'))
+                imgstr = doc.xpath('//ImageStream[@objid="%s" and @file]' %
+                                   image[0].get('refstream'))
                 if imgstr:
                     plot_map[ro] = imgstr[0].get('file')
 
@@ -58,21 +56,19 @@ class LRFInput(InputFormatPlugin):
                   resource_filename('ebook_converter',
                                     'data/lrf.xsl')) as fobj:
             # TODO(gryf): change this nonsense to etree.parse() instead.
-            styledoc = safe_xml_fromstring(fobj.read())
+            styledoc = etree.fromstring(fobj.read())
         media_type = MediaType()
         styles = Styles()
         text_block = TextBlock(styles, char_button_map, plot_map, log)
         canvas = Canvas(doc, styles, text_block, log)
         image_block = ImageBlock(canvas)
         ruled_line = RuledLine()
-        extensions = {
-                ('calibre', 'media-type') : media_type,
-                ('calibre', 'text-block') : text_block,
-                ('calibre', 'ruled-line') : ruled_line,
-                ('calibre', 'styles')     : styles,
-                ('calibre', 'canvas')     : canvas,
-                ('calibre', 'image-block'): image_block,
-                }
+        extensions = {('calibre', 'media-type'): media_type,
+                      ('calibre', 'text-block'): text_block,
+                      ('calibre', 'ruled-line'): ruled_line,
+                      ('calibre', 'styles'): styles,
+                      ('calibre', 'canvas'): canvas,
+                      ('calibre', 'image-block'): image_block}
         transform = etree.XSLT(styledoc, extensions=extensions)
         try:
             result = transform(doc)
diff --git a/ebook_converter/ebooks/conversion/plugins/rtf_input.py b/ebook_converter/ebooks/conversion/plugins/rtf_input.py
index 3c4192d..5d36fee 100644
--- a/ebook_converter/ebooks/conversion/plugins/rtf_input.py
+++ b/ebook_converter/ebooks/conversion/plugins/rtf_input.py
@@ -1,57 +1,58 @@
-import os, glob, re, textwrap
+import glob
+import os
 import pkg_resources
+import re
+import textwrap
 
-from ebook_converter.customize.conversion import InputFormatPlugin, OptionRecommendation
+from lxml import etree
+
+from ebook_converter.customize.conversion import InputFormatPlugin
+from ebook_converter.customize.conversion import OptionRecommendation
 from ebook_converter.polyglot.builtins import as_bytes
 
-__license__ = 'GPL v3'
-__copyright__ = '2008, Kovid Goyal '
 
-border_style_map = {
-        'single' : 'solid',
-        'double-thickness-border' : 'double',
-        'shadowed-border': 'outset',
-        'double-border': 'double',
-        'dotted-border': 'dotted',
-        'dashed': 'dashed',
-        'hairline': 'solid',
-        'inset': 'inset',
-        'dash-small': 'dashed',
-        'dot-dash': 'dotted',
-        'dot-dot-dash': 'dotted',
-        'outset': 'outset',
-        'tripple': 'double',
-        'triple': 'double',
-        'thick-thin-small': 'solid',
-        'thin-thick-small': 'solid',
-        'thin-thick-thin-small': 'solid',
-        'thick-thin-medium': 'solid',
-        'thin-thick-medium': 'solid',
-        'thin-thick-thin-medium': 'solid',
-        'thick-thin-large': 'solid',
-        'thin-thick-thin-large': 'solid',
-        'wavy': 'ridge',
-        'double-wavy': 'ridge',
-        'striped': 'ridge',
-        'emboss': 'inset',
-        'engrave': 'inset',
-        'frame': 'ridge',
-}
+border_style_map = {'single': 'solid',
+                    'double-thickness-border': 'double',
+                    'shadowed-border': 'outset',
+                    'double-border': 'double',
+                    'dotted-border': 'dotted',
+                    'dashed': 'dashed',
+                    'hairline': 'solid',
+                    'inset': 'inset',
+                    'dash-small': 'dashed',
+                    'dot-dash': 'dotted',
+                    'dot-dot-dash': 'dotted',
+                    'outset': 'outset',
+                    'tripple': 'double',
+                    'triple': 'double',
+                    'thick-thin-small': 'solid',
+                    'thin-thick-small': 'solid',
+                    'thin-thick-thin-small': 'solid',
+                    'thick-thin-medium': 'solid',
+                    'thin-thick-medium': 'solid',
+                    'thin-thick-thin-medium': 'solid',
+                    'thick-thin-large': 'solid',
+                    'thin-thick-thin-large': 'solid',
+                    'wavy': 'ridge',
+                    'double-wavy': 'ridge',
+                    'striped': 'ridge',
+                    'emboss': 'inset',
+                    'engrave': 'inset',
+                    'frame': 'ridge'}
 
 
 class RTFInput(InputFormatPlugin):
 
-    name        = 'RTF Input'
-    author      = 'Kovid Goyal'
+    name = 'RTF Input'
+    author = 'Kovid Goyal'
     description = 'Convert RTF files to HTML'
-    file_types  = {'rtf'}
+    file_types = {'rtf'}
     commit_name = 'rtf_input'
 
-    options = {
-        OptionRecommendation(name='ignore_wmf', recommended_value=False,
-            help='Ignore WMF images instead of replacing them with a '
-                 'placeholder image.'),
-    }
+    options = {OptionRecommendation(name='ignore_wmf', recommended_value=False,
+                                    help='Ignore WMF images instead of '
+                                    'replacing them with a placeholder '
+                                    'image.')}
 
     def generate_xml(self, stream):
         from ebook_converter.ebooks.rtf2xml.ParseRtf import ParseRtf
@@ -64,7 +65,7 @@ class RTFInput(InputFormatPlugin):
                 run_lev = 4
                 indent_out = 1
                 self.log('Running RTFParser in debug mode')
-            except:
+            except Exception:
                 self.log.warn('Impossible to run RTFParser in debug mode')
         parser = ParseRtf(
             in_file=stream,
@@ -108,7 +109,8 @@ class RTFInput(InputFormatPlugin):
             deb_dir=debug_dir,
 
             # Default encoding
-            default_encoding=getattr(self.opts, 'input_encoding', 'cp1252') or 'cp1252',
+            default_encoding=getattr(self.opts, 'input_encoding',
+                                     'cp1252') or 'cp1252',
 
             # Run level
             run_level=run_lev,
@@ -151,7 +153,7 @@ class RTFInput(InputFormatPlugin):
         for count, val in imap.items():
             try:
                 imap[count] = self.convert_image(val)
-            except:
+            except Exception:
                 self.log.exception('Failed to convert', val)
         return imap
 
@@ -161,7 +163,7 @@ class RTFInput(InputFormatPlugin):
         try:
             return self.rasterize_wmf(name)
         except Exception:
-            self.log.exception('Failed to convert WMF image %r'%name)
+            self.log.exception('Failed to convert WMF image %r' % name)
         return self.replace_wmf(name)
 
     def replace_wmf(self, name):
@@ -170,9 +172,11 @@ class RTFInput(InputFormatPlugin):
             return '__REMOVE_ME__'
         from ebook_converter.ebooks.covers import message_image
         if self.default_img is None:
-            self.default_img = message_image('Conversion of WMF images is not supported.'
-            ' Use Microsoft Word or OpenOffice to save this RTF file'
-            ' as HTML and convert that in calibre.')
+            self.default_img = message_image('Conversion of WMF images is not '
+                                             'supported. Use Microsoft Word '
+                                             'or OpenOffice to save this RTF '
+                                             'file as HTML and convert that '
+                                             'in calibre.')
         name = name.replace('.wmf', '.jpg')
         with open(name, 'wb') as f:
             f.write(self.default_img)
@@ -189,10 +193,10 @@ class RTFInput(InputFormatPlugin):
         return name
 
     def write_inline_css(self, ic, border_styles):
-        font_size_classes = ['span.fs%d { font-size: %spt }'%(i, x) for i, x in
-                enumerate(ic.font_sizes)]
-        color_classes = ['span.col%d { color: %s }'%(i, x) for i, x in
-                enumerate(ic.colors) if x != 'false']
+        font_size_classes = ['span.fs%d { font-size: %spt }' % (i, x)
+                             for i, x in enumerate(ic.font_sizes)]
+        color_classes = ['span.col%d { color: %s }' % (i, x)
+                         for i, x in enumerate(ic.colors) if x != 'false']
         css = textwrap.dedent('''
         span.none {
             text-decoration: none; font-weight: normal;
@@ -210,11 +214,11 @@ class RTFInput(InputFormatPlugin):
         span.strike-through { text-decoration: line-through }
 
         ''')
-        css += '\n'+'\n'.join(font_size_classes)
-        css += '\n' +'\n'.join(color_classes)
+        css += '\n' + '\n'.join(font_size_classes)
+        css += '\n' + '\n'.join(color_classes)
 
         for cls, val in border_styles.items():
-            css += '\n\n.%s {\n%s\n}'%(cls, val)
+            css += '\n\n.%s {\n%s\n}' % (cls, val)
 
         with open(u'styles.css', 'ab') as f:
             f.write(css.encode('utf-8'))
@@ -224,35 +228,34 @@ class RTFInput(InputFormatPlugin):
         style_map = {}
         for elem in doc.xpath(r'//*[local-name()="cell"]'):
             style = ['border-style: hidden', 'border-width: 1px',
-                    'border-color: black']
+                     'border-color: black']
             for x in ('bottom', 'top', 'left', 'right'):
-                bs = elem.get('border-cell-%s-style'%x, None)
+                bs = elem.get('border-cell-%s-style' % x, None)
                 if bs:
                     cbs = border_style_map.get(bs, 'solid')
-                    style.append('border-%s-style: %s'%(x, cbs))
-                bw = elem.get('border-cell-%s-line-width'%x, None)
+                    style.append('border-%s-style: %s' % (x, cbs))
+                bw = elem.get('border-cell-%s-line-width' % x, None)
                 if bw:
-                    style.append('border-%s-width: %spt'%(x, bw))
-                bc = elem.get('border-cell-%s-color'%x, None)
+                    style.append('border-%s-width: %spt' % (x, bw))
+                bc = elem.get('border-cell-%s-color' % x, None)
                 if bc:
-                    style.append('border-%s-color: %s'%(x, bc))
+                    style.append('border-%s-color: %s' % (x, bc))
             style = ';\n'.join(style)
             if style not in border_styles:
                 border_styles.append(style)
             idx = border_styles.index(style)
-            cls = 'border_style%d'%idx
+            cls = 'border_style%d' % idx
             style_map[cls] = style
             elem.set('class', cls)
         return style_map
 
     def convert(self, stream, options, file_ext, log,
                 accelerators):
-        from lxml import etree
         from ebook_converter.ebooks.metadata.meta import get_metadata
         from ebook_converter.ebooks.metadata.opf2 import OPFCreator
-        from ebook_converter.ebooks.rtf2xml.ParseRtf import RtfInvalidCodeException
+        from ebook_converter.ebooks.rtf2xml.ParseRtf import \
+            RtfInvalidCodeException
         from ebook_converter.ebooks.rtf.input import InlineClass
-        from ebook_converter.utils.xml_parse import safe_xml_fromstring
         self.opts = options
         self.log = log
         self.log('Converting RTF to XML...')
@@ -269,14 +272,15 @@ class RTFInput(InputFormatPlugin):
             imap = {}
             try:
                 imap = self.extract_images(d[0])
-            except:
+            except Exception:
                 self.log.exception('Failed to extract images...')
 
         self.log('Parsing XML...')
-        doc = safe_xml_fromstring(xml)
+        doc = etree.fromstring(xml)
         border_styles = self.convert_borders(doc)
         for pict in doc.xpath('//rtf:pict[@num]',
-                namespaces={'rtf':'http://rtf2xml.sourceforge.net/'}):
+                              namespaces={'rtf':
+                                          'http://rtf2xml.sourceforge.net/'}):
             num = int(pict.get('num'))
             name = imap.get(num, None)
             if name is not None:
@@ -286,8 +290,8 @@ class RTFInput(InputFormatPlugin):
         inline_class = InlineClass(self.log)
         with open(pkg_resources.resource_filename('ebook_converter',
                                                   'data/rtf.xsl')) as fobj:
-            styledoc = safe_xml_fromstring(fobj.read())
-        extensions = {('calibre', 'inline-class') : inline_class}
+            styledoc = etree.fromstring(fobj.read())
+        extensions = {('calibre', 'inline-class'): inline_class}
         transform = etree.XSLT(styledoc, extensions=extensions)
         result = transform(doc)
         html = u'index.xhtml'
@@ -296,7 +300,8 @@ class RTFInput(InputFormatPlugin):
             # res = res[:100].replace('xmlns:html', 'xmlns') + res[100:]
             # clean multiple \n
             res = re.sub(b'\n+', b'\n', res)
-            # Replace newlines inserted by the 'empty_paragraphs' option in rtf2xml with html blank lines
+            # Replace newlines inserted by the 'empty_paragraphs' option in
+            # rtf2xml with html blank lines
             # res = re.sub('\s*', '', res)
             # res = re.sub('(?<=\n)\n{2}',
             # u'
\u00a0
\n'.encode('utf-8'), res)
@@ -316,7 +321,8 @@ class RTFInput(InputFormatPlugin):
 
     def postprocess_book(self, oeb, opts, log):
         for item in oeb.spine:
-            for img in item.data.xpath('//*[local-name()="img" and @src="__REMOVE_ME__"]'):
+            for img in item.data.xpath('//*[local-name()="img" and '
+                                       '@src="__REMOVE_ME__"]'):
                 p = img.getparent()
                 idx = p.index(img)
                 p.remove(img)
diff --git a/ebook_converter/ebooks/conversion/plugins/snb_input.py b/ebook_converter/ebooks/conversion/plugins/snb_input.py
index 51c7d7e..fb16039 100644
--- a/ebook_converter/ebooks/conversion/plugins/snb_input.py
+++ b/ebook_converter/ebooks/conversion/plugins/snb_input.py
@@ -1,27 +1,33 @@
 import os
 
+from lxml import etree
+
 from ebook_converter.customize.conversion import InputFormatPlugin
 from ebook_converter.ptempfile import TemporaryDirectory
 from ebook_converter.utils.filenames import ascii_filename
 
 
-__license__ = 'GPL 3'
-__copyright__ = '2010, Li Fanxi '
-__docformat__ = 'restructuredtext en'
-
-HTML_TEMPLATE = '%s\n%s\n'
+HTML_TEMPLATE = ('%s'
+                 '\n%s\n')
 
 
 def html_encode(s):
-    return s.replace('&', '&').replace('<', '<').replace('>', '>').replace('"', '"').replace("'", ''').replace('\n', '
').replace(' ', ' ')  # noqa
+    return (s.replace('&', '&')
+            .replace('<', '<')
+            .replace('>', '>')
+            .replace('"', '"')
+            .replace("'", ''')
+            .replace('\n', '
')
+            .replace(' ', ' '))
 
 
 class SNBInput(InputFormatPlugin):
 
-    name        = 'SNB Input'
-    author      = 'Li Fanxi'
+    name = 'SNB Input'
+    author = 'Li Fanxi'
     description = 'Convert SNB files to OEB'
-    file_types  = {'snb'}
+    file_types = {'snb'}
     commit_name = 'snb_input'
 
     options = set()
@@ -32,13 +38,12 @@ class SNBInput(InputFormatPlugin):
 
         from ebook_converter.ebooks.oeb.base import DirContainer
         from ebook_converter.ebooks.snb.snbfile import SNBFile
-        from ebook_converter.utils.xml_parse import safe_xml_fromstring
 
         log.debug("Parsing SNB file...")
         snbFile = SNBFile()
         try:
             snbFile.Parse(stream)
-        except:
+        except Exception:
             raise ValueError("Invalid SNB file")
         if not snbFile.IsValid():
             log.debug("Invalid SNB file")
@@ -46,27 +51,28 @@ class SNBInput(InputFormatPlugin):
         log.debug("Handle meta data ...")
         from ebook_converter.ebooks.conversion.plumber import create_oebbook
         oeb = create_oebbook(log, None, options,
-                encoding=options.input_encoding, populate=False)
+                             encoding=options.input_encoding, populate=False)
         meta = snbFile.GetFileStream('snbf/book.snbf')
         if meta is not None:
-            meta = safe_xml_fromstring(meta)
-            l = {'title'    : './/head/name',
-                  'creator'  : './/head/author',
-                  'language' : './/head/language',
-                  'generator': './/head/generator',
-                  'publisher': './/head/publisher',
-                  'cover'    : './/head/cover', }
+            meta = etree.fromstring(meta)
+            item_map = {'title': './/head/name',
+                        'creator': './/head/author',
+                        'language': './/head/language',
+                        'generator': './/head/generator',
+                        'publisher': './/head/publisher',
+                        'cover': './/head/cover'}
             d = {}
-            for item in l:
-                node = meta.find(l[item])
+            for key, item in item_map.items():
+                node = meta.find(item)
                 if node is not None:
-                    d[item] = node.text if node.text is not None else ''
+                    d[key] = node.text if node.text is not None else ''
                 else:
-                    d[item] = ''
+                    d[key] = ''
 
             oeb.metadata.add('title', d['title'])
-            oeb.metadata.add('creator', d['creator'], attrib={'role':'aut'})
-            oeb.metadata.add('language', d['language'].lower().replace('_', '-'))
+            oeb.metadata.add('creator', d['creator'], attrib={'role': 'aut'})
+            oeb.metadata.add('language',
+                             d['language'].lower().replace('_', '-'))
             oeb.metadata.add('generator', d['generator'])
             oeb.metadata.add('publisher', d['publisher'])
             if d['cover'] != '':
@@ -84,7 +90,7 @@ class SNBInput(InputFormatPlugin):
             toc = snbFile.GetFileStream('snbf/toc.snbf')
             oeb.container = DirContainer(tdir, log)
             if toc is not None:
-                toc = safe_xml_fromstring(toc)
+                toc = etree.fromstring(toc)
                 i = 1
                 for ch in toc.find('.//body'):
                     chapterName = ch.text
@@ -93,18 +99,22 @@ class SNBInput(InputFormatPlugin):
                     data = snbFile.GetFileStream('snbc/' + chapterSrc)
                     if data is None:
                         continue
-                    snbc = safe_xml_fromstring(data)
+                    snbc = etree.fromstring(data)
                     lines = []
                     for line in snbc.find('.//body'):
                         if line.tag == 'text':
                             lines.append('
%s
' % html_encode(line.text))
                         elif line.tag == 'img':
-                            lines.append('
' % html_encode(line.text))
+                            lines.append('
' %
+                                         html_encode(line.text))
                     with open(os.path.join(tdir, fname), 'wb') as f:
-                        f.write((HTML_TEMPLATE % (chapterName, '\n'.join(lines))).encode('utf-8', 'replace'))
+                        f.write((HTML_TEMPLATE %
+                                 (chapterName,
+                                  '\n'.join(lines))).encode('utf-8',
+                                                            'replace'))
                     oeb.toc.add(ch.text, fname)
-                    id, href = oeb.manifest.generate(id='html',
-                        href=ascii_filename(fname))
+                    id, href = oeb.manifest.generate(
+                        id='html', href=ascii_filename(fname))
                     item = oeb.manifest.add(id, href, 'text/html')
                     item.html_input_href = fname
                     oeb.spine.add(item, True)
@@ -112,7 +122,7 @@ class SNBInput(InputFormatPlugin):
                 imageFiles = snbFile.OutputImageFiles(tdir)
                 for f, m in imageFiles:
                     id, href = oeb.manifest.generate(id='image',
-                        href=ascii_filename(f))
+                                                     href=ascii_filename(f))
                     item = oeb.manifest.add(id, href, m)
                     item.html_input_href = f
 
diff --git a/ebook_converter/ebooks/docx/container.py b/ebook_converter/ebooks/docx/container.py
index e9768f4..db11f3d 100644
--- a/ebook_converter/ebooks/docx/container.py
+++ b/ebook_converter/ebooks/docx/container.py
@@ -1,9 +1,12 @@
-import os, sys, shutil
+import os
+import shutil
+import sys
 
 from lxml import etree
 
 from ebook_converter import walk, guess_type
-from ebook_converter.ebooks.metadata import string_to_authors, authors_to_sort_string
+from ebook_converter.ebooks.metadata import authors_to_sort_string
+from ebook_converter.ebooks.metadata import string_to_authors
 from ebook_converter.ebooks.metadata.book.base import Metadata
 from ebook_converter.ebooks.docx import InvalidDOCX
 from ebook_converter.ebooks.docx.names import DOCXNamespace
@@ -11,21 +14,11 @@ from ebook_converter.ptempfile import PersistentTemporaryDirectory
 from ebook_converter.utils.localization import canonicalize_lang
 from ebook_converter.utils.logging import default_log
 from ebook_converter.utils.zipfile import ZipFile
-from ebook_converter.utils.xml_parse import safe_xml_fromstring
 
 
-__license__ = 'GPL v3'
-__copyright__ = '2013, Kovid Goyal '
-
-
-def fromstring(raw, parser=None):
-    return safe_xml_fromstring(raw)
-
 # Read metadata {{{
-
-
 def read_doc_props(raw, mi, XPath):
-    root = fromstring(raw)
+    root = etree.fromstring(raw)
     titles = XPath('//dc:title')(root)
     if titles:
         title = titles[0].text
@@ -53,29 +46,31 @@ def read_doc_props(raw, mi, XPath):
     desc = XPath('//dc:description')(root)
     if desc:
         raw = etree.tostring(desc[0], method='text', encoding='unicode')
-        raw = raw.replace('_x000d_', '')  # Word 2007 mangles newlines in the summary
+        # Word 2007 mangles newlines in the summary
+        raw = raw.replace('_x000d_', '')
         mi.comments = raw.strip()
 
     langs = []
     for lang in XPath('//dc:language')(root):
         if lang.text and lang.text.strip():
-            l = canonicalize_lang(lang.text)
-            if l:
-                langs.append(l)
+            canonic_lang = canonicalize_lang(lang.text)
+            if canonic_lang:
+                langs.append(canonic_lang)
     if langs:
         mi.languages = langs
 
 
 def read_app_props(raw, mi):
-    root = fromstring(raw)
+    root = etree.fromstring(raw)
     company = root.xpath('//*[local-name()="Company"]')
     if company and company[0].text and company[0].text.strip():
         mi.publisher = company[0].text.strip()
 
 
 def read_default_style_language(raw, mi, XPath):
-    root = fromstring(raw)
-    for lang in XPath('/w:styles/w:docDefaults/w:rPrDefault/w:rPr/w:lang/@w:val')(root):
+    root = etree.fromstring(raw)
+    for lang in XPath('/w:styles/w:docDefaults/w:rPrDefault/w:rPr/w:lang/'
+                      '@w:val')(root):
         lang = canonicalize_lang(lang)
         if lang:
             mi.languages = [lang]
@@ -87,7 +82,9 @@ class DOCX(object):
 
     def __init__(self, path_or_stream, log=None, extract=True):
         self.docx_is_transitional = True
-        stream = path_or_stream if hasattr(path_or_stream, 'read') else open(path_or_stream, 'rb')
+        stream = path_or_stream
+        if not hasattr(path_or_stream, 'read'):
+            stream = open(path_or_stream, 'rb')
         self.name = getattr(stream, 'name', None) or ''
         self.log = log or default_log
         if extract:
@@ -107,9 +104,9 @@ class DOCX(object):
         try:
             zf = ZipFile(stream)
             zf.extractall(self.tdir)
-        except:
+        except Exception:
             self.log.exception('DOCX appears to be invalid ZIP file, trying a'
-                    ' more forgiving ZIP parser')
+                               ' more forgiving ZIP parser')
             from ebook_converter.utils.localunzip import extractall
             stream.seek(0)
             extractall(stream, self.tdir)
@@ -133,13 +130,17 @@ class DOCX(object):
         try:
             raw = self.read('[Content_Types].xml')
         except KeyError:
-            raise InvalidDOCX('The file %s docx file has no [Content_Types].xml' % self.name)
-        root = fromstring(raw)
+            raise InvalidDOCX('The file %s docx file has no '
+                              '[Content_Types].xml' % self.name)
+        root = etree.fromstring(raw)
         self.content_types = {}
         self.default_content_types = {}
-        for item in root.xpath('//*[local-name()="Types"]/*[local-name()="Default" and @Extension and @ContentType]'):
-            self.default_content_types[item.get('Extension').lower()] = item.get('ContentType')
-        for item in root.xpath('//*[local-name()="Types"]/*[local-name()="Override" and @PartName and @ContentType]'):
+        for item in root.xpath('//*[local-name()="Types"]/*[local-name()='
+                               '"Default" and @Extension and @ContentType]'):
+            self.default_content_types[item.get('Extension').lower()] = \
+                    item.get('ContentType')
+        for item in root.xpath('//*[local-name()="Types"]/*[local-name()='
+                               '"Override" and @PartName and @ContentType]'):
             name = item.get('PartName').lstrip('/')
             self.content_types[name] = item.get('ContentType')
 
@@ -155,15 +156,19 @@ class DOCX(object):
         try:
             raw = self.read('_rels/.rels')
         except KeyError:
-            raise InvalidDOCX('The file %s docx file has no _rels/.rels' % self.name)
-        root = fromstring(raw)
+            raise InvalidDOCX('The file %s docx file has no _rels/.rels' %
+                              self.name)
+        root = etree.fromstring(raw)
         self.relationships = {}
         self.relationships_rmap = {}
-        for item in root.xpath('//*[local-name()="Relationships"]/*[local-name()="Relationship" and @Type and @Target]'):
+        for item in root.xpath('//*[local-name()="Relationships"]/*[local-name'
+                               '()="Relationship" and @Type and @Target]'):
             target = item.get('Target').lstrip('/')
             typ = item.get('Type')
             if target == 'word/document.xml':
-                self.docx_is_transitional = typ != 'http://purl.oclc.org/ooxml/officeDocument/relationships/officeDocument'
+                self.docx_is_transitional = (typ != 'http://purl.oclc.org/'
+                                             'ooxml/officeDocument/'
+                                             'relationships/officeDocument')
             self.relationships[typ] = target
             self.relationships_rmap[target] = typ
 
@@ -171,15 +176,17 @@ class DOCX(object):
     def document_name(self):
         name = self.relationships.get(self.namespace.names['DOCUMENT'], None)
         if name is None:
-            names = tuple(n for n in self.names if n == 'document.xml' or n.endswith('/document.xml'))
+            names = tuple(n for n in self.names if n == 'document.xml' or
+                          n.endswith('/document.xml'))
             if not names:
-                raise InvalidDOCX('The file %s docx file has no main document' % self.name)
+                raise InvalidDOCX('The file %s docx file has no main '
+                                  'document' % self.name)
             name = names[0]
         return name
 
     @property
     def document(self):
-        return fromstring(self.read(self.document_name))
+        return etree.fromstring(self.read(self.document_name))
 
     @property
     def document_relationships(self):
@@ -195,10 +202,13 @@ class DOCX(object):
         except KeyError:
             pass
         else:
-            root = fromstring(raw)
-            for item in root.xpath('//*[local-name()="Relationships"]/*[local-name()="Relationship" and @Type and @Target]'):
+            root = etree.fromstring(raw)
+            for item in root.xpath('//*[local-name()="Relationships"]/*'
+                                   '[local-name()="Relationship" and @Type '
+                                   'and @Target]'):
                 target = item.get('Target')
-                if item.get('TargetMode', None) != 'External' and not target.startswith('#'):
+                if (item.get('TargetMode', None) != 'External' and not
+                        target.startswith('#')):
                     target = '/'.join((base, target.lstrip('/')))
                 typ = item.get('Type')
                 Id = item.get('Id')
@@ -209,13 +219,15 @@ class DOCX(object):
     def get_document_properties_names(self):
         name = self.relationships.get(self.namespace.names['DOCPROPS'], None)
         if name is None:
-            names = tuple(n for n in self.names if n.lower() == 'docprops/core.xml')
+            names = tuple(n for n in self.names
+                          if n.lower() == 'docprops/core.xml')
             if names:
                 name = names[0]
         yield name
         name = self.relationships.get(self.namespace.names['APPPROPS'], None)
         if name is None:
-            names = tuple(n for n in self.names if n.lower() == 'docprops/app.xml')
+            names = tuple(n for n in self.names
+                          if n.lower() == 'docprops/app.xml')
             if names:
                 name = names[0]
         yield name
@@ -239,7 +251,8 @@ class DOCX(object):
             else:
                 read_default_style_language(raw, mi, self.namespace.XPath)
 
-        ap_name = self.relationships.get(self.namespace.names['APPPROPS'], None)
+        ap_name = self.relationships.get(self.namespace.names['APPPROPS'],
+                                         None)
         if ap_name:
             try:
                 raw = self.read(ap_name)
diff --git a/ebook_converter/ebooks/docx/to_html.py b/ebook_converter/ebooks/docx/to_html.py
index 11ce9e2..30c2a64 100644
--- a/ebook_converter/ebooks/docx/to_html.py
+++ b/ebook_converter/ebooks/docx/to_html.py
@@ -1,12 +1,13 @@
 import sys, os, re, math, errno, uuid, numbers
 from collections import OrderedDict, defaultdict
 
+from lxml import etree
 from lxml import html
 from lxml.html.builder import (
     HTML, HEAD, TITLE, BODY, LINK, META, P, SPAN, BR, DIV, A, DT, DL, DD, H1)
 
 from ebook_converter import guess_type
-from ebook_converter.ebooks.docx.container import DOCX, fromstring
+from ebook_converter.ebooks.docx.container import DOCX
 from ebook_converter.ebooks.docx.names import XML, generate_anchor
 from ebook_converter.ebooks.docx.styles import Styles, inherit, PageProperties
 from ebook_converter.ebooks.docx.numbering import Numbering
@@ -311,7 +312,7 @@ class Convert(object):
                     raise
                 self.log.warn('Settings %s file missing' % sename)
             else:
-                self.settings(fromstring(seraw))
+                self.settings(etree.fromstring(seraw))
 
         if foname is not None:
             try:
@@ -327,7 +328,7 @@ class Convert(object):
                 self.log.warn('Endnotes %s do not exist' % enname)
             else:
                 enrel = self.docx.get_relationships(enname)
-        footnotes(fromstring(foraw) if foraw else None, forel, fromstring(enraw) if enraw else None, enrel)
+        footnotes(etree.fromstring(foraw) if foraw else None, forel, etree.fromstring(enraw) if enraw else None, enrel)
 
         if fname is not None:
             embed_relationships = self.docx.get_relationships(fname)[0]
@@ -336,7 +337,7 @@ class Convert(object):
             except KeyError:
                 self.log.warn('Fonts table %s does not exist' % fname)
             else:
-                fonts(fromstring(raw), embed_relationships, self.docx, self.dest_dir)
+                fonts(etree.fromstring(raw), embed_relationships, self.docx, self.dest_dir)
 
         if tname is not None:
             try:
@@ -344,7 +345,7 @@ class Convert(object):
             except KeyError:
                 self.log.warn('Styles %s do not exist' % sname)
             else:
-                self.theme(fromstring(raw))
+                self.theme(etree.fromstring(raw))
 
         styles_loaded = False
         if sname is not None:
@@ -353,7 +354,7 @@ class Convert(object):
             except KeyError:
                 self.log.warn('Styles %s do not exist' % sname)
             else:
-                self.styles(fromstring(raw), fonts, self.theme)
+                self.styles(etree.fromstring(raw), fonts, self.theme)
                 styles_loaded = True
         if not styles_loaded:
             self.styles(None, fonts, self.theme)
@@ -364,7 +365,7 @@ class Convert(object):
             except KeyError:
                 self.log.warn('Numbering styles %s do not exist' % nname)
             else:
-                numbering(fromstring(raw), self.styles, self.docx.get_relationships(nname)[0])
+                numbering(etree.fromstring(raw), self.styles, self.docx.get_relationships(nname)[0])
 
         self.styles.resolve_numbering(numbering)
 
diff --git a/ebook_converter/ebooks/docx/writer/from_html.py b/ebook_converter/ebooks/docx/writer/from_html.py
index c9894a5..b13a0bb 100644
--- a/ebook_converter/ebooks/docx/writer/from_html.py
+++ b/ebook_converter/ebooks/docx/writer/from_html.py
@@ -1,22 +1,19 @@
+import collections
 import re
-from collections import Counter
 
 from ebook_converter.ebooks.docx.writer.container import create_skeleton, page_size, page_effective_area
-from ebook_converter.ebooks.docx.writer.styles import StylesManager, FloatSpec
-from ebook_converter.ebooks.docx.writer.links import LinksManager
-from ebook_converter.ebooks.docx.writer.images import ImagesManager
 from ebook_converter.ebooks.docx.writer.fonts import FontsManager
-from ebook_converter.ebooks.docx.writer.tables import Table
+from ebook_converter.ebooks.docx.writer.images import ImagesManager
+from ebook_converter.ebooks.docx.writer.links import LinksManager
 from ebook_converter.ebooks.docx.writer.lists import ListsManager
+from ebook_converter.ebooks.docx.writer.styles import StylesManager, FloatSpec
+from ebook_converter.ebooks.docx.writer.tables import Table
+from ebook_converter.ebooks.oeb import base
+from ebook_converter.ebooks.oeb import parse_utils
 from ebook_converter.ebooks.oeb.stylizer import Stylizer as Sz, Style as St
-from ebook_converter.ebooks.oeb.base import XPath, barename
 from ebook_converter.utils.localization import lang_as_iso639_1
 
 
-__license__ = 'GPL v3'
-__copyright__ = '2013, Kovid Goyal '
-
-
 def lang_for_tag(tag):
     for attr in ('lang', '{http://www.w3.org/XML/1998/namespace}lang'):
         val = lang_as_iso639_1(tag.get(attr))
@@ -140,7 +137,7 @@ class Block(object):
         self.numbering_id = None
         self.parent_items = None
         self.html_block = html_block
-        self.html_tag = barename(html_block.tag)
+        self.html_tag = parse_utils.barename(html_block.tag)
         self.float_spec = float_spec
         if float_spec is not None:
             float_spec.blocks.append(self)
@@ -387,7 +384,7 @@ class Blocks(object):
     def resolve_language(self):
         default_lang = self.styles_manager.document_lang
         for block in self.all_blocks:
-            count = Counter()
+            count = collections.Counter()
             for run in block.runs:
                 count[run.lang] += 1
             if count:
@@ -473,13 +470,13 @@ class Convert(object):
         self.abshref = self.images_manager.abshref = item.abshref
 
         self.current_lang = lang_for_tag(item.data) or self.styles_manager.document_lang
-        for i, body in enumerate(XPath('//h:body')(item.data)):
+        for i, body in enumerate(base.XPath('//h:body')(item.data)):
             with self.blocks:
                 self.blocks.top_bookmark = self.links_manager.bookmark_for_anchor(self.links_manager.top_anchor, self.current_item, body)
                 self.process_tag(body, stylizer, is_first_tag=i == 0)
 
     def process_tag(self, html_tag, stylizer, is_first_tag=False, float_spec=None):
-        tagname = barename(html_tag.tag)
+        tagname = parse_utils.barename(html_tag.tag)
         tag_style = stylizer.style(html_tag)
         ignore_tag_contents = tagname in {'script', 'style', 'title', 'meta'} or tag_style.is_hidden
         display = tag_style._get('display')
@@ -573,7 +570,7 @@ class Convert(object):
             text = html_tag.text
             if text:
                 block.add_text(text, tag_style, ignore_leading_whitespace=True, is_parent_style=True, link=self.current_link, lang=self.current_lang)
-            elif tagname == 'li' and len(html_tag) and barename(html_tag[0].tag) in ('ul', 'ol') and len(html_tag[0]):
+            elif tagname == 'li' and len(html_tag) and parse_utils.barename(html_tag[0].tag) in ('ul', 'ol') and len(html_tag[0]):
                 block.force_not_empty = True
 
     def add_inline_tag(self, tagname, html_tag, tag_style, stylizer):
diff --git a/ebook_converter/ebooks/fb2/fb2ml.py b/ebook_converter/ebooks/fb2/fb2ml.py
index 9db7965..0749be2 100644
--- a/ebook_converter/ebooks/fb2/fb2ml.py
+++ b/ebook_converter/ebooks/fb2/fb2ml.py
@@ -9,10 +9,10 @@ import uuid
 
 from lxml import etree
 
+from ebook_converter import constants as const
 from ebook_converter import prepare_string_for_xml
 from ebook_converter.constants_old import __appname__, __version__
 from ebook_converter.utils.localization import lang_as_iso639_1
-from ebook_converter.utils.xml_parse import safe_xml_fromstring
 from ebook_converter.utils.img import save_cover_data_to
 from ebook_converter.ebooks.oeb.base import urlnormalize
 from ebook_converter.polyglot.binary import as_base64_unicode
@@ -36,9 +36,10 @@ class FB2MLizer(object):
     def reset_state(self):
         # Used to ensure text and tags are always within 
 and 

         self.in_p = False
-        # Mapping of image names. OEB allows for images to have the same name but be stored
-        # in different directories. FB2 images are all in a flat layout so we rename all images
-        # into a sequential numbering system to ensure there are no collisions between image names.
+        # Mapping of image names. OEB allows for images to have the same name
+        # but be stored in different directories. FB2 images are all in a flat
+        # layout so we rename all images into a sequential numbering system to
+        # ensure there are no collisions between image names.
         self.image_hrefs = {}
         # Mapping of toc items and their
         self.toc = {}
@@ -68,13 +69,15 @@ class FB2MLizer(object):
         output = self.clean_text('\n'.join(output))
 
         if self.opts.pretty_print:
-            output = etree.tostring(safe_xml_fromstring(output), encoding='unicode', pretty_print=True)
+            output = etree.tostring(etree.fromstring(output),
+                                    encoding='unicode', pretty_print=True)
 
         return '\n' + output
 
     def clean_text(self, text):
         # Remove pointless tags, but keep their contents.
-        text = re.sub(r'(?mu)<(strong|emphasis|strikethrough|sub|sup)>(\s*)', r'\2', text)
+        text = re.sub(r'(?mu)<(strong|emphasis|strikethrough|sub|sup)>'
+                      r'(\s*)', r'\2', text)
 
         # Clean up paragraphs endings.
         text = re.sub(r'(?mu)\s+
', '
', text)
@@ -96,7 +99,8 @@ class FB2MLizer(object):
         text = re.sub(r'(?mu)\s*
', '\n
', text)
 
         # Put line breaks between paragraphs on a separate line.
-        text = re.sub(r'(?mu)\s*', r'\n', text)
+        text = re.sub(r'(?mu)\s*',
+                      r'\n', text)
         text = re.sub(r'(?mu)\s*
', '\n
', text)
 
         # Remove empty sections.
@@ -115,7 +119,9 @@ class FB2MLizer(object):
         metadata['title'] = self.oeb_book.metadata.title[0].value
         metadata['appname'] = __appname__
         metadata['version'] = __version__
-        metadata['date'] = '%i.%i.%i' % (datetime.now().day, datetime.now().month, datetime.now().year)
+        metadata['date'] = '%i.%i.%i' % (datetime.now().day,
+                                         datetime.now().month,
+                                         datetime.now().year)
         if self.oeb_book.metadata.language:
             lc = lang_as_iso639_1(self.oeb_book.metadata.language[0].value)
             if not lc:
@@ -143,31 +149,38 @@ class FB2MLizer(object):
                 author_middle = ' '.join(author_parts[1:-1])
                 author_last = author_parts[-1]
             metadata['author'] += ''
-            metadata['author'] += '%s' % prepare_string_for_xml(author_first)
+            metadata['author'] += ('%s' %
+                                   prepare_string_for_xml(author_first))
             if author_middle:
-                metadata['author'] += '%s' % prepare_string_for_xml(author_middle)
-            metadata['author'] += '%s' % prepare_string_for_xml(author_last)
+                metadata['author'] += ('%s' %
+                                       prepare_string_for_xml(author_middle))
+            metadata['author'] += ('%s' %
+                                   prepare_string_for_xml(author_last))
             metadata['author'] += ''
         if not metadata['author']:
-            metadata['author'] = ''
+            metadata['author'] = (''
+                                  '')
 
         metadata['keywords'] = ''
         tags = list(map(str, self.oeb_book.metadata.subject))
         if tags:
             tags = ', '.join(prepare_string_for_xml(x) for x in tags)
-            metadata['keywords'] = '%s'%tags
+            metadata['keywords'] = '%s' % tags
 
         metadata['sequence'] = ''
         if self.oeb_book.metadata.series:
             index = '1'
             if self.oeb_book.metadata.series_index:
                 index = self.oeb_book.metadata.series_index[0]
-            metadata['sequence'] = '' % (prepare_string_for_xml('%s' % self.oeb_book.metadata.series[0]), index)
+            seq = prepare_string_for_xml(str(self.oeb_book.metadata.series[0]))
+            metadata['sequence'] = ('' %
+                                    (seq, index))
 
         year = publisher = isbn = ''
         identifiers = self.oeb_book.metadata['identifier']
         for x in identifiers:
-            if x.get(OPF('scheme'), None).lower() == 'uuid' or str(x).startswith('urn:uuid:'):
+            if (x.get(OPF('scheme'), None).lower() == 'uuid' or
+                    str(x).startswith('urn:uuid:')):
                 metadata['id'] = str(x).split(':')[-1]
                 break
         if metadata['id'] is None:
@@ -179,22 +192,27 @@ class FB2MLizer(object):
         except IndexError:
             pass
         else:
-            year = '%s' % prepare_string_for_xml(date.value.partition('-')[0])
+            year = ('%s' %
+                    prepare_string_for_xml(date.value.partition('-')[0]))
 
         try:
             publisher = self.oeb_book.metadata['publisher'][0]
         except IndexError:
             pass
         else:
-            publisher = '%s' % prepare_string_for_xml(publisher.value)
+            publisher = ('%s' %
+                         prepare_string_for_xml(publisher.value))
 
         for x in identifiers:
             if x.get(OPF('scheme'), None).lower() == 'isbn':
                 isbn = '%s' % prepare_string_for_xml(x.value)
 
-        metadata['year'], metadata['isbn'], metadata['publisher'] = year, isbn, publisher
+        metadata['year'] = year
+        metadata['isbn'] = isbn
+        metadata['publisher'] = publisher
         for key, value in metadata.items():
-            if key not in ('author', 'cover', 'sequence', 'keywords', 'year', 'publisher', 'isbn'):
+            if key not in ('author', 'cover', 'sequence', 'keywords', 'year',
+                           'publisher', 'isbn'):
                 metadata[key] = prepare_string_for_xml(value)
 
         try:
@@ -203,7 +221,8 @@ class FB2MLizer(object):
             metadata['comments'] = ''
         else:
             from ebook_converter.utils.html2text import html2text
-            metadata['comments'] = '
{}
'.format(prepare_string_for_xml(html2text(comments.value).strip()))
+            annot = prepare_string_for_xml(html2text(comments.value).strip())
+            metadata['comments'] = f'
{annot}
'
 
         # Keep the indentation level of the description the same as the body.
         header = textwrap.dedent('''\
@@ -245,7 +264,9 @@ class FB2MLizer(object):
         cover_href = None
 
         # Get the raster cover if it's available.
-        if self.oeb_book.metadata.cover and str(self.oeb_book.metadata.cover[0]) in self.oeb_book.manifest.ids:
+        if (self.oeb_book.metadata.cover and
+                str(self.oeb_book.metadata.cover[0]) in
+                self.oeb_book.manifest.ids):
             id = str(self.oeb_book.metadata.cover[0])
             cover_item = self.oeb_book.manifest.ids[id]
             if cover_item.media_type in OEB_RASTER_IMAGES:
@@ -259,7 +280,8 @@ class FB2MLizer(object):
                 page_name = 'cover'
 
             if page_name:
-                cover_item = self.oeb_book.manifest.hrefs[self.oeb_book.guide[page_name].href]
+                key = self.oeb_book.guide[page_name].href
+                cover_item = self.oeb_book.manifest.hrefs[key]
                 # Get the first image in the page
                 for img in cover_item.xpath('//img'):
                     cover_href = cover_item.abshref(img.get('src'))
@@ -267,10 +289,11 @@ class FB2MLizer(object):
 
         if cover_href:
             # Only write the image tag if it is in the manifest.
-            if cover_href in self.oeb_book.manifest.hrefs and cover_href not in self.image_hrefs:
+            if (cover_href in self.oeb_book.manifest.hrefs and
+                    cover_href not in self.image_hrefs):
                 self.image_hrefs[cover_href] = 'img_%s' % len(self.image_hrefs)
-            return '' % self.image_hrefs[cover_href]
-
+            return ('' %
+                    self.image_hrefs[cover_href])
         return ''
 
     def get_text(self):
@@ -285,16 +308,20 @@ class FB2MLizer(object):
 
         for item in self.oeb_book.spine:
             self.log.debug('Converting %s to FictionBook2 XML' % item.href)
-            stylizer = Stylizer(item.data, item.href, self.oeb_book, self.opts, self.opts.output_profile)
+            stylizer = Stylizer(item.data, item.href, self.oeb_book, self.opts,
+                                self.opts.output_profile)
 
-            # Start a 
 if we must sectionize each file or if the TOC references this page
+            # Start a 
 if we must sectionize each file or if the TOC
+            # references this page
             page_section_open = False
-            if self.opts.sectionize == 'files' or None in self.toc.get(item.href, ()):
+            if (self.opts.sectionize == 'files' or
+                    None in self.toc.get(item.href, ())):
                 text.append('
')
                 page_section_open = True
                 self.section_level += 1
 
-            text += self.dump_text(item.data.find(XHTML('body')), stylizer, item)
+            text += self.dump_text(item.data.find(XHTML('body')), stylizer,
+                                   item)
 
             if page_section_open:
                 text.append('
')
@@ -309,20 +336,23 @@ class FB2MLizer(object):
         return ''.join(text)
 
     def fb2mlize_images(self):
-        '''
-        This function uses the self.image_hrefs dictionary mapping. It is populated by the dump_text function.
-        '''
+        """
+        This function uses the self.image_hrefs dictionary mapping. It is
+        populated by the dump_text function.
+        """
         from ebook_converter.ebooks.oeb.base import OEB_RASTER_IMAGES
 
         images = []
         for item in self.oeb_book.manifest:
-            # Don't write the image if it's not referenced in the document's text.
+            # Don't write the image if it's not referenced in the document's
+            # text.
             if item.href not in self.image_hrefs:
                 continue
             if item.media_type in OEB_RASTER_IMAGES:
                 try:
                     if item.media_type not in ('image/jpeg', 'image/png'):
-                        imdata = save_cover_data_to(item.data, compression_quality=70)
+                        imdata = save_cover_data_to(item.data,
+                                                    compression_quality=70)
                         raw_data = as_base64_unicode(imdata)
                         content_type = 'image/jpeg'
                     else:
@@ -330,11 +360,14 @@ class FB2MLizer(object):
                         content_type = item.media_type
                     # Don't put the encoded image on a single line.
                     step = 72
-                    data = '\n'.join(raw_data[i:i+step] for i in range(0, len(raw_data), step))
-                    images.append('%s' % (self.image_hrefs[item.href], content_type, data))
+                    data = '\n'.join(raw_data[i:i+step]
+                                     for i in range(0, len(raw_data), step))
+                    images.append('%s'
+                                  '' % (self.image_hrefs[item.href],
+                                                 content_type, data))
                 except Exception as e:
                     self.log.error('Error: Could not include file %s because '
-                        '%s.' % (item.href, e))
+                                   '%s.' % (item.href, e))
         return '\n'.join(images)
 
     def create_flat_toc(self, nodes, level):
@@ -391,26 +424,31 @@ class FB2MLizer(object):
 
     def dump_text(self, elem_tree, stylizer, page, tag_stack=[]):
         '''
-        This function is intended to be used in a recursive manner. dump_text will
-        run though all elements in the elem_tree and call itself on each element.
+        This function is intended to be used in a recursive manner. dump_text
+        will run though all elements in the elem_tree and call itself on each
+        element.
 
         self.image_hrefs will be populated by calling this function.
 
-        @param elem_tree: etree representation of XHTML content to be transformed.
+        @param elem_tree: etree representation of XHTML content to be
+            transformed.
         @param stylizer: Used to track the style of elements within the tree.
         @param page: OEB page used to determine absolute urls.
         @param tag_stack: List of open FB2 tags to take into account.
 
         @return: List of string representing the XHTML converted to FB2 markup.
         '''
-        from ebook_converter.ebooks.oeb.base import XHTML_NS, barename, namespace
+        from ebook_converter.ebooks.oeb.base import barename
+        from ebook_converter.ebooks.oeb.base import namespace
         elem = elem_tree
 
-        # Ensure what we are converting is not a string and that the fist tag is part of the XHTML namespace.
-        if not isinstance(elem_tree.tag, (str, bytes)) or namespace(elem_tree.tag) != XHTML_NS:
+        # Ensure what we are converting is not a string and that the fist tag
+        # is part of the XHTML namespace.
+        if (not isinstance(elem_tree.tag, (str, bytes)) or
+                namespace(elem_tree.tag) != const.XHTML_NS):
             p = elem.getparent()
-            if p is not None and isinstance(p.tag, (str, bytes)) and namespace(p.tag) == XHTML_NS \
-                    and elem.tail:
+            if (p is not None and isinstance(p.tag, (str, bytes)) and
+                    namespace(p.tag) == const.XHTML_NS and elem.tail):
                 return [elem.tail]
             return []
 
@@ -423,7 +461,8 @@ class FB2MLizer(object):
 
         # FB2 generated output.
         fb2_out = []
-        # FB2 tags in the order they are opened. This will be used to close the tags.
+        # FB2 tags in the order they are opened. This will be used to close
+        # the tags.
         tags = []
         # First tag in tree
         tag = barename(elem_tree.tag)
@@ -432,26 +471,31 @@ class FB2MLizer(object):
             ems = int(round((float(style.marginTop) / style.fontSize) - 1))
             if ems < 0:
                 ems = 0
-        except:
+        except Exception:
             ems = 0
 
         # Convert TOC entries to s and add <section>s
         if self.opts.sectionize == 'toc':
-            # A section cannot be a child of any other element than another section,
-            # so leave the tag alone if there are parents
+            # A section cannot be a child of any other element than another
+            # section, so leave the tag alone if there are parents
             if not tag_stack:
-                # There are two reasons to start a new section here: the TOC pointed to
-                # this page (then we use the first non-<body> on the page as a <title>), or
-                # the TOC pointed to a specific element
+                # There are two reasons to start a new section here: the TOC
+                # pointed to this page (then we use the first non-<body> on
+                # the page as a <title>), or the TOC pointed to a specific
+                # element
                 newlevel = 0
                 toc_entry = self.toc.get(page.href, None)
                 if toc_entry is not None:
                     if None in toc_entry:
-                        if tag != 'body' and hasattr(elem_tree, 'text') and elem_tree.text:
+                        if (tag != 'body' and hasattr(elem_tree, 'text') and
+                                elem_tree.text):
                             newlevel = 1
                             self.toc[page.href] = None
-                    if not newlevel and elem_tree.attrib.get('id', None) is not None:
-                        newlevel = toc_entry.get(elem_tree.attrib.get('id', None), None)
+                    if (not newlevel and
+                            elem_tree.attrib.get('id', None) is not None):
+                        newlevel = toc_entry.get(elem_tree.attrib.get('id',
+                                                                      None),
+                                                 None)
 
                 # Start a new section if necessary
                 if newlevel:
@@ -463,13 +507,14 @@ class FB2MLizer(object):
                     fb2_out.append('<title>')
                     tags.append('title')
             if self.section_level == 0:
-                # If none of the prior processing made a section, make one now to be FB2 spec compliant
+                # If none of the prior processing made a section, make one now
+                # to be FB2 spec compliant
                 fb2_out.append('<section>')
                 self.section_level += 1
 
         # Process the XHTML tag and styles. Converted to an FB2 tag.
-        # Use individual if statement not if else. There can be
-        # only one XHTML tag but it can have multiple styles.
+        # Use individual if statement not if else. There can be only one XHTML
+        # tag but it can have multiple styles.
         if tag == 'img' and elem_tree.attrib.get('src', None):
             # Only write the image tag if it is in the manifest.
             ihref = urlnormalize(page.abshref(elem_tree.attrib['src']))
@@ -479,7 +524,8 @@ class FB2MLizer(object):
                 p_txt, p_tag = self.ensure_p()
                 fb2_out += p_txt
                 tags += p_tag
-                fb2_out.append('<image l:href="#%s"/>' % self.image_hrefs[ihref])
+                fb2_out.append('<image l:href="#%s"/>' %
+                               self.image_hrefs[ihref])
             else:
                 self.log.warn(u'Ignoring image not in manifest: %s' % ihref)
         if tag in ('br', 'hr') or ems >= 1:
@@ -513,7 +559,8 @@ class FB2MLizer(object):
                 p_txt, p_tag = self.ensure_p()
                 fb2_out += p_txt
                 tags += p_tag
-                fb2_out.append('<a l:href="%s">' % urlnormalize(elem_tree.attrib['href']))
+                fb2_out.append('<a l:href="%s">' %
+                               urlnormalize(elem_tree.attrib['href']))
                 tags.append('a')
         if tag == 'b' or style['font-weight'] in ('bold', 'bolder'):
             s_out, s_tags = self.handle_simple_tag('strong', tag_stack+tags)
@@ -523,8 +570,10 @@ class FB2MLizer(object):
             s_out, s_tags = self.handle_simple_tag('emphasis', tag_stack+tags)
             fb2_out += s_out
             tags += s_tags
-        if tag in ('del', 'strike') or style['text-decoration'] == 'line-through':
-            s_out, s_tags = self.handle_simple_tag('strikethrough', tag_stack+tags)
+        if (tag in ('del', 'strike') or
+                style['text-decoration'] == 'line-through'):
+            s_out, s_tags = self.handle_simple_tag('strikethrough',
+                                                   tag_stack+tags)
             fb2_out += s_out
             tags += s_tags
         if tag == 'sub':
@@ -552,7 +601,8 @@ class FB2MLizer(object):
         tags.reverse()
         fb2_out += self.close_tags(tags)
 
-        # Process element text that comes after the close of the XHTML tag but before the next XHTML tag.
+        # Process element text that comes after the close of the XHTML tag but
+        # before the next XHTML tag.
         if hasattr(elem_tree, 'tail') and elem_tree.tail:
             if not self.in_p:
                 fb2_out.append('<p>')
diff --git a/ebook_converter/ebooks/htmlz/oeb2html.py b/ebook_converter/ebooks/htmlz/oeb2html.py
index f0c5640..42b0742 100644
--- a/ebook_converter/ebooks/htmlz/oeb2html.py
+++ b/ebook_converter/ebooks/htmlz/oeb2html.py
@@ -9,8 +9,9 @@ from functools import partial
 from lxml import html
 
 from ebook_converter import prepare_string_for_xml
-from ebook_converter.ebooks.oeb.base import (
-    XHTML, XHTML_NS, SVG_NS, barename, namespace, OEB_IMAGES, XLINK, rewrite_links, urlnormalize)
+from ebook_converter import constants as const
+from ebook_converter.ebooks.oeb import base
+from ebook_converter.ebooks.oeb import parse_utils
 from ebook_converter.ebooks.oeb.stylizer import Stylizer
 from ebook_converter.utils.logging import default_log
 from ebook_converter.polyglot.builtins import as_bytes
@@ -61,9 +62,9 @@ class OEB2HTML(object):
         for item in oeb_book.spine:
             self.log.debug('Converting %s to HTML...' % item.href)
             self.rewrite_ids(item.data, item)
-            rewrite_links(item.data, partial(self.rewrite_link, page=item))
+            base.rewrite_links(item.data, partial(self.rewrite_link, page=item))
             stylizer = Stylizer(item.data, item.href, oeb_book, self.opts)
-            output += self.dump_text(item.data.find(XHTML('body')), stylizer, item)
+            output += self.dump_text(item.data.find(base.tag('xhtml', 'body')), stylizer, item)
             output.append('\n\n')
         output.append('</body></html>')
         return ''.join(output)
@@ -80,7 +81,7 @@ class OEB2HTML(object):
 
     def map_resources(self, oeb_book):
         for item in oeb_book.manifest:
-            if item.media_type in OEB_IMAGES:
+            if item.media_type in base.OEB_IMAGES:
                 if item.href not in self.images:
                     ext = os.path.splitext(item.href)[1]
                     fname = '%s%s' % (len(self.images), ext)
@@ -88,9 +89,9 @@ class OEB2HTML(object):
                     self.images[item.href] = fname
             if item in oeb_book.spine:
                 self.get_link_id(item.href)
-                root = item.data.find(XHTML('body'))
+                root = item.data.find(base.tag('xhtml', 'body'))
                 link_attrs = set(html.defs.link_attrs)
-                link_attrs.add(XLINK('href'))
+                link_attrs.add(base.tag('xlink', 'href'))
                 for el in root.iter():
                     attribs = el.attrib
                     try:
@@ -108,7 +109,7 @@ class OEB2HTML(object):
     def rewrite_link(self, url, page=None):
         if not page:
             return url
-        abs_url = page.abshref(urlnormalize(url))
+        abs_url = page.abshref(base.urlnormalize(url))
         if abs_url in self.images:
             return 'images/%s' % self.images[abs_url]
         if abs_url in self.links:
@@ -121,7 +122,7 @@ class OEB2HTML(object):
                 tag = el.tag
             except UnicodeDecodeError:
                 continue
-            if tag == XHTML('body'):
+            if tag == base.tag('xhtml', 'body'):
                 el.attrib['id'] = self.get_link_id(page.href)[1:]
                 continue
             if 'id' in el.attrib:
@@ -156,9 +157,9 @@ class OEB2HTMLNoCSSizer(OEB2HTML):
 
         # We can only processes tags. If there isn't a tag return any text.
         if not isinstance(elem.tag, (str, bytes)) \
-           or namespace(elem.tag) not in (XHTML_NS, SVG_NS):
+           or parse_utils.namespace(elem.tag) not in (const.XHTML_NS, const.SVG_NS):
             p = elem.getparent()
-            if p is not None and isinstance(p.tag, (str, bytes)) and namespace(p.tag) in (XHTML_NS, SVG_NS) \
+            if p is not None and isinstance(p.tag, (str, bytes)) and parse_utils.namespace(p.tag) in (const.XHTML_NS, const.SVG_NS) \
                     and elem.tail:
                 return [elem.tail]
             return ['']
@@ -167,7 +168,7 @@ class OEB2HTMLNoCSSizer(OEB2HTML):
         text = ['']
         style = stylizer.style(elem)
         tags = []
-        tag = barename(elem.tag)
+        tag = parse_utils.barename(elem.tag)
         attribs = elem.attrib
 
         if tag == 'body':
@@ -245,9 +246,9 @@ class OEB2HTMLInlineCSSizer(OEB2HTML):
 
         # We can only processes tags. If there isn't a tag return any text.
         if not isinstance(elem.tag, (str, bytes)) \
-           or namespace(elem.tag) not in (XHTML_NS, SVG_NS):
+           or parse_utils.namespace(elem.tag) not in (const.XHTML_NS, const.SVG_NS):
             p = elem.getparent()
-            if p is not None and isinstance(p.tag, (str, bytes)) and namespace(p.tag) in (XHTML_NS, SVG_NS) \
+            if p is not None and isinstance(p.tag, (str, bytes)) and parse_utils.namespace(p.tag) in (const.XHTML_NS, const.SVG_NS) \
                     and elem.tail:
                 return [elem.tail]
             return ['']
@@ -256,7 +257,7 @@ class OEB2HTMLInlineCSSizer(OEB2HTML):
         text = ['']
         style = stylizer.style(elem)
         tags = []
-        tag = barename(elem.tag)
+        tag = parse_utils.barename(elem.tag)
         attribs = elem.attrib
 
         style_a = '%s' % style
@@ -327,9 +328,9 @@ class OEB2HTMLClassCSSizer(OEB2HTML):
         for item in oeb_book.spine:
             self.log.debug('Converting %s to HTML...' % item.href)
             self.rewrite_ids(item.data, item)
-            rewrite_links(item.data, partial(self.rewrite_link, page=item))
+            base.rewrite_links(item.data, partial(self.rewrite_link, page=item))
             stylizer = Stylizer(item.data, item.href, oeb_book, self.opts)
-            output += self.dump_text(item.data.find(XHTML('body')), stylizer, item)
+            output += self.dump_text(item.data.find(base.tag('xhtml', 'body')), stylizer, item)
             output.append('\n\n')
         if self.opts.htmlz_class_style == 'external':
             css = u'<link href="style.css" rel="stylesheet" type="text/css" />'
@@ -348,9 +349,9 @@ class OEB2HTMLClassCSSizer(OEB2HTML):
 
         # We can only processes tags. If there isn't a tag return any text.
         if not isinstance(elem.tag, (str, bytes)) \
-           or namespace(elem.tag) not in (XHTML_NS, SVG_NS):
+           or parse_utils.namespace(elem.tag) not in (const.XHTML_NS, const.SVG_NS):
             p = elem.getparent()
-            if p is not None and isinstance(p.tag, (str, bytes)) and namespace(p.tag) in (XHTML_NS, SVG_NS) \
+            if p is not None and isinstance(p.tag, (str, bytes)) and parse_utils.namespace(p.tag) in (const.XHTML_NS, const.SVG_NS) \
                     and elem.tail:
                 return [elem.tail]
             return ['']
@@ -358,7 +359,7 @@ class OEB2HTMLClassCSSizer(OEB2HTML):
         # Setup our variables.
         text = ['']
         tags = []
-        tag = barename(elem.tag)
+        tag = parse_utils.barename(elem.tag)
         attribs = elem.attrib
 
         if tag == 'body':
diff --git a/ebook_converter/ebooks/metadata/fb2.py b/ebook_converter/ebooks/metadata/fb2.py
index f38c8e0..43bc493 100644
--- a/ebook_converter/ebooks/metadata/fb2.py
+++ b/ebook_converter/ebooks/metadata/fb2.py
@@ -1,38 +1,32 @@
 """
 Read meta information from fb2 files
 """
-import os, random
-from functools import partial
-from string import ascii_letters, digits
+import functools
+import os
+import random
+import string
 
 from lxml import etree
 
 from ebook_converter.utils.date import parse_only_date
 from ebook_converter.utils.img import save_cover_data_to
-from ebook_converter.utils.xml_parse import safe_xml_fromstring
 from ebook_converter.utils.imghdr import identify
-from ebook_converter import guess_type, guess_all_extensions, prints, force_unicode
+from ebook_converter import guess_type, guess_all_extensions, prints, \
+        force_unicode
 from ebook_converter.ebooks.metadata import MetaInformation, check_isbn
 from ebook_converter.ebooks.chardet import xml_to_unicode
 from ebook_converter.polyglot.binary import as_base64_unicode
 
 
-__license__ = 'GPL v3'
-__copyright__ = ('2011, Roman Mukhin <ramses_ru at hotmail.com>, '
-                 '2008, Anatoly Shipitsin <norguhtar at gmail.com>')
+NAMESPACES = {'fb2': 'http://www.gribuser.ru/xml/fictionbook/2.0',
+              'fb21': 'http://www.gribuser.ru/xml/fictionbook/2.1',
+              'xlink': 'http://www.w3.org/1999/xlink'}
 
-
-NAMESPACES = {
-    'fb2'   :   'http://www.gribuser.ru/xml/fictionbook/2.0',
-    'fb21'  :   'http://www.gribuser.ru/xml/fictionbook/2.1',
-    'xlink' :   'http://www.w3.org/1999/xlink'
-}
-
-tostring = partial(etree.tostring, method='text', encoding='unicode')
+tostring = functools.partial(etree.tostring, method='text', encoding='unicode')
 
 
 def XLINK(tag):
-    return '{%s}%s'%(NAMESPACES['xlink'], tag)
+    return '{%s}%s' % (NAMESPACES['xlink'], tag)
 
 
 class Context(object):
@@ -52,7 +46,7 @@ class Context(object):
         return etree.XPath(*args, namespaces=self.namespaces)
 
     def get_or_create(self, parent, tag, attribs={}, at_start=True):
-        xpathstr='./fb:'+tag
+        xpathstr = './fb:'+tag
         for n, v in attribs.items():
             xpathstr += '[@%s="%s"]' % (n, v)
         ans = self.XPath(xpathstr)(parent)
@@ -73,7 +67,7 @@ class Context(object):
 
     def clear_meta_tags(self, doc, tag):
         for parent in ('title-info', 'src-title-info', 'publish-info'):
-            for x in self.XPath('//fb:%s/fb:%s'%(parent, tag))(doc):
+            for x in self.XPath('//fb:%s/fb:%s' % (parent, tag))(doc):
                 x.getparent().remove(x)
 
     def text2fb2(self, parent, text):
@@ -117,42 +111,41 @@ def get_metadata(stream):
         book_title = str(book_title)
     else:
         book_title = force_unicode(os.path.splitext(
-            os.path.basename(getattr(stream, 'name',
-                'Unknown')))[0])
+            os.path.basename(getattr(stream, 'name', 'Unknown')))[0])
     mi = MetaInformation(book_title, authors)
 
     try:
         _parse_cover(root, mi, ctx)
-    except:
+    except Exception:
         pass
     try:
         _parse_comments(root, mi, ctx)
-    except:
+    except Exception:
         pass
     try:
         _parse_tags(root, mi, ctx)
-    except:
+    except Exception:
         pass
     try:
         _parse_series(root, mi, ctx)
-    except:
+    except Exception:
         pass
     try:
         _parse_isbn(root, mi, ctx)
-    except:
+    except Exception:
         pass
     try:
         _parse_publisher(root, mi, ctx)
-    except:
+    except Exception:
         pass
     try:
         _parse_pubdate(root, mi, ctx)
-    except:
+    except Exception:
         pass
 
     try:
         _parse_language(root, mi, ctx)
-    except:
+    except Exception:
         pass
 
     return mi
@@ -160,11 +153,11 @@ def get_metadata(stream):
 
 def _parse_authors(root, ctx):
     authors = []
-    # pick up authors but only from 1 secrion <title-info>; otherwise it is not consistent!
-    # Those are fallbacks: <src-title-info>, <document-info>
+    # pick up authors but only from 1 secrion <title-info>; otherwise it is
+    # not consistent! Those are fallbacks: <src-title-info>, <document-info>
     author = None
     for author_sec in ['title-info', 'src-title-info', 'document-info']:
-        for au in ctx.XPath('//fb:%s/fb:author'%author_sec)(root):
+        for au in ctx.XPath('//fb:%s/fb:author' % author_sec)(root):
             author = _parse_author(au, ctx)
             if author:
                 authors.append(author)
@@ -207,24 +200,26 @@ def _parse_book_title(root, ctx):
     xp_ti = '//fb:title-info/fb:book-title/text()'
     xp_pi = '//fb:publish-info/fb:book-title/text()'
     xp_si = '//fb:src-title-info/fb:book-title/text()'
-    book_title = ctx.XPath('normalize-space(%s|%s|%s)' % (xp_ti, xp_pi, xp_si))(root)
+    book_title = ctx.XPath('normalize-space(%s|%s|%s)' %
+                           (xp_ti, xp_pi, xp_si))(root)
 
     return book_title
 
 
 def _parse_cover(root, mi, ctx):
     # pickup from <title-info>, if not exists it fallbacks to <src-title-info>
-    imgid = ctx.XPath('substring-after(string(//fb:coverpage/fb:image/@xlink:href), "#")')(root)
+    imgid = ctx.XPath('substring-after(string(//fb:coverpage/fb:image/'
+                      '@xlink:href), "#")')(root)
     if imgid:
         try:
             _parse_cover_data(root, imgid, mi, ctx)
-        except:
+        except Exception:
             pass
 
 
 def _parse_cover_data(root, imgid, mi, ctx):
     from ebook_converter.ebooks.fb2 import base64_decode
-    elm_binary = ctx.XPath('//fb:binary[@id="%s"]'%imgid)(root)
+    elm_binary = ctx.XPath('//fb:binary[@id="%s"]' % imgid)(root)
     if elm_binary:
         mimetype = elm_binary[0].get('content-type', 'image/jpeg')
         mime_extensions = guess_all_extensions(mimetype)
@@ -241,12 +236,13 @@ def _parse_cover_data(root, imgid, mi, ctx):
                 fmt = identify(cdata)[0]
                 mi.cover_data = (fmt, cdata)
         else:
-            prints("WARNING: Unsupported coverpage mime-type '%s' (id=#%s)" % (mimetype, imgid))
+            prints("WARNING: Unsupported coverpage mime-type '%s' (id=#%s)" %
+                   (mimetype, imgid))
 
 
 def _parse_tags(root, mi, ctx):
-    # pick up genre but only from 1 secrion <title-info>; otherwise it is not consistent!
-    # Those are fallbacks: <src-title-info>
+    # pick up genre but only from 1 secrion <title-info>; otherwise it is not
+    # consistent! Those are fallbacks: <src-title-info>
     for genre_sec in ['title-info', 'src-title-info']:
         # -- i18n Translations-- ?
         tags = ctx.XPath('//fb:%s/fb:genre/text()' % genre_sec)(root)
@@ -267,16 +263,20 @@ def _parse_series(root, mi, ctx):
         mi.series = elms_sequence[0].get('name', None)
         if mi.series:
             try:
-                mi.series_index = float('.'.join(elms_sequence[0].get('number', None).split()[:2]))
+                i = float('.'.join(elms_sequence[0].get('number',
+                                                        None).split()[:2]))
+                mi.series_index = i
             except Exception:
                 pass
 
 
 def _parse_isbn(root, mi, ctx):
-    # some people try to put several isbn in this field, but it is not allowed.  try to stick to the 1-st one in this case
+    # some people try to put several isbn in this field, but it is not
+    # allowed. try to stick to the 1-st one in this case
     isbn = ctx.XPath('normalize-space(//fb:publish-info/fb:isbn/text())')(root)
     if isbn:
-        # some people try to put several isbn in this field, but it is not allowed.  try to stick to the 1-st one in this case
+        # some people try to put several isbn in this field, but it is not
+        # allowed. try to stick to the 1-st one in this case
         if ',' in isbn:
             isbn = isbn[:isbn.index(',')]
         if check_isbn(isbn):
@@ -284,9 +284,11 @@ def _parse_isbn(root, mi, ctx):
 
 
 def _parse_comments(root, mi, ctx):
-    # pick up annotation but only from 1 section <title-info>;  fallback: <src-title-info>
+    # pick up annotation but only from 1 section <title-info>;
+    # fallback: <src-title-info>
     for annotation_sec in ['title-info', 'src-title-info']:
-        elms_annotation = ctx.XPath('//fb:%s/fb:annotation' % annotation_sec)(root)
+        elms_annotation = ctx.XPath('//fb:%s/fb:annotation' %
+                                    annotation_sec)(root)
         if elms_annotation:
             mi.comments = tostring(elms_annotation[0])
             # TODO: tags i18n, xslt?
@@ -294,7 +296,8 @@ def _parse_comments(root, mi, ctx):
 
 
 def _parse_publisher(root, mi, ctx):
-    publisher = ctx.XPath('string(//fb:publish-info/fb:publisher/text())')(root)
+    publisher = ctx.XPath('string(//fb:publish-info/fb:publisher/'
+                          'text())')(root)
     if publisher:
         mi.publisher = publisher
 
@@ -315,7 +318,7 @@ def _parse_language(root, mi, ctx):
 
 def _get_fbroot(raw):
     raw = xml_to_unicode(raw, strip_encoding_pats=True)[0]
-    root = safe_xml_fromstring(raw)
+    root = etree.fromstring(raw)
     return ensure_namespace(root)
 
 
@@ -348,10 +351,12 @@ def _set_authors(title_info, mi, ctx):
                 ctx.create_tag(atag, 'first-name').text = author_parts[0]
                 author_parts = author_parts[1:]
                 if len(author_parts) > 1:
-                    ctx.create_tag(atag, 'middle-name', at_start=False).text = author_parts[0]
+                    ctx.create_tag(atag, 'middle-name',
+                                   at_start=False).text = author_parts[0]
                     author_parts = author_parts[1:]
                 if author_parts:
-                    ctx.create_tag(atag, 'last-name', at_start=False).text = ' '.join(author_parts)
+                    a = ' '.join(author_parts)
+                    ctx.create_tag(atag, 'last-name', at_start=False).text = a
 
 
 def _set_tags(title_info, mi, ctx):
@@ -368,12 +373,12 @@ def _set_series(title_info, mi, ctx):
         seq = ctx.get_or_create(title_info, 'sequence')
         seq.set('name', mi.series)
         try:
-            seq.set('number', '%g'%mi.series_index)
-        except:
+            seq.set('number', '%g' % mi.series_index)
+        except Exception:
             seq.set('number', '1')
 
 
-def _rnd_name(size=8, chars=ascii_letters + digits):
+def _rnd_name(size=8, chars=string.ascii_letters + string.digits):
     return ''.join(random.choice(chars) for x in range(size))
 
 
@@ -396,7 +401,9 @@ def _set_cover(title_info, mi, ctx):
             cim_filename = _rnd_pic_file_name('cover')
             cim_tag.attrib[XLINK('href')] = '#' + cim_filename
         fb2_root = cim_tag.getroottree().getroot()
-        cim_binary = ctx.get_or_create(fb2_root, 'binary', attribs={'id': cim_filename}, at_start=False)
+        cim_binary = ctx.get_or_create(fb2_root, 'binary',
+                                       attribs={'id': cim_filename},
+                                       at_start=False)
         cim_binary.attrib['content-type'] = 'image/jpeg'
         cim_binary.text = _encode_into_jpeg(mi.cover_data[1])
 
@@ -425,7 +432,8 @@ def set_metadata(stream, mi, apply_null=False, update_timestamp=False):
     # single quotes in xml declaration. Sigh. See
     # https://www.mobileread.com/forums/showthread.php?p=2273184#post2273184
     raw = b'<?xml version="1.0" encoding="UTF-8"?>\n'
-    raw += etree.tostring(root, method='xml', encoding='utf-8', xml_declaration=False)
+    raw += etree.tostring(root, method='xml', encoding='utf-8',
+                          xml_declaration=False)
 
     stream.seek(0)
     stream.truncate()
@@ -449,6 +457,7 @@ def ensure_namespace(doc):
     if bare_tags:
         import re
         raw = etree.tostring(doc, encoding='unicode')
-        raw = re.sub(r'''<(description|body)\s+xmlns=['"]['"]>''', r'<\1>', raw)
-        doc = safe_xml_fromstring(raw)
+        raw = re.sub(r'''<(description|body)\s+xmlns=['"]['"]>''', r'<\1>',
+                     raw)
+        doc = etree.fromstring(raw)
     return doc
diff --git a/ebook_converter/ebooks/metadata/opf2.py b/ebook_converter/ebooks/metadata/opf2.py
index 586164b..9c85d23 100644
--- a/ebook_converter/ebooks/metadata/opf2.py
+++ b/ebook_converter/ebooks/metadata/opf2.py
@@ -12,29 +12,31 @@ import sys
 import unittest
 import urllib.parse
 import uuid
+import traceback
+import textwrap
 
 from lxml import etree
+from lxml.builder import ElementMaker
 
+from ebook_converter.ebooks.oeb import base
+from ebook_converter import constants as const
 from ebook_converter.ebooks import escape_xpath_attr
 from ebook_converter.constants_old import __appname__, __version__, \
         filesystem_encoding
 from ebook_converter.ebooks.metadata.toc import TOC
-from ebook_converter.ebooks.metadata.utils import parse_opf, pretty_print_opf as _pretty_print
-from ebook_converter.ebooks.metadata import string_to_authors, MetaInformation, check_isbn
+from ebook_converter.ebooks.metadata.utils import parse_opf, \
+        pretty_print_opf as _pretty_print
+from ebook_converter.ebooks.metadata import string_to_authors, \
+        MetaInformation, check_isbn
 from ebook_converter.ebooks.metadata.book.base import Metadata
 from ebook_converter.utils.date import parse_date, isoformat
 from ebook_converter.utils.localization import get_lang, canonicalize_lang
 from ebook_converter import prints, guess_type
 from ebook_converter.utils.cleantext import clean_ascii_chars, clean_xml_chars
 from ebook_converter.utils.config import tweaks
-from ebook_converter.utils.xml_parse import safe_xml_fromstring
 from ebook_converter.polyglot.urllib import unquote
 
 
-__license__ = 'GPL v3'
-__copyright__ = '2008, Kovid Goyal kovid@kovidgoyal.net'
-__docformat__ = 'restructuredtext en'
-
 pretty_print_opf = False
 
 
@@ -73,7 +75,7 @@ class Resource(object):  # {{{
         self.fragment = ''
         try:
             self.mime_type = guess_type(href_or_path)[0]
-        except:
+        except Exception:
             self.mime_type = None
         if self.mime_type is None:
             self.mime_type = 'application/octet-stream'
@@ -94,17 +96,21 @@ class Resource(object):  # {{{
                 if isinstance(pc, str):
                     pc = pc.encode('utf-8')
                 pc = pc.decode('utf-8')
-                self.path = os.path.abspath(os.path.join(basedir, pc.replace('/', os.sep)))
+                self.path = os.path.abspath(os.path.join(basedir,
+                                                         pc.replace('/',
+                                                                    os.sep)))
                 self.fragment = url[-1]
 
     def href(self, basedir=None):
-        '''
-        Return a URL pointing to this resource. If it is a file on the filesystem
-        the URL is relative to `basedir`.
+        """
+        Return a URL pointing to this resource. If it is a file on the
+        filesystem the URL is relative to `basedir`.
 
-        `basedir`: If None, the basedir of this resource is used (see :method:`set_basedir`).
-        If this resource has no basedir, then the current working directory is used as the basedir.
-        '''
+        `basedir`: If None, the basedir of this resource is used (see
+            :method:`set_basedir`).
+        If this resource has no basedir, then the current working directory is
+        used as the basedir.
+        """
         if basedir is None:
             if self._basedir:
                 basedir = self._basedir
@@ -117,7 +123,8 @@ class Resource(object):  # {{{
             return frag
         try:
             rpath = os.path.relpath(self.path, basedir)
-        except ValueError:  # On windows path and basedir could be on different drives
+        except ValueError:
+            # On windows path and basedir could be on different drives
             rpath = self.path
         if isinstance(rpath, bytes):
             rpath = rpath.decode(filesystem_encoding)
@@ -130,7 +137,7 @@ class Resource(object):  # {{{
         return self._basedir
 
     def __repr__(self):
-        return 'Resource(%s, %s)'%(repr(self.path), repr(self.href()))
+        return 'Resource(%s, %s)' % (repr(self.path), repr(self.href()))
 
 # }}}
 
@@ -155,7 +162,7 @@ class ResourceCollection(object):  # {{{
 
     def __str__(self):
         resources = map(repr, self)
-        return '[%s]'%', '.join(resources)
+        return '[%s]' % ', '.join(resources)
     __unicode__ = __str__
 
     def __repr__(self):
@@ -211,7 +218,8 @@ class ManifestItem(Resource):  # {{{
         self.mime_type = val
 
     def __unicode__representation__(self):
-        return u'<item id="%s" href="%s" media-type="%s" />'%(self.id, self.href(), self.media_type)
+        return (u'<item id="%s" href="%s" media-type="%s" />' %
+                (self.id, self.href(), self.media_type))
 
     __str__ = __unicode__representation__
 
@@ -223,7 +231,7 @@ class ManifestItem(Resource):  # {{{
             return self.href()
         if index == 1:
             return self.media_type
-        raise IndexError('%d out of bounds.'%index)
+        raise IndexError('%d out of bounds.' % index)
 
 # }}}
 
@@ -234,7 +242,7 @@ class Manifest(ResourceCollection):  # {{{
         self.append(ManifestItem.from_opf_manifest_item(item, dir))
         id = item.get('id', '')
         if not id:
-            id = 'id%d'%self.next_id
+            id = 'id%d' % self.next_id
         self[-1].id = id
         self.next_id += 1
 
@@ -250,15 +258,16 @@ class Manifest(ResourceCollection):  # {{{
 
     @staticmethod
     def from_paths(entries):
-        '''
-        `entries`: List of (path, mime-type) If mime-type is None it is autodetected
-        '''
+        """
+        `entries`: List of (path, mime-type) If mime-type is None it is
+        autodetected
+        """
         m = Manifest()
         for path, mt in entries:
             mi = ManifestItem(path, is_path=True)
             if mt:
                 mi.mime_type = mt
-            mi.id = 'id%d'%m.next_id
+            mi.id = 'id%d' % m.next_id
             m.next_id += 1
             m.append(mi)
         return m
@@ -267,7 +276,7 @@ class Manifest(ResourceCollection):  # {{{
         mi = ManifestItem(path, is_path=True)
         if mime_type:
             mi.mime_type = mime_type
-        mi.id = 'id%d'%self.next_id
+        mi.id = 'id%d' % self.next_id
         self.next_id += 1
         self.append(mi)
         return mi.id
@@ -318,13 +327,13 @@ class Spine(ResourceCollection):  # {{{
     def from_opf_spine_element(itemrefs, manifest):
         s = Spine(manifest)
         seen = set()
-        path_map = {i.id:i.path for i in s.manifest}
+        path_map = {i.id: i.path for i in s.manifest}
         for itemref in itemrefs:
             idref = itemref.get('idref', None)
             if idref is not None:
                 path = path_map.get(idref)
                 if path and path not in seen:
-                    r = Spine.Item(lambda x:idref, path, is_path=True)
+                    r = Spine.Item(lambda x: idref, path, is_path=True)
                     r.is_linear = itemref.get('linear', 'yes') == 'yes'
                     r.idref = idref
                     s.append(r)
@@ -336,8 +345,9 @@ class Spine(ResourceCollection):  # {{{
         s = Spine(manifest)
         for path in paths:
             try:
-                s.append(Spine.Item(s.manifest.id_for_path, path, is_path=True))
-            except:
+                s.append(Spine.Item(s.manifest.id_for_path, path,
+                                    is_path=True))
+            except Exception:
                 continue
         return s
 
@@ -346,10 +356,10 @@ class Spine(ResourceCollection):  # {{{
         self.manifest = manifest
 
     def replace(self, start, end, ids):
-        '''
-        Replace the items between start (inclusive) and end (not inclusive) with
+        """
+        Replace the items between start (inclusive) and end (not inclusive)
         with the items identified by ids. ids can be a list of any length.
-        '''
+        """
         items = []
         for id in ids:
             path = self.manifest.path_for_id(id)
@@ -381,16 +391,18 @@ class Guide(ResourceCollection):  # {{{
 
         @staticmethod
         def from_opf_resource_item(ref, basedir):
-            title, href, type = ref.get('title', ''), ref.get('href'), ref.get('type')
+            title = ref.get('title', '')
+            href = ref.get('href')
+            type = ref.get('type')
             res = Guide.Reference(href, basedir, is_path=True)
             res.title = title
             res.type = type
             return res
 
         def __repr__(self):
-            ans = '<reference type="%s" href="%s" '%(self.type, self.href())
+            ans = '<reference type="%s" href="%s" ' % (self.type, self.href())
             if self.title:
-                ans += 'title="%s" '%self.title
+                ans += 'title="%s" ' % self.title
             return ans + '/>'
 
     @staticmethod
@@ -400,7 +412,7 @@ class Guide(ResourceCollection):  # {{{
             try:
                 ref = Guide.Reference.from_opf_resource_item(ref, base_dir)
                 coll.append(ref)
-            except:
+            except Exception:
                 continue
         return coll
 
@@ -408,7 +420,8 @@ class Guide(ResourceCollection):  # {{{
         for i in tuple(self):
             if 'cover' in i.type.lower():
                 self.remove(i)
-        for typ in ('cover', 'other.ms-coverimage-standard', 'other.ms-coverimage'):
+        for typ in ('cover', 'other.ms-coverimage-standard',
+                    'other.ms-coverimage'):
             self.append(Guide.Reference(path, is_path=True))
             self[-1].type = typ
             self[-1].title = ''
@@ -419,12 +432,12 @@ class Guide(ResourceCollection):  # {{{
 class MetadataField(object):
 
     def __init__(self, name, is_dc=True, formatter=None, none_is=None,
-            renderer=lambda x: str(x)):
-        self.name      = name
-        self.is_dc     = is_dc
+                 renderer=lambda x: str(x)):
+        self.name = name
+        self.is_dc = is_dc
         self.formatter = formatter
-        self.none_is   = none_is
-        self.renderer  = renderer
+        self.none_is = none_is
+        self.renderer = renderer
 
     def __real_get__(self, obj, type=None):
         ans = obj.get_metadata_element(self.name)
@@ -436,7 +449,7 @@ class MetadataField(object):
         if self.formatter is not None:
             try:
                 ans = self.formatter(ans)
-            except:
+            except Exception:
                 return None
         if hasattr(ans, 'strip'):
             ans = ans.strip()
@@ -467,7 +480,8 @@ class TitleSortField(MetadataField):
             matches = obj.title_path(obj.metadata)
             if matches:
                 for match in matches:
-                    ans = match.get('{%s}file-as'%obj.NAMESPACES['opf'], None)
+                    ans = match.get('{%s}file-as' %
+                                    obj.NAMESPACES['opf'], None)
                     if not ans:
                         ans = match.get('file-as', None)
                     if ans:
@@ -488,10 +502,11 @@ class TitleSortField(MetadataField):
                         del match.attrib[attr]
 
 
-def serialize_user_metadata(metadata_elem, all_user_metadata, tail='\n'+(' '*8)):
+def serialize_user_metadata(metadata_elem, all_user_metadata,
+                            tail='\n'+(' '*8)):
     from ebook_converter.utils.config import to_json
-    from ebook_converter.ebooks.metadata.book.json_codec import (object_to_unicode,
-                                                         encode_is_multiple)
+    from ebook_converter.ebooks.metadata.book.json_codec import \
+        object_to_unicode, encode_is_multiple
 
     for name, fm in all_user_metadata.items():
         try:
@@ -499,9 +514,8 @@ def serialize_user_metadata(metadata_elem, all_user_metadata, tail='\n'+(' '*8))
             encode_is_multiple(fm)
             fm = object_to_unicode(fm)
             fm = json.dumps(fm, default=to_json, ensure_ascii=False)
-        except:
+        except Exception:
             prints('Failed to write user metadata:', name)
-            import traceback
             traceback.print_exc()
             continue
         meta = metadata_elem.makeelement('meta')
@@ -514,91 +528,109 @@ def serialize_user_metadata(metadata_elem, all_user_metadata, tail='\n'+(' '*8))
 def dump_dict(cats):
     if not cats:
         cats = {}
-    from ebook_converter.ebooks.metadata.book.json_codec import object_to_unicode
+    from ebook_converter.ebooks.metadata.book.json_codec import \
+        object_to_unicode
     return json.dumps(object_to_unicode(cats), ensure_ascii=False,
-            skipkeys=True)
+                      skipkeys=True)
 
 
 class OPF(object):  # {{{
 
-    MIMETYPE         = 'application/oebps-package+xml'
-    NAMESPACES       = {
-                        None: "http://www.idpf.org/2007/opf",
-                        'dc': "http://purl.org/dc/elements/1.1/",
-                        'opf': "http://www.idpf.org/2007/opf",
-                       }
-    META             = '{%s}meta' % NAMESPACES['opf']
+    MIMETYPE = 'application/oebps-package+xml'
+    NAMESPACES = {None: "http://www.idpf.org/2007/opf",
+                  'dc': "http://purl.org/dc/elements/1.1/",
+                  'opf': "http://www.idpf.org/2007/opf"}
+    META = '{%s}meta' % NAMESPACES['opf']
     xpn = NAMESPACES.copy()
     xpn.pop(None)
     xpn['re'] = 'http://exslt.org/regular-expressions'
     XPath = functools.partial(etree.XPath, namespaces=xpn)
-    CONTENT          = XPath('self::*[re:match(name(), "meta$", "i")]/@content')
-    TEXT             = XPath('string()')
+    CONTENT = XPath('self::*[re:match(name(), "meta$", "i")]/@content')
+    TEXT = XPath('string()')
 
-    metadata_path   = XPath('descendant::*[re:match(name(), "metadata", "i")]')
-    metadata_elem_path = XPath(
-        'descendant::*[re:match(name(), concat($name, "$"), "i") or (re:match(name(), "meta$", "i") '
-        'and re:match(@name, concat("^calibre:", $name, "$"), "i"))]')
-    title_path      = XPath('descendant::*[re:match(name(), "title", "i")]')
-    authors_path    = XPath('descendant::*[re:match(name(), "creator", "i") and (@role="aut" or @opf:role="aut" or (not(@role) and not(@opf:role)))]')
-    bkp_path        = XPath('descendant::*[re:match(name(), "contributor", "i") and (@role="bkp" or @opf:role="bkp")]')
-    tags_path       = XPath('descendant::*[re:match(name(), "subject", "i")]')
-    isbn_path       = XPath('descendant::*[re:match(name(), "identifier", "i") and '
-                            '(re:match(@scheme, "isbn", "i") or re:match(@opf:scheme, "isbn", "i"))]')
-    pubdate_path    = XPath('descendant::*[re:match(name(), "date", "i")]')
-    raster_cover_path = XPath('descendant::*[re:match(name(), "meta", "i") and '
-            're:match(@name, "cover", "i") and @content]')
-    guide_cover_path = XPath('descendant::*[local-name()="guide"]/*[local-name()="reference" and re:match(@type, "cover", "i")]/@href')
-    identifier_path = XPath('descendant::*[re:match(name(), "identifier", "i")]')
-    application_id_path = XPath('descendant::*[re:match(name(), "identifier", "i") and '
-                            '(re:match(@opf:scheme, "calibre|libprs500", "i") or re:match(@scheme, "calibre|libprs500", "i"))]')
-    uuid_id_path    = XPath('descendant::*[re:match(name(), "identifier", "i") and '
-                            '(re:match(@opf:scheme, "uuid", "i") or re:match(@scheme, "uuid", "i"))]')
-    languages_path  = XPath('descendant::*[local-name()="language"]')
+    metadata_path = XPath('descendant::*[re:match(name(), "metadata", "i")]')
+    metadata_elem_path = XPath('descendant::*[re:match(name(), '
+                               'concat($name, "$"), "i") or (re:match(name(), '
+                               '"meta$", "i") and re:match(@name, '
+                               'concat("^calibre:", $name, "$"), "i"))]')
+    title_path = XPath('descendant::*[re:match(name(), "title", "i")]')
+    authors_path = XPath('descendant::*[re:match(name(), "creator", "i") '
+                         'and (@role="aut" or @opf:role="aut" or (not(@role) '
+                         'and not(@opf:role)))]')
+    bkp_path = XPath('descendant::*[re:match(name(), "contributor", "i") and '
+                     '(@role="bkp" or @opf:role="bkp")]')
+    tags_path = XPath('descendant::*[re:match(name(), "subject", "i")]')
+    isbn_path = XPath('descendant::*[re:match(name(), "identifier", "i") and '
+                      '(re:match(@scheme, "isbn", "i") or re:match(@opf:'
+                      'scheme, "isbn", "i"))]')
+    pubdate_path = XPath('descendant::*[re:match(name(), "date", "i")]')
+    raster_cover_path = XPath('descendant::*[re:match(name(), "meta", "i") '
+                              'and re:match(@name, "cover", "i") and '
+                              '@content]')
+    guide_cover_path = XPath('descendant::*[local-name()="guide"]/*[local-'
+                             'name()="reference" and re:match(@type, "cover", '
+                             '"i")]/@href')
+    identifier_path = XPath('descendant::*[re:match(name(), "identifier", '
+                            '"i")]')
+    application_id_path = XPath('descendant::*[re:match(name(), "identifier",'
+                                ' "i") and (re:match(@opf:scheme, '
+                                '"calibre|libprs500", "i") or re:match(@'
+                                'scheme, "calibre|libprs500", "i"))]')
+    uuid_id_path = XPath('descendant::*[re:match(name(), "identifier", "i") '
+                         'and (re:match(@opf:scheme, "uuid", "i") or re:match'
+                         '(@scheme, "uuid", "i"))]')
+    languages_path = XPath('descendant::*[local-name()="language"]')
 
-    manifest_path   = XPath('descendant::*[re:match(name(), "manifest", "i")]/*[re:match(name(), "item", "i")]')
-    manifest_ppath  = XPath('descendant::*[re:match(name(), "manifest", "i")]')
-    spine_path      = XPath('descendant::*[re:match(name(), "spine", "i")]/*[re:match(name(), "itemref", "i")]')
-    guide_path      = XPath('descendant::*[re:match(name(), "guide", "i")]/*[re:match(name(), "reference", "i")]')
+    manifest_path = XPath('descendant::*[re:match(name(), "manifest", "i")]/'
+                          '*[re:match(name(), "item", "i")]')
+    manifest_ppath = XPath('descendant::*[re:match(name(), "manifest", "i")]')
+    spine_path = XPath('descendant::*[re:match(name(), "spine", "i")]/*[re:'
+                       'match(name(), "itemref", "i")]')
+    guide_path = XPath('descendant::*[re:match(name(), "guide", "i")]/*[re:'
+                       'match(name(), "reference", "i")]')
 
-    publisher       = MetadataField('publisher')
-    comments        = MetadataField('description')
-    category        = MetadataField('type')
-    rights          = MetadataField('rights')
-    series          = MetadataField('series', is_dc=False)
+    publisher = MetadataField('publisher')
+    comments = MetadataField('description')
+    category = MetadataField('type')
+    rights = MetadataField('rights')
+    series = MetadataField('series', is_dc=False)
     if tweaks['use_series_auto_increment_tweak_when_importing']:
-        series_index    = MetadataField('series_index', is_dc=False,
-                                        formatter=float, none_is=None)
+        series_index = MetadataField('series_index', is_dc=False,
+                                     formatter=float, none_is=None)
     else:
-        series_index    = MetadataField('series_index', is_dc=False,
-                                        formatter=float, none_is=1)
-    title_sort      = TitleSortField('title_sort', is_dc=False)
-    rating          = MetadataField('rating', is_dc=False, formatter=float)
+        series_index = MetadataField('series_index', is_dc=False,
+                                     formatter=float, none_is=1)
+    title_sort = TitleSortField('title_sort', is_dc=False)
+    rating = MetadataField('rating', is_dc=False, formatter=float)
     publication_type = MetadataField('publication_type', is_dc=False)
-    timestamp       = MetadataField('timestamp', is_dc=False,
-                                    formatter=parse_date, renderer=isoformat)
+    timestamp = MetadataField('timestamp', is_dc=False,
+                              formatter=parse_date, renderer=isoformat)
     user_categories = MetadataField('user_categories', is_dc=False,
                                     formatter=json.loads,
                                     renderer=dump_dict)
     author_link_map = MetadataField('author_link_map', is_dc=False,
-                                formatter=json.loads, renderer=dump_dict)
+                                    formatter=json.loads, renderer=dump_dict)
 
     def __init__(self, stream, basedir=os.getcwd(), unquote_urls=True,
-            populate_spine=True, try_to_guess_cover=True, preparsed_opf=None, read_toc=True):
+                 populate_spine=True, try_to_guess_cover=True,
+                 preparsed_opf=None, read_toc=True):
         self.try_to_guess_cover = try_to_guess_cover
-        self.basedir  = self.base_dir = basedir
+        self.basedir = self.base_dir = basedir
         self.path_to_html_toc = self.html_toc_fragment = None
-        self.root = parse_opf(stream) if preparsed_opf is None else preparsed_opf
+        self.root = preparsed_opf
+        if preparsed_opf is None:
+            self.root = parse_opf(stream)
         try:
             self.package_version = float(self.root.get('version', None))
         except (AttributeError, TypeError, ValueError):
             self.package_version = 0
         self.metadata = self.metadata_path(self.root)
         if not self.metadata:
-            self.metadata = [self.root.makeelement('{http://www.idpf.org/2007/opf}metadata')]
+            self.metadata = [self.root.makeelement('{http://www.idpf.org/2007'
+                                                   '/opf}metadata')]
             self.root.insert(0, self.metadata[0])
             self.metadata[0].tail = '\n'
-        self.metadata      = self.metadata[0]
+        self.metadata = self.metadata[0]
         if unquote_urls:
             self.unquote_urls()
         self.manifest = Manifest()
@@ -623,9 +655,10 @@ class OPF(object):  # {{{
         self._user_metadata_ = {}
         temp = Metadata('x', ['x'])
         from ebook_converter.utils.config import from_json
-        from ebook_converter.ebooks.metadata.book.json_codec import decode_is_multiple
+        from ebook_converter.ebooks.metadata.book.json_codec import \
+            decode_is_multiple
         elems = self.root.xpath('//*[name() = "meta" and starts-with(@name,'
-                '"calibre:user_metadata:") and @content]')
+                                '"calibre:user_metadata:") and @content]')
         for elem in elems:
             name = elem.get('name')
             name = ':'.join(name.split(':')[2:])
@@ -636,9 +669,8 @@ class OPF(object):  # {{{
                 fm = json.loads(fm, object_hook=from_json)
                 decode_is_multiple(fm)
                 temp.set_user_metadata(name, fm)
-            except:
+            except Exception:
                 prints('Failed to read user metadata:', name)
-                import traceback
                 traceback.print_exc()
                 continue
         self._user_metadata_ = temp.get_all_user_metadata(True)
@@ -657,16 +689,16 @@ class OPF(object):  # {{{
 
     def write_user_metadata(self):
         elems = self.root.xpath('//*[name() = "meta" and starts-with(@name,'
-                '"calibre:user_metadata:") and @content]')
+                                '"calibre:user_metadata:") and @content]')
         for elem in elems:
             elem.getparent().remove(elem)
-        serialize_user_metadata(self.metadata,
-                self._user_metadata_)
+        serialize_user_metadata(self.metadata, self._user_metadata_)
 
     def find_toc(self):
         self.toc = None
         try:
-            spine = self.XPath('descendant::*[re:match(name(), "spine", "i")]')(self.root)
+            spine = self.XPath('descendant::*[re:match(name(), "spine", '
+                               '"i")]')(self.root)
             toc = None
             if spine:
                 spine = spine[0]
@@ -682,9 +714,9 @@ class OPF(object):  # {{{
             if toc is None:
                 return
             self.toc = TOC(base_path=self.base_dir)
-            is_ncx = getattr(self, 'manifest', None) is not None and \
-                     self.manifest.type_for_id(toc) is not None and \
-                     'dtbncx' in self.manifest.type_for_id(toc)
+            is_ncx = (getattr(self, 'manifest', None) is not None and
+                      self.manifest.type_for_id(toc) is not None and
+                      'dtbncx' in self.manifest.type_for_id(toc))
             if is_ncx or toc.lower() in ('ncx', 'ncxtoc'):
                 path = self.manifest.path_for_id(toc)
                 if path:
@@ -700,7 +732,7 @@ class OPF(object):  # {{{
                         not os.path.isfile(self.path_to_html_toc):
                     self.path_to_html_toc = None
                 self.toc.read_html_toc(toc)
-        except:
+        except Exception:
             pass
 
     def get_text(self, elem):
@@ -721,11 +753,12 @@ class OPF(object):  # {{{
         c = 1
         while manifest_id in ids:
             c += 1
-            manifest_id = 'id%d'%c
+            manifest_id = 'id%d' % c
         if not media_type:
             media_type = 'application/xhtml+xml'
-        ans = etree.Element('{%s}item'%self.NAMESPACES['opf'],
-                             attrib={'id':manifest_id, 'href':href, 'media-type':media_type})
+        ans = etree.Element('{%s}item' % self.NAMESPACES['opf'],
+                            attrib={'id': manifest_id,
+                                    'href': href, 'media-type': media_type})
         ans.tail = '\n\t\t'
         if append:
             manifest = self.manifest_ppath(self.root)[0]
@@ -735,7 +768,7 @@ class OPF(object):  # {{{
     def replace_manifest_item(self, item, items):
         items = [self.create_manifest_item(*i) for i in items]
         for i, item2 in enumerate(items):
-            item2.set('id', item.get('id')+'.%d'%(i+1))
+            item2.set('id', item.get('id')+'.%d' % (i+1))
         manifest = item.getparent()
         index = manifest.index(item)
         manifest[index:index+1] = items
@@ -761,37 +794,43 @@ class OPF(object):  # {{{
                 return x.get('href', None)
 
     def create_spine_item(self, idref):
-        ans = etree.Element('{%s}itemref'%self.NAMESPACES['opf'], idref=idref)
+        ans = etree.Element('{%s}itemref' % self.NAMESPACES['opf'],
+                            idref=idref)
         ans.tail = '\n\t\t'
         return ans
 
     def replace_spine_items_by_idref(self, idref, new_idrefs):
         items = list(map(self.create_spine_item, new_idrefs))
-        spine = self.XPath('/opf:package/*[re:match(name(), "spine", "i")]')(self.root)[0]
+        spine = self.XPath('/opf:package/*[re:match(name(), "spine", '
+                           '"i")]')(self.root)[0]
         old = [i for i in self.iterspine() if i.get('idref', None) == idref]
         for x in old:
             i = spine.index(x)
             spine[i:i+1] = items
 
     def create_guide_element(self):
-        e = etree.SubElement(self.root, '{%s}guide'%self.NAMESPACES['opf'])
+        e = etree.SubElement(self.root, '{%s}guide' % self.NAMESPACES['opf'])
         e.text = '\n        '
-        e.tail =  '\n'
+        e.tail = '\n'
         return e
 
     def remove_guide(self):
         self.guide = None
-        for g in self.root.xpath('./*[re:match(name(), "guide", "i")]', namespaces={'re':'http://exslt.org/regular-expressions'}):
+        for g in self.root.xpath('./*[re:match(name(), "guide", "i")]',
+                                 namespaces={'re': 'http://exslt.org/regular-'
+                                             'expressions'}):
             self.root.remove(g)
 
     def create_guide_item(self, type, title, href):
-        e = etree.Element('{%s}reference'%self.NAMESPACES['opf'],
-                             type=type, title=title, href=href)
-        e.tail='\n'
+        e = etree.Element('{%s}reference' % self.NAMESPACES['opf'], type=type,
+                          title=title, href=href)
+        e.tail = '\n'
         return e
 
     def add_guide_item(self, type, title, href):
-        g = self.root.xpath('./*[re:match(name(), "guide", "i")]', namespaces={'re':'http://exslt.org/regular-expressions'})[0]
+        g = self.root.xpath('./*[re:match(name(), "guide", "i")]',
+                            namespaces={'re': 'http://exslt.org/regular-'
+                                        'expressions'})[0]
         g.append(self.create_guide_item(type, title, href))
 
     def iterguide(self):
@@ -828,7 +867,10 @@ class OPF(object):  # {{{
                 title.getparent().remove(title)
             titles = ()
         if val:
-            title = titles[0] if titles else self.create_metadata_element('title')
+            if titles:
+                title = titles[0]
+            else:
+                title = self.create_metadata_element('title')
             title.text = re.sub(r'\s+', ' ', str(val))
 
     @property
@@ -847,11 +889,12 @@ class OPF(object):  # {{{
         # for broken implementations that always use the first
         # <dc:creator> element with no attention to the role
         for author in reversed(val):
-            elem = self.metadata.makeelement('{%s}creator'%
-                    self.NAMESPACES['dc'], nsmap=self.NAMESPACES)
+            elem = self.metadata.makeelement('{%s}creator' %
+                                             self.NAMESPACES['dc'],
+                                             nsmap=self.NAMESPACES)
             elem.tail = '\n'
             self.metadata.insert(0, elem)
-            elem.set('{%s}role'%self.NAMESPACES['opf'], 'aut')
+            elem.set('{%s}role' % self.NAMESPACES['opf'], 'aut')
             self.set_text(elem, author.strip())
 
     @property
@@ -859,7 +902,7 @@ class OPF(object):  # {{{
         matches = self.authors_path(self.metadata)
         if matches:
             for match in matches:
-                ans = match.get('{%s}file-as'%self.NAMESPACES['opf'], None)
+                ans = match.get('{%s}file-as' % self.NAMESPACES['opf'], None)
                 if not ans:
                     ans = match.get('file-as', None)
                 if ans:
@@ -872,7 +915,7 @@ class OPF(object):  # {{{
             for key in matches[0].attrib:
                 if key.endswith('file-as'):
                     matches[0].attrib.pop(key)
-            matches[0].set('{%s}file-as'%self.NAMESPACES['opf'], str(val))
+            matches[0].set('{%s}file-as' % self.NAMESPACES['opf'], str(val))
 
     @property
     def tags(self):
@@ -897,8 +940,9 @@ class OPF(object):  # {{{
         for match in self.pubdate_path(self.metadata):
             try:
                 val = parse_date(etree.tostring(match, encoding='unicode',
-                    method='text', with_tail=False).strip())
-            except:
+                                                method='text',
+                                                with_tail=False).strip())
+            except Exception:
                 continue
             if ans is None or val < ans:
                 ans = val
@@ -910,8 +954,9 @@ class OPF(object):  # {{{
         for match in self.pubdate_path(self.metadata):
             try:
                 cval = parse_date(etree.tostring(match, encoding='unicode',
-                    method='text', with_tail=False).strip())
-            except:
+                                                 method='text',
+                                                 with_tail=False).strip())
+            except Exception:
                 match.getparent().remove(match)
             else:
                 if not val:
@@ -953,7 +998,7 @@ class OPF(object):  # {{{
                     x.getparent().remove(x)
             return
         if not matches:
-            attrib = {'{%s}scheme'%self.NAMESPACES['opf']: 'ISBN'}
+            attrib = {'{%s}scheme' % self.NAMESPACES['opf']: 'ISBN'}
             matches = [self.create_metadata_element('identifier',
                                                     attrib=attrib)]
         self.set_text(matches[0], str(val))
@@ -967,17 +1012,19 @@ class OPF(object):  # {{{
             for attr, val in x.attrib.items():
                 if attr.endswith('scheme'):
                     typ = val.lower()
-                    val = etree.tostring(x, with_tail=False, encoding='unicode',
-                            method='text').strip()
+                    val = etree.tostring(x, with_tail=False,
+                                         encoding='unicode',
+                                         method='text').strip()
                     if val and typ not in ('calibre', 'uuid'):
-                        if typ == 'isbn' and val.lower().startswith('urn:isbn:'):
+                        if (typ == 'isbn' and
+                                val.lower().startswith('urn:isbn:')):
                             val = val[len('urn:isbn:'):]
                         identifiers[typ] = val
                     found_scheme = True
                     break
             if not found_scheme:
                 val = etree.tostring(x, with_tail=False, encoding='unicode',
-                            method='text').strip()
+                                     method='text').strip()
                 if val.lower().startswith('urn:isbn:'):
                     val = check_isbn(val.split(':')[-1])
                     if val is not None:
@@ -997,7 +1044,8 @@ class OPF(object):  # {{{
                     self.metadata):
             xid = x.get('id', None)
             is_package_identifier = uuid_id is not None and uuid_id == xid
-            typ = {val.lower() for attr, val in x.attrib.items() if attr.endswith('scheme')}
+            typ = {val.lower() for attr, val in x.attrib.items()
+                   if attr.endswith('scheme')}
             if is_package_identifier:
                 typ = tuple(typ)
                 if typ and typ[0] in identifiers:
@@ -1007,7 +1055,7 @@ class OPF(object):  # {{{
                 x.getparent().remove(x)
 
         for typ, val in identifiers.items():
-            attrib = {'{%s}scheme'%self.NAMESPACES['opf']: typ.upper()}
+            attrib = {'{%s}scheme' % self.NAMESPACES['opf']: typ.upper()}
             self.set_text(self.create_metadata_element(
                 'identifier', attrib=attrib), str(val))
 
@@ -1028,7 +1076,7 @@ class OPF(object):  # {{{
             if attr.endswith('unique-identifier'):
                 uuid_id = self.root.attrib[attr]
                 break
-        attrib = {'{%s}scheme'%self.NAMESPACES['opf']: 'calibre'}
+        attrib = {'{%s}scheme' % self.NAMESPACES['opf']: 'calibre'}
         if uuid_id and uuid_id in removed_ids:
             attrib['id'] = uuid_id
         self.set_text(self.create_metadata_element(
@@ -1043,7 +1091,7 @@ class OPF(object):  # {{{
     def uuid(self, val):
         matches = self.uuid_id_path(self.metadata)
         if not matches:
-            attrib = {'{%s}scheme'%self.NAMESPACES['opf']: 'uuid'}
+            attrib = {'{%s}scheme' % self.NAMESPACES['opf']: 'uuid'}
             matches = [self.create_metadata_element('identifier',
                                                     attrib=attrib)]
         self.set_text(matches[0], str(val))
@@ -1064,9 +1112,9 @@ class OPF(object):  # {{{
         for match in self.languages_path(self.metadata):
             t = self.get_text(match)
             if t and t.strip():
-                l = canonicalize_lang(t.strip())
-                if l:
-                    ans.append(l)
+                cl = canonicalize_lang(t.strip())
+                if cl:
+                    ans.append(cl)
         return ans
 
     @languages.setter
@@ -1076,8 +1124,8 @@ class OPF(object):  # {{{
             x.getparent().remove(x)
 
         for lang in val:
-            l = self.create_metadata_element('language')
-            self.set_text(l, str(lang))
+            cl = self.create_metadata_element('language')
+            self.set_text(cl, str(lang))
 
     @property
     def raw_languages(self):
@@ -1096,7 +1144,7 @@ class OPF(object):  # {{{
         matches = self.bkp_path(self.metadata)
         if not matches:
             matches = [self.create_metadata_element('contributor')]
-            matches[0].set('{%s}role'%self.NAMESPACES['opf'], 'bkp')
+            matches[0].set('{%s}role' % self.NAMESPACES['opf'], 'bkp')
         self.set_text(matches[0], str(val))
 
     def identifier_iter(self):
@@ -1111,7 +1159,8 @@ class OPF(object):  # {{{
                 uuid_elem = self.root.attrib[attr]
                 break
         if uuid_elem:
-            matches = self.root.xpath('//*[@id=%s]'%escape_xpath_attr(uuid_elem))
+            matches = self.root.xpath('//*[@id=%s]' %
+                                      escape_xpath_attr(uuid_elem))
             if matches:
                 for m in matches:
                     raw = m.text
@@ -1126,15 +1175,18 @@ class OPF(object):  # {{{
 
     @property
     def page_progression_direction(self):
-        spine = self.XPath('descendant::*[re:match(name(), "spine", "i")][1]')(self.root)
+        spine = self.XPath('descendant::*[re:match(name(), "spine", '
+                           '"i")][1]')(self.root)
         if spine:
             for k, v in spine[0].attrib.items():
-                if k == 'page-progression-direction' or k.endswith('}page-progression-direction'):
+                if (k == 'page-progression-direction' or
+                        k.endswith('}page-progression-direction')):
                     return v
 
     @property
     def primary_writing_mode(self):
-        for m in self.XPath('//*[local-name()="meta" and @name="primary-writing-mode" and @content]')(self.root):
+        for m in self.XPath('//*[local-name()="meta" and @name="primary-'
+                            'writing-mode" and @content]')(self.root):
             return m.get('content')
 
     def guess_cover(self):
@@ -1153,8 +1205,11 @@ class OPF(object):  # {{{
                 if item.text:
                     prefix = item.text.replace('-', '')
                     for suffix in ['.jpg', '.jpeg', '.gif', '.png', '.bmp']:
-                        cpath = os.access(os.path.join(self.base_dir, prefix+suffix), os.R_OK)
-                        if os.access(os.path.join(self.base_dir, prefix+suffix), os.R_OK):
+                        cpath = os.access(os.path.join(self.base_dir,
+                                                       prefix + suffix),
+                                          os.R_OK)
+                        if os.access(os.path.join(self.base_dir,
+                                                  prefix+suffix), os.R_OK):
                             return cpath
 
     @property
@@ -1188,13 +1243,16 @@ class OPF(object):  # {{{
     def guide_raster_cover(self):
         covers = self.guide_cover_path(self.root)
         if covers:
-            mt_map = {i.get('href'):i for i in self.itermanifest()}
+            mt_map = {i.get('href'): i for i in self.itermanifest()}
             for href in covers:
                 if href:
                     i = mt_map.get(href)
                     if i is not None:
                         iid, mt = i.get('id'), i.get('media-type')
-                        if iid and mt and mt.lower() in {'image/png', 'image/jpeg', 'image/jpg', 'image/gif'}:
+                        if iid and mt and mt.lower() in {'image/png',
+                                                         'image/jpeg',
+                                                         'image/jpg',
+                                                         'image/gif'}:
                             return i
 
     @property
@@ -1214,14 +1272,15 @@ class OPF(object):  # {{{
     @property
     def cover(self):
         if self.guide is not None:
-            for t in ('cover', 'other.ms-coverimage-standard', 'other.ms-coverimage'):
+            for t in ('cover', 'other.ms-coverimage-standard',
+                      'other.ms-coverimage'):
                 for item in self.guide:
                     if item.type and item.type.lower() == t:
                         return item.path
         try:
             if self.try_to_guess_cover:
                 return self.guess_cover()
-        except:
+        except Exception:
             pass
 
     @cover.setter
@@ -1237,13 +1296,16 @@ class OPF(object):  # {{{
             self.guide = Guide()
             self.guide.set_cover(path)
             etree.SubElement(g, 'opf:reference', nsmap=self.NAMESPACES,
-                                attrib={'type':'cover', 'href':self.guide[-1].href()})
+                             attrib={'type': 'cover',
+                                     'href': self.guide[-1].href()})
         id = self.manifest.id_for_path(self.cover)
         if id is None:
-            for t in ('cover', 'other.ms-coverimage-standard', 'other.ms-coverimage'):
+            for t in ('cover', 'other.ms-coverimage-standard',
+                      'other.ms-coverimage'):
                 for item in self.guide:
                     if item.type.lower() == t:
-                        self.create_manifest_item(item.href(), guess_type(path)[0])
+                        self.create_manifest_item(item.href(),
+                                                  guess_type(path)[0])
 
     def get_metadata_element(self, name):
         matches = self.metadata_elem_path(self.metadata, name=name)
@@ -1278,9 +1340,11 @@ class OPF(object):  # {{{
         # We swap attributes instead of elements, as that avoids namespace
         # re-declarations
         smap = {}
-        for child in self.metadata.xpath('./*[@name="calibre:series" or @name="calibre:series_index"]'):
+        for child in self.metadata.xpath('./*[@name="calibre:series" or @name'
+                                         '="calibre:series_index"]'):
             smap[child.get('name')] = (child, self.metadata.index(child))
-        if len(smap) == 2 and smap['calibre:series'][1] > smap['calibre:series_index'][1]:
+        if (len(smap) == 2 and
+                smap['calibre:series'][1] > smap['calibre:series_index'][1]):
             s, si = smap['calibre:series'][0], smap['calibre:series_index'][0]
 
             def swap(attr):
@@ -1293,7 +1357,8 @@ class OPF(object):  # {{{
             _pretty_print(self.root)
         raw = etree.tostring(self.root, encoding=encoding, pretty_print=True)
         if not raw.lstrip().startswith(b'<?xml '):
-            raw = ('<?xml version="1.0"  encoding="%s"?>\n'%encoding.upper()).encode('ascii') + raw
+            raw = ('<?xml version="1.0"  encoding="%s"?>\n' %
+                   encoding.upper()).encode('ascii') + raw
         return raw
 
     def smart_update(self, mi, replace_metadata=False, apply_null=False):
@@ -1302,9 +1367,12 @@ class OPF(object):  # {{{
                      'isbn', 'tags', 'category', 'comments', 'book_producer',
                      'pubdate', 'user_categories', 'author_link_map'):
             val = getattr(mi, attr, None)
-            is_null = val is None or val in ((), [], (None, None), {}) or (attr == 'rating' and val < 0.1)
+            is_null = val is None or val in ((), [], (None, None),
+                                             {}) or (attr == 'rating' and
+                                                     val < 0.1)
             if is_null:
-                if apply_null and attr in {'series', 'tags', 'isbn', 'comments', 'publisher', 'rating'}:
+                if apply_null and attr in {'series', 'tags', 'isbn',
+                                           'comments', 'publisher', 'rating'}:
                     setattr(self, attr, ([] if attr == 'tags' else None))
             else:
                 setattr(self, attr, val)
@@ -1315,7 +1383,9 @@ class OPF(object):  # {{{
             self.languages = langs or []
         temp = self.to_book_metadata()
         temp.smart_update(mi, replace_metadata=replace_metadata)
-        if not replace_metadata and callable(getattr(temp, 'custom_field_keys', None)):
+        if not replace_metadata and callable(getattr(temp,
+                                                     'custom_field_keys',
+                                                     None)):
             # We have to replace non-null fields regardless of the value of
             # replace_metadata to match the behavior of the builtin fields
             # above.
@@ -1346,12 +1416,12 @@ class OPF(object):  # {{{
 class OPFCreator(Metadata):
 
     def __init__(self, base_path, other):
-        '''
+        """
         Initialize.
-        @param base_path: An absolute path to the directory in which this OPF file
-        will eventually be. This is used by the L{create_manifest} method
+        @param base_path: An absolute path to the directory in which this OPF
+        file will eventually be. This is used by the L{create_manifest} method
         to convert paths to files into relative paths.
-        '''
+        """
         Metadata.__init__(self, title='', other=other)
         self.base_path = os.path.abspath(base_path)
         self.page_progression_direction = None
@@ -1368,19 +1438,21 @@ class OPFCreator(Metadata):
             self.guide.set_cover(self.cover)
 
     def create_manifest(self, entries):
-        '''
+        """
         Create <manifest>
 
-        `entries`: List of (path, mime-type) If mime-type is None it is autodetected
-        '''
+        `entries`: List of (path, mime-type) If mime-type is None it is
+        autodetected
+        """
         entries = list(map(lambda x: x if os.path.isabs(x[0]) else
-                      (os.path.abspath(os.path.join(self.base_path, x[0])), x[1]),
-                      entries))
+                           (os.path.abspath(os.path.join(self.base_path,
+                                                         x[0])), x[1]),
+                           entries))
         self.manifest = Manifest.from_paths(entries)
         self.manifest.set_basedir(self.base_path)
 
     def create_manifest_from_files_in(self, files_and_dirs,
-            exclude=lambda x:False):
+                                      exclude=lambda x: False):
         entries = []
 
         def dodir(dir):
@@ -1406,7 +1478,8 @@ class OPFCreator(Metadata):
         `entries`: List of paths
         '''
         entries = list(map(lambda x: x if os.path.isabs(x) else
-                      os.path.abspath(os.path.join(self.base_path, x)), entries))
+                           os.path.abspath(os.path.join(self.base_path, x)),
+                           entries))
         self.spine = Spine.from_paths(entries, self.manifest)
 
     def set_toc(self, toc):
@@ -1431,11 +1504,13 @@ class OPFCreator(Metadata):
             self.manifest.set_basedir(self.base_path)
             if ncx_manifest_entry is not None and toc is not None:
                 if not os.path.isabs(ncx_manifest_entry):
-                    ncx_manifest_entry = os.path.join(self.base_path, ncx_manifest_entry)
+                    ncx_manifest_entry = os.path.join(self.base_path,
+                                                      ncx_manifest_entry)
                 remove = [i for i in self.manifest if i.id == 'ncx']
                 for item in remove:
                     self.manifest.remove(item)
-                self.manifest.append(ManifestItem(ncx_manifest_entry, self.base_path))
+                self.manifest.append(ManifestItem(ncx_manifest_entry,
+                                                  self.base_path))
                 self.manifest[-1].id = 'ncx'
                 self.manifest[-1].mime_type = 'application/x-dtbncx+xml'
         if self.guide is None:
@@ -1448,13 +1523,13 @@ class OPFCreator(Metadata):
         self.guide.set_basedir(self.base_path)
 
         # Actual rendering
-        from lxml.builder import ElementMaker
-        from ebook_converter.ebooks.oeb.base import OPF2_NS, DC11_NS, CALIBRE_NS
-        DNS = OPF2_NS+'___xx___'
-        E = ElementMaker(namespace=DNS, nsmap={None:DNS})
+        DNS = const.OPF2_NS + '___xx___'
+        E = ElementMaker(namespace=DNS, nsmap={None: DNS})
         M = ElementMaker(namespace=DNS,
-                nsmap={'dc':DC11_NS, 'calibre':CALIBRE_NS, 'opf':OPF2_NS})
-        DC = ElementMaker(namespace=DC11_NS)
+                         nsmap={'dc': const.DC11_NS,
+                                'calibre': const.CALIBRE_NS,
+                                'opf': const.OPF2_NS})
+        DC = ElementMaker(namespace=const.DC11_NS)
 
         def DC_ELEM(tag, text, dc_attrs={}, opf_attrs={}):
             if text:
@@ -1462,7 +1537,7 @@ class OPFCreator(Metadata):
             else:
                 elem = getattr(DC, tag)(**dc_attrs)
             for k, v in opf_attrs.items():
-                elem.set('{%s}%s'%(OPF2_NS, k), v)
+                elem.set('{%s}%s' % (const.OPF2_NS, k), v)
             return elem
 
         def CAL_ELEM(name, content):
@@ -1472,18 +1547,18 @@ class OPFCreator(Metadata):
         a = metadata.append
         role = {}
         a(DC_ELEM('title', self.title if self.title else 'Unknown',
-            opf_attrs=role))
+                  opf_attrs=role))
         for i, author in enumerate(self.authors):
-            fa = {'role':'aut'}
+            fa = {'role': 'aut'}
             if i == 0 and self.author_sort:
                 fa['file-as'] = self.author_sort
             a(DC_ELEM('creator', author, opf_attrs=fa))
-        a(DC_ELEM('contributor', '%s (%s) [%s]'%(__appname__, __version__,
-            'https://calibre-ebook.com'), opf_attrs={'role':'bkp',
-                'file-as':__appname__}))
+        a(DC_ELEM('contributor', '%s (%s) [%s]' %
+                  (__appname__, __version__, 'https://calibre-ebook.com'),
+                  opf_attrs={'role': 'bkp', 'file-as': __appname__}))
         a(DC_ELEM('identifier', str(self.application_id),
-            opf_attrs={'scheme':__appname__},
-            dc_attrs={'id':__appname__+'_id'}))
+                  opf_attrs={'scheme': __appname__},
+                  dc_attrs={'id': __appname__+'_id'}))
         if getattr(self, 'pubdate', None) is not None:
             a(DC_ELEM('date', self.pubdate.isoformat()))
         langs = self.languages
@@ -1496,7 +1571,7 @@ class OPFCreator(Metadata):
         if self.publisher:
             a(DC_ELEM('publisher', self.publisher))
         for key, val in self.get_identifiers().items():
-            a(DC_ELEM('identifier', val, opf_attrs={'scheme':key.upper()}))
+            a(DC_ELEM('identifier', val, opf_attrs={'scheme': key.upper()}))
         if self.rights:
             a(DC_ELEM('rights', self.rights))
         if self.tags:
@@ -1515,11 +1590,13 @@ class OPFCreator(Metadata):
         if self.publication_type is not None:
             a(CAL_ELEM('calibre:publication_type', self.publication_type))
         if self.user_categories:
-            from ebook_converter.ebooks.metadata.book.json_codec import object_to_unicode
+            from ebook_converter.ebooks.metadata.book.json_codec import \
+                object_to_unicode
             a(CAL_ELEM('calibre:user_categories',
                        json.dumps(object_to_unicode(self.user_categories))))
         if self.primary_writing_mode:
-            a(M.meta(name='primary-writing-mode', content=self.primary_writing_mode))
+            a(M.meta(name='primary-writing-mode',
+                     content=self.primary_writing_mode))
         manifest = E.manifest()
         if self.manifest is not None:
             for ref in self.manifest:
@@ -1533,7 +1610,8 @@ class OPFCreator(Metadata):
         if self.toc is not None:
             spine.set('toc', 'ncx')
         if self.page_progression_direction is not None:
-            spine.set('page-progression-direction', self.page_progression_direction)
+            spine.set('page-progression-direction',
+                      self.page_progression_direction)
         if self.spine is not None:
             for ref in self.spine:
                 if ref.id is not None:
@@ -1562,8 +1640,8 @@ class OPFCreator(Metadata):
         root.set('unique-identifier', __appname__+'_id')
         root.set('version', '2.0')
         raw = etree.tostring(root, pretty_print=True, xml_declaration=True,
-                encoding=encoding)
-        raw = raw.replace(DNS.encode('utf-8'), OPF2_NS.encode('utf-8'))
+                             encoding=encoding)
+        raw = raw.replace(DNS.encode('utf-8'), const.OPF2_NS.encode('utf-8'))
         opf_stream.write(raw)
         opf_stream.flush()
         if toc is not None and ncx_stream is not None:
@@ -1572,9 +1650,6 @@ class OPFCreator(Metadata):
 
 
 def metadata_to_opf(mi, as_string=True, default_lang=None):
-    from lxml import etree
-    import textwrap
-    from ebook_converter.ebooks.oeb.base import OPF, DC
 
     if not mi.application_id:
         mi.application_id = str(uuid.uuid4())
@@ -1583,15 +1658,15 @@ def metadata_to_opf(mi, as_string=True, default_lang=None):
         mi.uuid = str(uuid.uuid4())
 
     if not mi.book_producer:
-        mi.book_producer = __appname__ + ' (%s) '%__version__ + \
-            '[https://calibre-ebook.com]'
+        mi.book_producer = (__appname__ + ' (%s) ' %
+                            __version__ + '[https://calibre-ebook.com]')
 
     if not mi.languages:
         lang = (get_lang().replace('_', '-').partition('-')[0] if default_lang
                 is None else default_lang)
         mi.languages = [lang]
 
-    root = safe_xml_fromstring(textwrap.dedent(
+    root = etree.fromstring(textwrap.dedent(
     '''
     <package xmlns="http://www.idpf.org/2007/opf" unique-identifier="uuid_id" version="2.0">
         <metadata xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:opf="http://www.idpf.org/2007/opf">
@@ -1600,20 +1675,20 @@ def metadata_to_opf(mi, as_string=True, default_lang=None):
             </metadata>
         <guide/>
     </package>
-    '''%dict(a=__appname__, id=mi.application_id, uuid=mi.uuid)))
+    ''' % dict(a=__appname__, id=mi.application_id, uuid=mi.uuid)))
     metadata = root[0]
     guide = root[1]
     metadata[0].tail = '\n'+(' '*8)
 
     def factory(tag, text=None, sort=None, role=None, scheme=None, name=None,
-            content=None):
+                content=None):
         attrib = {}
         if sort:
-            attrib[OPF('file-as')] = sort
+            attrib[base.tag('opf', 'file-as')] = sort
         if role:
-            attrib[OPF('role')] = role
+            attrib[base.tag('opf', 'role')] = role
         if scheme:
-            attrib[OPF('scheme')] = scheme
+            attrib[base.tag('opf', 'scheme')] = scheme
         if name:
             attrib['name'] = name
         if content:
@@ -1621,7 +1696,9 @@ def metadata_to_opf(mi, as_string=True, default_lang=None):
         try:
             elem = metadata.makeelement(tag, attrib=attrib)
         except ValueError:
-            elem = metadata.makeelement(tag, attrib={k:clean_xml_chars(v) for k, v in attrib.items()})
+            elem = metadata.makeelement(tag, attrib={k: clean_xml_chars(v)
+                                                     for k, v in
+                                                     attrib.items()})
         elem.tail = '\n'+(' '*8)
         if text:
             try:
@@ -1630,29 +1707,29 @@ def metadata_to_opf(mi, as_string=True, default_lang=None):
                 elem.text = clean_ascii_chars(text.strip())
         metadata.append(elem)
 
-    factory(DC('title'), mi.title)
+    factory(base.tag('dc', 'title'), mi.title)
     for au in mi.authors:
-        factory(DC('creator'), au, mi.author_sort, 'aut')
-    factory(DC('contributor'), mi.book_producer, __appname__, 'bkp')
+        factory(base.tag('dc', 'creator'), au, mi.author_sort, 'aut')
+    factory(base.tag('dc', 'contributor'), mi.book_producer, __appname__, 'bkp')
     if hasattr(mi.pubdate, 'isoformat'):
-        factory(DC('date'), isoformat(mi.pubdate))
+        factory(base.tag('dc', 'date'), isoformat(mi.pubdate))
     if hasattr(mi, 'category') and mi.category:
-        factory(DC('type'), mi.category)
+        factory(base.tag('dc', 'type'), mi.category)
     if mi.comments:
-        factory(DC('description'), clean_ascii_chars(mi.comments))
+        factory(base.tag('dc', 'description'), clean_ascii_chars(mi.comments))
     if mi.publisher:
-        factory(DC('publisher'), mi.publisher)
+        factory(base.tag('dc', 'publisher'), mi.publisher)
     for key, val in mi.get_identifiers().items():
-        factory(DC('identifier'), val, scheme=key.upper())
+        factory(base.tag('dc', 'identifier'), val, scheme=key.upper())
     if mi.rights:
-        factory(DC('rights'), mi.rights)
+        factory(base.tag('dc', 'rights'), mi.rights)
     for lang in mi.languages:
         if not lang or lang.lower() == 'und':
             continue
-        factory(DC('language'), lang)
+        factory(base.tag('dc', 'language'), lang)
     if mi.tags:
         for tag in mi.tags:
-            factory(DC('subject'), tag)
+            factory(base.tag('dc', 'subject'), tag)
     meta = lambda n, c: factory('meta', name='calibre:'+n, content=c)
     if getattr(mi, 'author_link_map', None) is not None:
         meta('author_link_map', dump_dict(mi.author_link_map))
@@ -1673,21 +1750,22 @@ def metadata_to_opf(mi, as_string=True, default_lang=None):
 
     serialize_user_metadata(metadata, mi.get_all_user_metadata(False))
 
-    metadata[-1].tail = '\n' +(' '*4)
+    metadata[-1].tail = '\n' + (' '*4)
 
     if mi.cover:
         if not isinstance(mi.cover, str):
             mi.cover = mi.cover.decode(filesystem_encoding)
         guide.text = '\n'+(' '*8)
-        r = guide.makeelement(OPF('reference'),
-                attrib={'type': 'cover', 'title': 'Cover', 'href': mi.cover})
-        r.tail = '\n' +(' '*4)
+        r = guide.makeelement(base.tag('opf', 'reference'),
+                              attrib={'type': 'cover', 'title': 'Cover',
+                                      'href': mi.cover})
+        r.tail = '\n' + (' '*4)
         guide.append(r)
     if pretty_print_opf:
         _pretty_print(root)
 
     return etree.tostring(root, pretty_print=True, encoding='utf-8',
-            xml_declaration=True) if as_string else root
+                          xml_declaration=True) if as_string else root
 
 
 def test_m2o():
@@ -1699,7 +1777,7 @@ def test_m2o():
     mi.language = 'en'
     mi.comments = 'what a fun book\n\n'
     mi.publisher = 'publisher'
-    mi.set_identifiers({'isbn':'booo', 'dummy':'dummy'})
+    mi.set_identifiers({'isbn': 'booo', 'dummy': 'dummy'})
     mi.tags = ['a', 'b']
     mi.series = 's"c\'l&<>'
     mi.series_index = 3.34
@@ -1711,15 +1789,15 @@ def test_m2o():
     opf = metadata_to_opf(mi)
     print(opf)
     newmi = MetaInformation(OPF(io.BytesIO(opf)))
-    for attr in ('author_sort', 'title_sort', 'comments',
-                    'publisher', 'series', 'series_index', 'rating',
-                    'isbn', 'tags', 'cover_data', 'application_id',
-                    'language', 'cover',
-                    'book_producer', 'timestamp',
-                    'pubdate', 'rights', 'publication_type'):
+    for attr in ('author_sort', 'title_sort', 'comments', 'publisher',
+                 'series', 'series_index', 'rating', 'isbn', 'tags',
+                 'cover_data', 'application_id', 'language', 'cover',
+                 'book_producer', 'timestamp', 'pubdate', 'rights',
+                 'publication_type'):
         o, n = getattr(mi, attr), getattr(newmi, attr)
         if o != n and o.strip() != n.strip():
-            print('FAILED:', attr, getattr(mi, attr), '!=', getattr(newmi, attr))
+            print('FAILED:', attr, getattr(mi, attr), '!=',
+                  getattr(newmi, attr))
     if mi.get_identifiers() != newmi.get_identifiers():
         print('FAILED:', 'identifiers', mi.get_identifiers(), end=' ')
         print('!=', newmi.get_identifiers())
@@ -1766,8 +1844,8 @@ b'''\
         self.assertEqual(opf.rating, 4)
         self.assertEqual(opf.publication_type, 'test')
         self.assertEqual(list(opf.itermanifest())[0].get('href'), 'a ~ b')
-        self.assertEqual(opf.get_identifiers(), {'isbn':'123456789',
-            'dummy':'dummy'})
+        self.assertEqual(opf.get_identifiers(), {'isbn': '123456789',
+                                                 'dummy': 'dummy'})
 
     def testWriting(self):
         for test in [('title', 'New & Title'), ('authors', ['One', 'Two']),
@@ -1802,14 +1880,12 @@ def test():
 
 def test_user_metadata():
     mi = Metadata('Test title', ['test author1', 'test author2'])
-    um = {
-        '#myseries': {'#value#': u'test series\xe4', 'datatype':'text',
-            'is_multiple': None, 'name': u'My Series'},
-        '#myseries_index': {'#value#': 2.45, 'datatype': 'float',
-            'is_multiple': None},
-        '#mytags': {'#value#':['t1','t2','t3'], 'datatype':'text',
-            'is_multiple': '|', 'name': u'My Tags'}
-        }
+    um = {'#myseries': {'#value#': u'test series\xe4', 'datatype': 'text',
+                        'is_multiple': None, 'name': u'My Series'},
+          '#myseries_index': {'#value#': 2.45, 'datatype': 'float',
+                              'is_multiple': None},
+          '#mytags': {'#value#': ['t1', 't2', 't3'], 'datatype': 'text',
+                      'is_multiple': '|', 'name': u'My Tags'}}
     mi.set_all_user_metadata(um)
     raw = metadata_to_opf(mi)
     opfc = OPFCreator(os.getcwd(), other=mi)
diff --git a/ebook_converter/ebooks/metadata/opf3.py b/ebook_converter/ebooks/metadata/opf3.py
index f12a79d..dd146a6 100644
--- a/ebook_converter/ebooks/metadata/opf3.py
+++ b/ebook_converter/ebooks/metadata/opf3.py
@@ -5,6 +5,7 @@ from functools import wraps
 
 from lxml import etree
 
+from ebook_converter import constants as const
 from ebook_converter import prints
 from ebook_converter.ebooks.metadata import authors_to_string, check_isbn, string_to_authors
 from ebook_converter.ebooks.metadata.book.base import Metadata
@@ -15,7 +16,6 @@ from ebook_converter.ebooks.metadata.utils import (
     create_manifest_item, ensure_unique, normalize_languages, parse_opf,
     pretty_print_opf
 )
-from ebook_converter.ebooks.oeb.base import DC, OPF, OPF2_NSMAP
 from ebook_converter.utils.config import from_json, to_json
 from ebook_converter.utils.date import (
     fix_only_date, is_date_undefined, isoformat, parse_date as parse_date_, utcnow,
@@ -46,7 +46,7 @@ def XPath(x):
     try:
         return _xpath_cache[x]
     except KeyError:
-        _xpath_cache[x] = ans = etree.XPath(x, namespaces=OPF2_NSMAP)
+        _xpath_cache[x] = ans = etree.XPath(x, namespaces=const.OPF2_NSMAP)
         return ans
 
 
@@ -213,7 +213,7 @@ def set_refines(elem, existing_refines, *new_refines):
     remove_refines(elem, existing_refines)
     for ref in reversed(new_refines):
         prop, val, scheme = ref
-        r = elem.makeelement(OPF('meta'))
+        r = elem.makeelement(const.OPF_META)
         r.set('refines', '#' + eid), r.set('property', prop)
         r.text = val.strip()
         if scheme:
@@ -249,7 +249,7 @@ def parse_identifier(ident, val, refines):
     # Try the OPF 2 style opf:scheme attribute, which will be present, for
     # example, in EPUB 3 files that have had their metadata set by an
     # application that only understands EPUB 2.
-    scheme = ident.get(OPF('scheme'))
+    scheme = ident.get(const.OPF_SCHEME)
     if scheme and not lval.startswith('urn:'):
         return finalize(scheme, val)
 
@@ -294,7 +294,7 @@ def set_identifiers(root, prefixes, refines, new_identifiers, force_identifiers=
             continue
     metadata = XPath('./opf:metadata')(root)[0]
     for scheme, val in new_identifiers.items():
-        ident = metadata.makeelement(DC('identifier'))
+        ident = metadata.makeelement(const.DC_IDENT)
         ident.text = '%s:%s' % (scheme, val)
         if package_identifier is None:
             metadata.append(ident)
@@ -312,11 +312,11 @@ def identifier_writer(name):
             if is_package_id:
                 package_identifier = ident
             val = (ident.text or '').strip()
-            if (val.startswith(name + ':') or ident.get(OPF('scheme')) == name) and not is_package_id:
+            if (val.startswith(name + ':') or ident.get(const.OPF_SCHEME) == name) and not is_package_id:
                 remove_element(ident, refines)
         metadata = XPath('./opf:metadata')(root)[0]
         if ival:
-            ident = metadata.makeelement(DC('identifier'))
+            ident = metadata.makeelement(const.DC_IDENT)
             ident.text = '%s:%s' % (name, ival)
             if package_identifier is None:
                 metadata.append(ident)
@@ -376,7 +376,7 @@ def set_title(root, prefixes, refines, title, title_sort=None):
     main_title = find_main_title(root, refines, remove_blanks=True)
     if main_title is None:
         m = XPath('./opf:metadata')(root)[0]
-        main_title = m.makeelement(DC('title'))
+        main_title = m.makeelement(const.DC_TITLE)
         m.insert(0, main_title)
     main_title.text = title or None
     ts = [refdef('file-as', title_sort)] if title_sort else ()
@@ -411,7 +411,7 @@ def set_languages(root, prefixes, refines, languages):
         languages = ['und']
     metadata = XPath('./opf:metadata')(root)[0]
     for lang in uniq(languages):
-        l = metadata.makeelement(DC('language'))
+        l = metadata.makeelement(const.DC_LANG)
         l.text = lang
         metadata.append(l)
 # }}}
@@ -440,7 +440,7 @@ def read_authors(root, prefixes, refines):
         if file_as:
             aus = file_as[0][-1]
         else:
-            aus = item.get(OPF('file-as')) or None
+            aus = item.get(const.OPF_FILE_AS) or None
         return Author(normalize_whitespace(val), normalize_whitespace(aus))
 
     for item in XPath('./opf:metadata/dc:creator')(root):
@@ -448,7 +448,7 @@ def read_authors(root, prefixes, refines):
         if val:
             props = properties_for_id_with_scheme(item.get('id'), prefixes, refines)
             role = props.get('role')
-            opf_role = item.get(OPF('role'))
+            opf_role = item.get(const.OPF_ROLE)
             if role:
                 if is_relators_role(props, 'aut'):
                     roled_authors.append(author(item, props, val))
@@ -465,22 +465,22 @@ def set_authors(root, prefixes, refines, authors):
     ensure_prefix(root, prefixes, 'marc')
     for item in XPath('./opf:metadata/dc:creator')(root):
         props = properties_for_id_with_scheme(item.get('id'), prefixes, refines)
-        opf_role = item.get(OPF('role'))
+        opf_role = item.get(const.OPF_ROLE)
         if (opf_role and opf_role.lower() != 'aut') or (props.get('role') and not is_relators_role(props, 'aut')):
             continue
         remove_element(item, refines)
     metadata = XPath('./opf:metadata')(root)[0]
     for author in authors:
         if author.name:
-            a = metadata.makeelement(DC('creator'))
+            a = metadata.makeelement(const.DC_CREATOR)
             aid = ensure_id(a)
             a.text = author.name
             metadata.append(a)
-            m = metadata.makeelement(OPF('meta'), attrib={'refines':'#'+aid, 'property':'role', 'scheme':'marc:relators'})
+            m = metadata.makeelement(const.OPF_META, attrib={'refines':'#'+aid, 'property':'role', 'scheme':'marc:relators'})
             m.text = 'aut'
             metadata.append(m)
             if author.sort:
-                m = metadata.makeelement(OPF('meta'), attrib={'refines':'#'+aid, 'property':'file-as'})
+                m = metadata.makeelement(const.OPF_META, attrib={'refines':'#'+aid, 'property':'file-as'})
                 m.text = author.sort
                 metadata.append(m)
 
@@ -492,7 +492,7 @@ def read_book_producers(root, prefixes, refines):
         if val:
             props = properties_for_id_with_scheme(item.get('id'), prefixes, refines)
             role = props.get('role')
-            opf_role = item.get(OPF('role'))
+            opf_role = item.get(const.OPF_ROLE)
             if role:
                 if is_relators_role(props, 'bkp'):
                     ans.append(normalize_whitespace(val))
@@ -504,18 +504,18 @@ def read_book_producers(root, prefixes, refines):
 def set_book_producers(root, prefixes, refines, producers):
     for item in XPath('./opf:metadata/dc:contributor')(root):
         props = properties_for_id_with_scheme(item.get('id'), prefixes, refines)
-        opf_role = item.get(OPF('role'))
+        opf_role = item.get(const.OPF_ROLE)
         if (opf_role and opf_role.lower() != 'bkp') or (props.get('role') and not is_relators_role(props, 'bkp')):
             continue
         remove_element(item, refines)
     metadata = XPath('./opf:metadata')(root)[0]
     for bkp in producers:
         if bkp:
-            a = metadata.makeelement(DC('contributor'))
+            a = metadata.makeelement(const.DC_CONTRIBUTOR)
             aid = ensure_id(a)
             a.text = bkp
             metadata.append(a)
-            m = metadata.makeelement(OPF('meta'), attrib={'refines':'#'+aid, 'property':'role', 'scheme':'marc:relators'})
+            m = metadata.makeelement(const.OPF_META, attrib={'refines':'#'+aid, 'property':'role', 'scheme':'marc:relators'})
             m.text = 'bkp'
             metadata.append(m)
 # }}}
@@ -552,7 +552,7 @@ def set_pubdate(root, prefixes, refines, val):
     if not is_date_undefined(val):
         val = isoformat(val)
         m = XPath('./opf:metadata')(root)[0]
-        d = m.makeelement(DC('date'))
+        d = m.makeelement(const.DC_DATE)
         d.text = val
         m.append(d)
 
@@ -584,7 +584,7 @@ def create_timestamp(root, prefixes, m, val):
         ensure_prefix(root, prefixes, 'calibre', CALIBRE_PREFIX)
         ensure_prefix(root, prefixes, 'dcterms')
         val = w3cdtf(val)
-        d = m.makeelement(OPF('meta'), attrib={'property':'calibre:timestamp', 'scheme':'dcterms:W3CDTF'})
+        d = m.makeelement(const.OPF_META, attrib={'property':'calibre:timestamp', 'scheme':'dcterms:W3CDTF'})
         d.text = val
         m.append(d)
 
@@ -625,7 +625,7 @@ def set_last_modified(root, prefixes, refines, val=None):
     else:
         ensure_prefix(root, prefixes, 'dcterms')
         m = XPath('./opf:metadata')(root)[0]
-        meta = m.makeelement(OPF('meta'), attrib={'property':'dcterms:modified', 'scheme':'dcterms:W3CDTF'})
+        meta = m.makeelement(const.OPF_META, attrib={'property':'dcterms:modified', 'scheme':'dcterms:W3CDTF'})
         m.append(meta)
     meta.text = val
 # }}}
@@ -648,7 +648,7 @@ def set_comments(root, prefixes, refines, val):
     if val:
         val = val.strip()
         if val:
-            c = m.makeelement(DC('description'))
+            c = m.makeelement(const.DC_DESC)
             c.text = val
             m.append(c)
 # }}}
@@ -670,7 +670,7 @@ def set_publisher(root, prefixes, refines, val):
     if val:
         val = val.strip()
         if val:
-            c = m.makeelement(DC('publisher'))
+            c = m.makeelement(const.DC_PUBLISHER('publisher'))
             c.text = normalize_whitespace(val)
             m.append(c)
 # }}}
@@ -693,7 +693,7 @@ def set_tags(root, prefixes, refines, val):
     if val:
         val = uniq(list(filter(None, val)))
         for x in val:
-            c = m.makeelement(DC('subject'))
+            c = m.makeelement(const.DC_SUBJ)
             c.text = normalize_whitespace(x)
             if c.text:
                 m.append(c)
@@ -725,7 +725,7 @@ def read_rating(root, prefixes, refines):
 def create_rating(root, prefixes, val):
     ensure_prefix(root, prefixes, 'calibre', CALIBRE_PREFIX)
     m = XPath('./opf:metadata')(root)[0]
-    d = m.makeelement(OPF('meta'), attrib={'property':'calibre:rating'})
+    d = m.makeelement(const.OPF_META, attrib={'property':'calibre:rating'})
     d.text = val
     m.append(d)
 
@@ -772,7 +772,7 @@ def read_series(root, prefixes, refines):
 
 def create_series(root, refines, series, series_index):
     m = XPath('./opf:metadata')(root)[0]
-    d = m.makeelement(OPF('meta'), attrib={'property':'belongs-to-collection'})
+    d = m.makeelement(const.OPF_META, attrib={'property':'belongs-to-collection'})
     d.text = series
     m.append(d)
     set_refines(d, refines, refdef('collection-type', 'series'), refdef('group-position', series_index))
@@ -836,7 +836,7 @@ def dict_writer(name, serialize=dump_dict, remove2=True):
         if val:
             ensure_prefix(root, prefixes, 'calibre', CALIBRE_PREFIX)
             m = XPath('./opf:metadata')(root)[0]
-            d = m.makeelement(OPF('meta'), attrib={'property':'calibre:%s' % name})
+            d = m.makeelement(const.OPF_META, attrib={'property':'calibre:%s' % name})
             d.text = serialize(val)
             m.append(d)
     return writer
diff --git a/ebook_converter/ebooks/metadata/toc.py b/ebook_converter/ebooks/metadata/toc.py
index a9330a4..02f04d5 100644
--- a/ebook_converter/ebooks/metadata/toc.py
+++ b/ebook_converter/ebooks/metadata/toc.py
@@ -10,17 +10,13 @@ from lxml.builder import ElementMaker
 
 from ebook_converter.constants_old import __appname__, __version__
 from ebook_converter.ebooks.chardet import xml_to_unicode
-from ebook_converter.utils.xml_parse import safe_xml_fromstring
 from ebook_converter.utils.cleantext import clean_xml_chars
 from ebook_converter.polyglot.urllib import unquote
 
 
-__license__ = 'GPL v3'
-__copyright__ = '2010, Kovid Goyal <kovid at kovidgoyal.net>'
-
 NCX_NS = "http://www.daisy.org/z3986/2005/ncx/"
 CALIBRE_NS = "http://calibre.kovidgoyal.net/2009/metadata"
-NSMAP = {None: NCX_NS, 'calibre':CALIBRE_NS}
+NSMAP = {None: NCX_NS, 'calibre': CALIBRE_NS}
 E = ElementMaker(namespace=NCX_NS, nsmap=NSMAP)
 C = ElementMaker(namespace=CALIBRE_NS, nsmap=NSMAP)
 
@@ -30,8 +26,10 @@ def parse_html_toc(data):
     from ebook_converter.utils.cleantext import clean_xml_chars
     from lxml import etree
     if isinstance(data, bytes):
-        data = xml_to_unicode(data, strip_encoding_pats=True, resolve_entities=True)[0]
-    root = parse(clean_xml_chars(data), maybe_xhtml=True, keep_doctype=False, sanitize_names=True)
+        data = xml_to_unicode(data, strip_encoding_pats=True,
+                              resolve_entities=True)[0]
+    root = parse(clean_xml_chars(data), maybe_xhtml=True, keep_doctype=False,
+                 sanitize_names=True)
     for a in root.xpath('//*[@href and local-name()="a"]'):
         purl = urllib.parse.urlparse(unquote(a.get('href')))
         href, fragment = purl[2], purl[5]
@@ -48,8 +46,8 @@ def parse_html_toc(data):
 class TOC(list):
 
     def __init__(self, href=None, fragment=None, text=None, parent=None,
-            play_order=0, base_path=os.getcwd(), type='unknown', author=None,
-            description=None, toc_thumbnail=None):
+                 play_order=0, base_path=os.getcwd(), type='unknown',
+                 author=None, description=None, toc_thumbnail=None):
         self.href = href
         self.fragment = fragment
         if not self.fragment:
@@ -64,7 +62,7 @@ class TOC(list):
         self.toc_thumbnail = toc_thumbnail
 
     def __str__(self):
-        lines = ['TOC: %s#%s %s'%(self.href, self.fragment, self.text)]
+        lines = ['TOC: %s#%s %s' % (self.href, self.fragment, self.text)]
         for child in self:
             c = str(child).splitlines()
             for l in c:
@@ -91,12 +89,14 @@ class TOC(list):
         entry.parent = None
 
     def add_item(self, href, fragment, text, play_order=None, type='unknown',
-            author=None, description=None, toc_thumbnail=None):
+                 author=None, description=None, toc_thumbnail=None):
         if play_order is None:
-            play_order = (self[-1].play_order if len(self) else self.play_order) + 1
+            play_order = (self[-1].play_order
+                          if len(self) else self.play_order) + 1
         self.append(TOC(href=href, fragment=fragment, text=text, parent=self,
                         base_path=self.base_path, play_order=play_order,
-                        type=type, author=author, description=description, toc_thumbnail=toc_thumbnail))
+                        type=type, author=author, description=description,
+                        toc_thumbnail=toc_thumbnail))
         return self[-1]
 
     def top_level_items(self):
@@ -121,7 +121,10 @@ class TOC(list):
 
     @property
     def abspath(self):
-        'Return the file this toc entry points to as a absolute path to a file on the system.'
+        """
+        Return the file this toc entry points to as a absolute path to a file
+        on the system.
+        """
 
         if self.href is None:
             return None
@@ -136,8 +139,9 @@ class TOC(list):
             toc = toc['toc']
         if toc is None:
             try:
-                toc = opfreader.soup.find('guide').find('reference', attrs={'type':'toc'})['href']
-            except:
+                toc = (opfreader.soup.find('guide')
+                       .find('reference', attrs={'type': 'toc'})['href'])
+            except Exception:
                 for item in opfreader.manifest:
                     if 'toc' in item.href().lower():
                         toc = item.href()
@@ -151,13 +155,15 @@ class TOC(list):
                     toc = os.path.join(self.base_path, toc)
                 try:
                     if not os.path.exists(toc):
-                        bn  = os.path.basename(toc)
-                        bn  = bn.replace('_top.htm', '_toc.htm')  # Bug in BAEN OPF files
+                        bn = os.path.basename(toc)
+                        # Bug in BAEN OPF files
+                        bn = bn.replace('_top.htm', '_toc.htm')
                         toc = os.path.join(os.path.dirname(toc), bn)
 
                     self.read_html_toc(toc)
-                except:
-                    print('WARNING: Could not read Table of Contents. Continuing anyway.')
+                except Exception:
+                    print('WARNING: Could not read Table of Contents. '
+                          'Continuing anyway.')
             else:
                 path = opfreader.manifest.item(toc.lower())
                 path = getattr(path, 'path', path)
@@ -177,9 +183,9 @@ class TOC(list):
         self.base_path = os.path.dirname(toc)
         if root is None:
             with open(toc, 'rb') as f:
-                raw  = xml_to_unicode(f.read(), assume_utf8=True,
-                        strip_encoding_pats=True)[0]
-            root = safe_xml_fromstring(raw)
+                raw = xml_to_unicode(f.read(), assume_utf8=True,
+                                     strip_encoding_pats=True)[0]
+            root = etree.fromstring(raw)
         xpn = {'re': 'http://exslt.org/regular-expressions'}
         XPath = functools.partial(etree.XPath, namespaces=xpn)
 
@@ -197,7 +203,7 @@ class TOC(list):
         def process_navpoint(np, dest):
             try:
                 play_order = int(get_attr(np, 1))
-            except:
+            except Exception:
                 play_order = 1
             href = fragment = text = None
             nd = dest
@@ -207,7 +213,7 @@ class TOC(list):
                 text = ''
                 for txt in txt_path(nl):
                     text += etree.tostring(txt, method='text',
-                            encoding='unicode', with_tail=False)
+                                           encoding='unicode', with_tail=False)
                 content = content_path(np)
                 if content and text:
                     content = content[0]
@@ -242,17 +248,14 @@ class TOC(list):
                 self.add_item(href, fragment, txt)
 
     def render(self, stream, uid):
-        root = E.ncx(
-                E.head(
-                    E.meta(name='dtb:uid', content=str(uid)),
-                    E.meta(name='dtb:depth', content=str(self.depth())),
-                    E.meta(name='dtb:generator', content='%s (%s)'%(__appname__,
-                        __version__)),
-                    E.meta(name='dtb:totalPageCount', content='0'),
-                    E.meta(name='dtb:maxPageNumber', content='0'),
-                ),
-                E.docTitle(E.text('Table of Contents')),
-        )
+        root = E.ncx(E.head(E.meta(name='dtb:uid', content=str(uid)),
+                            E.meta(name='dtb:depth',
+                                   content=str(self.depth())),
+                            E.meta(name='dtb:generator', content='%s (%s)' %
+                                   (__appname__, __version__)),
+                            E.meta(name='dtb:totalPageCount', content='0'),
+                            E.meta(name='dtb:maxPageNumber', content='0')),
+                     E.docTitle(E.text('Table of Contents')))
         navmap = E.navMap()
         root.append(navmap)
         root.set('{http://www.w3.org/XML/1998/namespace}lang', 'en')
@@ -263,12 +266,12 @@ class TOC(list):
             if not text:
                 text = ''
             c[1] += 1
-            item_id = 'num_%d'%c[1]
+            item_id = 'num_%d' % c[1]
             text = clean_xml_chars(text)
             elem = E.navPoint(
                     E.navLabel(E.text(re.sub(r'\s+', ' ', text))),
                     E.content(src=str(np.href)+(('#' + str(np.fragment))
-                        if np.fragment else '')),
+                                                if np.fragment else '')),
                     id=item_id,
                     playOrder=str(np.play_order)
             )
@@ -282,7 +285,8 @@ class TOC(list):
                 try:
                     elem.append(C.meta(desc, name='description'))
                 except ValueError:
-                    elem.append(C.meta(clean_xml_chars(desc), name='description'))
+                    elem.append(C.meta(clean_xml_chars(desc),
+                                       name='description'))
             idx = getattr(np, 'toc_thumbnail', None)
             if idx:
                 elem.append(C.meta(idx, name='toc_thumbnail'))
@@ -293,5 +297,5 @@ class TOC(list):
         for np in self:
             navpoint(navmap, np)
         raw = etree.tostring(root, encoding='utf-8', xml_declaration=True,
-                pretty_print=True)
+                             pretty_print=True)
         stream.write(raw)
diff --git a/ebook_converter/ebooks/metadata/utils.py b/ebook_converter/ebooks/metadata/utils.py
index 0fbce09..031c917 100644
--- a/ebook_converter/ebooks/metadata/utils.py
+++ b/ebook_converter/ebooks/metadata/utils.py
@@ -1,12 +1,13 @@
 from collections import namedtuple
 
+from lxml import etree
+
 from ebook_converter.ebooks.chardet import xml_to_unicode
-from ebook_converter.ebooks.oeb.base import OPF
+from ebook_converter.ebooks.oeb import base
 from ebook_converter.ebooks.oeb.polish.utils import guess_type
 from ebook_converter.spell import parse_lang_code
 from ebook_converter.utils.cleantext import clean_xml_chars
 from ebook_converter.utils.localization import lang_as_iso639_1
-from ebook_converter.utils.xml_parse import safe_xml_fromstring
 
 
 OPFVersion = namedtuple('OPFVersion', 'major minor patch')
@@ -35,23 +36,26 @@ def parse_opf(stream_or_path):
     raw = stream.read()
     if not raw:
         raise ValueError('Empty file: '+getattr(stream, 'name', 'stream'))
-    raw, encoding = xml_to_unicode(raw, strip_encoding_pats=True, resolve_entities=True, assume_utf8=True)
+    raw, encoding = xml_to_unicode(raw, strip_encoding_pats=True,
+                                   resolve_entities=True, assume_utf8=True)
     raw = raw[raw.find('<'):]
-    root = safe_xml_fromstring(clean_xml_chars(raw))
+    root = etree.fromstring(clean_xml_chars(raw))
     if root is None:
         raise ValueError('Not an OPF file')
     return root
 
 
 def normalize_languages(opf_languages, mi_languages):
-    ' Preserve original country codes and use 2-letter lang codes where possible '
+    """
+    Preserve original country codes and use 2-letter lang codes where possible
+    """
     def parse(x):
         try:
             return parse_lang_code(x)
         except ValueError:
             return None
     opf_languages = filter(None, map(parse, opf_languages))
-    cc_map = {c.langcode:c.countrycode for c in opf_languages}
+    cc_map = {c.langcode: c.countrycode for c in opf_languages}
     mi_languages = filter(None, map(parse, mi_languages))
 
     def norm(x):
@@ -83,9 +87,9 @@ def create_manifest_item(root, href_template, id_template, media_type=None):
     all_hrefs = frozenset(root.xpath('//*/@href'))
     href = ensure_unique(href_template, all_hrefs)
     item_id = ensure_unique(id_template, all_ids)
-    manifest = root.find(OPF('manifest'))
+    manifest = root.find(base.tag('opf', 'manifest'))
     if manifest is not None:
-        i = manifest.makeelement(OPF('item'))
+        i = manifest.makeelement(base.tag('opf', 'item'))
         i.set('href', href), i.set('id', item_id)
         i.set('media-type', media_type or guess_type(href_template))
         manifest.append(i)
@@ -93,6 +97,7 @@ def create_manifest_item(root, href_template, id_template, media_type=None):
 
 
 def pretty_print_opf(root):
-    from ebook_converter.ebooks.oeb.polish.pretty import pretty_opf, pretty_xml_tree
+    from ebook_converter.ebooks.oeb.polish.pretty import pretty_opf, \
+            pretty_xml_tree
     pretty_opf(root)
     pretty_xml_tree(root)
diff --git a/ebook_converter/ebooks/metadata/xmp.py b/ebook_converter/ebooks/metadata/xmp.py
index fc6e154..e9a7faf 100644
--- a/ebook_converter/ebooks/metadata/xmp.py
+++ b/ebook_converter/ebooks/metadata/xmp.py
@@ -1,44 +1,43 @@
-import re, sys, copy, json
-from itertools import repeat
-from collections import defaultdict
+import collections
+import copy
+import itertools
+import json
+import re
+import sys
 
 from lxml import etree
 from lxml.builder import ElementMaker
 
 from ebook_converter import prints
 from ebook_converter.ebooks.metadata import check_isbn, check_doi
-from ebook_converter.utils.xml_parse import safe_xml_fromstring
 from ebook_converter.ebooks.metadata.book.base import Metadata
 from ebook_converter.ebooks.metadata.opf2 import dump_dict
 from ebook_converter.utils.date import parse_date, isoformat, now
-from ebook_converter.utils.localization import canonicalize_lang, lang_as_iso639_1
+from ebook_converter.utils.localization import canonicalize_lang, \
+        lang_as_iso639_1
 
 
-__license__ = 'GPL v3'
-__copyright__ = '2014, Kovid Goyal <kovid at kovidgoyal.net>'
+_xml_declaration = re.compile(r'<\?xml[^<>]+encoding\s*=\s*[\'"](.*?)'
+                              r'[\'"][^<>]*>', re.IGNORECASE)
 
-_xml_declaration = re.compile(r'<\?xml[^<>]+encoding\s*=\s*[\'"](.*?)[\'"][^<>]*>', re.IGNORECASE)
-
-NS_MAP = {
-    'rdf': 'http://www.w3.org/1999/02/22-rdf-syntax-ns#',
-    'dc': 'http://purl.org/dc/elements/1.1/',
-    'pdf': 'http://ns.adobe.com/pdf/1.3/',
-    'pdfx': 'http://ns.adobe.com/pdfx/1.3/',
-    'xmp': 'http://ns.adobe.com/xap/1.0/',
-    'xmpidq': 'http://ns.adobe.com/xmp/Identifier/qual/1.0/',
-    'xmpMM': 'http://ns.adobe.com/xap/1.0/mm/',
-    'xmpRights': 'http://ns.adobe.com/xap/1.0/rights/',
-    'xmpBJ': 'http://ns.adobe.com/xap/1.0/bj/',
-    'xmpTPg': 'http://ns.adobe.com/xap/1.0/t/pg/',
-    'xmpDM': 'http://ns.adobe.com/xmp/1.0/DynamicMedia/',
-    'prism': 'http://prismstandard.org/namespaces/basic/2.0/',
-    'crossmark': 'http://crossref.org/crossmark/1.0/',
-    'xml': 'http://www.w3.org/XML/1998/namespace',
-    'x': 'adobe:ns:meta/',
-    'calibre': 'http://calibre-ebook.com/xmp-namespace',
-    'calibreSI': 'http://calibre-ebook.com/xmp-namespace-series-index',
-    'calibreCC': 'http://calibre-ebook.com/xmp-namespace-custom-columns',
-}
+NS_MAP = {'rdf': 'http://www.w3.org/1999/02/22-rdf-syntax-ns#',
+          'dc': 'http://purl.org/dc/elements/1.1/',
+          'pdf': 'http://ns.adobe.com/pdf/1.3/',
+          'pdfx': 'http://ns.adobe.com/pdfx/1.3/',
+          'xmp': 'http://ns.adobe.com/xap/1.0/',
+          'xmpidq': 'http://ns.adobe.com/xmp/Identifier/qual/1.0/',
+          'xmpMM': 'http://ns.adobe.com/xap/1.0/mm/',
+          'xmpRights': 'http://ns.adobe.com/xap/1.0/rights/',
+          'xmpBJ': 'http://ns.adobe.com/xap/1.0/bj/',
+          'xmpTPg': 'http://ns.adobe.com/xap/1.0/t/pg/',
+          'xmpDM': 'http://ns.adobe.com/xmp/1.0/DynamicMedia/',
+          'prism': 'http://prismstandard.org/namespaces/basic/2.0/',
+          'crossmark': 'http://crossref.org/crossmark/1.0/',
+          'xml': 'http://www.w3.org/XML/1998/namespace',
+          'x': 'adobe:ns:meta/',
+          'calibre': 'http://calibre-ebook.com/xmp-namespace',
+          'calibreSI': 'http://calibre-ebook.com/xmp-namespace-series-index',
+          'calibreCC': 'http://calibre-ebook.com/xmp-namespace-custom-columns'}
 KNOWN_ID_SCHEMES = {'isbn', 'url', 'doi'}
 
 
@@ -63,7 +62,7 @@ def parse_xmp_packet(raw_bytes):
     pat = r'''<?xpacket\s+[^>]*?begin\s*=\s*['"]([^'"]*)['"]'''
     encodings = ('8', '16-le', '16-be', '32-le', '32-be')
     header = raw_bytes[:1024]
-    emap = {'\ufeff'.encode('utf-'+x):'utf-'+x for x in encodings}
+    emap = {'\ufeff'.encode('utf-'+x): 'utf-'+x for x in encodings}
     emap[b''] = 'utf-8'
     for q in encodings:
         m = re.search(pat.encode('utf-'+q), header)
@@ -71,15 +70,19 @@ def parse_xmp_packet(raw_bytes):
             enc = emap.get(m.group(1), enc)
             break
     if enc is None:
-        return safe_xml_fromstring(raw_bytes)
-    raw = _xml_declaration.sub('', raw_bytes.decode(enc))  # lxml barfs if encoding declaration present in unicode string
-    return safe_xml_fromstring(raw)
+        return etree.fromstring(raw_bytes)
+    # lxml barfs if encoding declaration present in unicode string
+    raw = _xml_declaration.sub('', raw_bytes.decode(enc))
+    return etree.fromstring(raw)
 
 
 def serialize_xmp_packet(root, encoding='utf-8'):
-    root.tail = '\n' + '\n'.join(repeat(' '*100, 30))  # Adobe spec recommends inserting padding at the end of the packet
-    raw_bytes = etree.tostring(root, encoding=encoding, pretty_print=True, with_tail=True, method='xml')
-    return b'<?xpacket begin="%s" id="W5M0MpCehiHzreSzNTczkc9d"?>\n%s\n<?xpacket end="w"?>' % ('\ufeff'.encode(encoding), raw_bytes)
+    # Adobe spec recommends inserting padding at the end of the packet
+    root.tail = '\n' + '\n'.join(itertools.repeat(' '*100, 30))
+    raw_bytes = etree.tostring(root, encoding=encoding, pretty_print=True,
+                               with_tail=True, method='xml')
+    return ('<?xpacket begin="%s" id="W5M0MpCehiHzreSzNTczkc9d"?>\n%s\n'
+            '<?xpacket end="w"?>' % ('\ufeff'.encode(encoding), raw_bytes))
 
 
 def read_simple_property(elem):
@@ -106,14 +109,15 @@ def read_sequence(parent):
         yield read_simple_property(item)
 
 
-def uniq(vals, kmap=lambda x:x):
+def uniq(vals, kmap=lambda x: x):
     ''' Remove all duplicates from vals, while preserving order. kmap must be a
     callable that returns a hashable value for every item in vals '''
     vals = vals or ()
     lvals = (kmap(x) for x in vals)
     seen = set()
     seen_add = seen.add
-    return tuple(x for x, k in zip(vals, lvals) if k not in seen and not seen_add(k))
+    return tuple(x for x, k in zip(vals, lvals) if k not in seen
+                 and not seen_add(k))
 
 
 def multiple_sequences(expr, root):
@@ -170,7 +174,8 @@ def read_series(root):
 
 def read_user_metadata(mi, root):
     from ebook_converter.utils.config import from_json
-    from ebook_converter.ebooks.metadata.book.json_codec import decode_is_multiple
+    from ebook_converter.ebooks.metadata.book.json_codec import \
+        decode_is_multiple
     fields = set()
     for item in XPath('//calibre:custom_metadata')(root):
         for li in XPath('./rdf:Bag/rdf:li')(item):
@@ -186,7 +191,7 @@ def read_user_metadata(mi, root):
                             decode_is_multiple(fm)
                             mi.set_user_metadata(name, fm)
                             fields.add(name)
-                        except:
+                        except Exception:
                             prints('Failed to read user metadata:', name)
                             import traceback
                             traceback.print_exc()
@@ -194,13 +199,17 @@ def read_user_metadata(mi, root):
 
 def read_xmp_identifers(parent):
     ''' For example:
-    <rdf:li rdf:parseType="Resource"><xmpidq:Scheme>URL</xmp:idq><rdf:value>http://foo.com</rdf:value></rdf:li>
+    <rdf:li rdf:parseType="Resource"><xmpidq:Scheme>URL</xmp:idq>
+    <rdf:value>http://foo.com</rdf:value></rdf:li>
     or the longer form:
-    <rdf:li><rdf:Description><xmpidq:Scheme>URL</xmp:idq><rdf:value>http://foo.com</rdf:value></rdf:Description></rdf:li>
+    <rdf:li><rdf:Description><xmpidq:Scheme>URL</xmp:idq>
+    <rdf:value>http://foo.com</rdf:value></rdf:Description></rdf:li>
     '''
     for li in XPath('./rdf:Bag/rdf:li')(parent):
-        is_resource = li.attrib.get(expand('rdf:parseType'), None) == 'Resource'
-        is_resource = is_resource or (len(li) == 1 and li[0].tag == expand('rdf:Description'))
+        is_resource = li.attrib.get(expand('rdf:parseType'),
+                                    None) == 'Resource'
+        is_resource = is_resource or (len(li) == 1 and
+                                      li[0].tag == expand('rdf:Description'))
         if not is_resource:
             yield None, li.text or ''
         value = XPath('descendant::rdf:value')(li)
@@ -241,12 +250,15 @@ def metadata_from_xmp_packet(raw_bytes):
         if title.startswith(r'\376\377'):
             # corrupted XMP packet generated by Nitro PDF. See
             # https://bugs.launchpad.net/calibre/+bug/1541981
-            raise ValueError('Corrupted XMP metadata packet detected, probably generated by Nitro PDF')
+            raise ValueError('Corrupted XMP metadata packet detected, '
+                             'probably generated by Nitro PDF')
         mi.title = title
     authors = multiple_sequences('//dc:creator', root)
     if authors:
         mi.authors = authors
-    tags = multiple_sequences('//dc:subject', root) or multiple_sequences('//pdf:Keywords', root)
+    tags = multiple_sequences('//dc:subject',
+                              root) or multiple_sequences('//pdf:Keywords',
+                                                          root)
     if tags:
         mi.tags = tags
     comments = first_alt('//dc:description', root)
@@ -256,8 +268,10 @@ def metadata_from_xmp_packet(raw_bytes):
     if publishers:
         mi.publisher = publishers[0]
     try:
-        pubdate = parse_date(first_sequence('//dc:date', root) or first_simple('//xmp:CreateDate', root), assume_utc=False)
-    except:
+        pubdate = (parse_date(first_sequence('//dc:date', root) or
+                              first_simple('//xmp:CreateDate', root),
+                              assume_utc=False))
+    except Exception:
         pass
     else:
         mi.pubdate = pubdate
@@ -291,7 +305,7 @@ def metadata_from_xmp_packet(raw_bytes):
         if val:
             try:
                 setattr(mi, x, json.loads(val))
-            except:
+            except Exception:
                 pass
 
     languages = multiple_sequences('//dc:language', root)
@@ -319,7 +333,7 @@ def metadata_from_xmp_packet(raw_bytes):
                     identifiers[scheme] = val
 
     # Check Dublin Core for recognizable identifier types
-    for scheme, check_func in {'doi':check_doi, 'isbn':check_isbn}.items():
+    for scheme, check_func in {'doi': check_doi, 'isbn': check_isbn}.items():
         if scheme not in identifiers:
             val = check_func(first_simple('//dc:identifier', root))
             if val:
@@ -359,17 +373,21 @@ def consolidate_metadata(info_mi, info):
         else:
             prefer_info = info_date > xmp_mi.metadata_date
     if prefer_info:
-        info_mi.title, info_mi.authors, info_mi.tags = info_title, info_authors, info_tags
+        info_mi.title = info_title
+        info_mi.authors = info_authors
+        info_mi.tags = info_tags
     else:
         # We'll use the xmp tags/authors but fallback to the info ones if the
-        # xmp does not have tags/authors. smart_update() should have taken care of
-        # the rest
-        info_mi.authors, info_mi.tags = (info_authors if xmp_mi.is_null('authors') else xmp_mi.authors), xmp_mi.tags or info_tags
+        # xmp does not have tags/authors. smart_update() should have taken care
+        # of the rest
+        info_mi.authors = (info_authors if xmp_mi.is_null('authors')
+                           else xmp_mi.authors)
+        info_mi.tags = xmp_mi.tags or info_tags
     return info_mi
 
 
 def nsmap(*args):
-    return {x:NS_MAP[x] for x in args}
+    return {x: NS_MAP[x] for x in args}
 
 
 def create_simple_property(parent, tag, value):
@@ -435,7 +453,8 @@ def create_series(calibre, series, series_index):
 
 def create_user_metadata(calibre, all_user_metadata):
     from ebook_converter.utils.config import to_json
-    from ebook_converter.ebooks.metadata.book.json_codec import object_to_unicode, encode_is_multiple
+    from ebook_converter.ebooks.metadata.book.json_codec import \
+        object_to_unicode, encode_is_multiple
 
     s = calibre.makeelement(expand('calibre:custom_metadata'))
     calibre.append(s)
@@ -447,7 +466,7 @@ def create_user_metadata(calibre, all_user_metadata):
             encode_is_multiple(fm)
             fm = object_to_unicode(fm)
             fm = json.dumps(fm, default=to_json, ensure_ascii=False)
-        except:
+        except Exception:
             prints('Failed to write user metadata:', name)
             import traceback
             traceback.print_exc()
@@ -471,7 +490,8 @@ def metadata_to_xmp_packet(mi):
     dc = rdf.makeelement(expand('rdf:Description'), nsmap=nsmap('dc'))
     dc.set(expand('rdf:about'), '')
     rdf.append(dc)
-    for prop, tag in {'title':'dc:title', 'comments':'dc:description'}.items():
+    for prop, tag in {'title': 'dc:title',
+                      'comments': 'dc:description'}.items():
         val = mi.get(prop) or ''
         create_alt_property(dc, tag, val)
     for prop, (tag, ordered) in {'authors': ('dc:creator', True),
@@ -482,18 +502,23 @@ def metadata_to_xmp_packet(mi):
             val = [val]
         create_sequence_property(dc, tag, val, ordered)
     if not mi.is_null('pubdate'):
-        create_sequence_property(dc, 'dc:date', [isoformat(mi.pubdate, as_utc=False)])  # Adobe spec recommends local time
+        # Adobe spec recommends local time
+        create_sequence_property(dc, 'dc:date',
+                                 [isoformat(mi.pubdate, as_utc=False)])
     if not mi.is_null('languages'):
-        langs = list(filter(None, map(lambda x:lang_as_iso639_1(x) or canonicalize_lang(x), mi.languages)))
+        langs = list(filter(None, map(lambda x: lang_as_iso639_1(x) or
+                                      canonicalize_lang(x), mi.languages)))
         if langs:
             create_sequence_property(dc, 'dc:language', langs, ordered=False)
 
-    xmp = rdf.makeelement(expand('rdf:Description'), nsmap=nsmap('xmp', 'xmpidq'))
+    xmp = rdf.makeelement(expand('rdf:Description'),
+                          nsmap=nsmap('xmp', 'xmpidq'))
     xmp.set(expand('rdf:about'), '')
     rdf.append(xmp)
     extra_ids = {}
     for x in ('prism', 'pdfx'):
-        p = extra_ids[x] = rdf.makeelement(expand('rdf:Description'), nsmap=nsmap(x))
+        p = extra_ids[x] = rdf.makeelement(expand('rdf:Description'),
+                                           nsmap=nsmap(x))
         p.set(expand('rdf:about'), '')
         rdf.append(p)
 
@@ -503,7 +528,7 @@ def metadata_to_xmp_packet(mi):
         for scheme, val in identifiers.items():
             if scheme in {'isbn', 'doi'}:
                 for prefix, parent in extra_ids.items():
-                    ie = parent.makeelement(expand('%s:%s'%(prefix, scheme)))
+                    ie = parent.makeelement(expand('%s:%s' % (prefix, scheme)))
                     ie.text = val
                     parent.append(ie)
 
@@ -511,7 +536,8 @@ def metadata_to_xmp_packet(mi):
     d.text = isoformat(now(), as_utc=False)
     xmp.append(d)
 
-    calibre = rdf.makeelement(expand('rdf:Description'), nsmap=nsmap('calibre', 'calibreSI', 'calibreCC'))
+    calibre = rdf.makeelement(expand('rdf:Description'),
+                              nsmap=nsmap('calibre', 'calibreSI', 'calibreCC'))
     calibre.set(expand('rdf:about'), '')
     rdf.append(calibre)
     if not mi.is_null('rating'):
@@ -524,7 +550,8 @@ def metadata_to_xmp_packet(mi):
     if not mi.is_null('series'):
         create_series(calibre, mi.series, mi.series_index)
     if not mi.is_null('timestamp'):
-        create_simple_property(calibre, 'calibre:timestamp', isoformat(mi.timestamp, as_utc=False))
+        create_simple_property(calibre, 'calibre:timestamp',
+                               isoformat(mi.timestamp, as_utc=False))
     for x in ('author_link_map', 'user_categories'):
         val = getattr(mi, x, None)
         if val:
@@ -550,10 +577,11 @@ def find_used_namespaces(elem):
 
 def find_preferred_prefix(namespace, elems):
     for elem in elems:
-        ans = {v:k for k, v in elem.nsmap.items()}.get(namespace, None)
+        ans = {v: k for k, v in elem.nsmap.items()}.get(namespace, None)
         if ans is not None:
             return ans
-        return find_preferred_prefix(namespace, elem.iterchildren(etree.Element))
+        return find_preferred_prefix(namespace,
+                                     elem.iterchildren(etree.Element))
 
 
 def find_nsmap(elems):
@@ -562,7 +590,7 @@ def find_nsmap(elems):
         used_namespaces |= find_used_namespaces(elem)
     ans = {}
     used_namespaces -= {NS_MAP['xml'], NS_MAP['x'], None, NS_MAP['rdf']}
-    rmap = {v:k for k, v in NS_MAP.items()}
+    rmap = {v: k for k, v in NS_MAP.items()}
     i = 0
     for ns in used_namespaces:
         if ns in rmap:
@@ -578,7 +606,10 @@ def find_nsmap(elems):
 
 
 def clone_into(parent, elem):
-    ' Clone the element, assuming that all namespace declarations are present in parent '
+    """
+    Clone the element, assuming that all namespace declarations are present
+    in parent
+    """
     clone = parent.makeelement(elem.tag)
     parent.append(clone)
     if elem.text and not elem.text.isspace():
@@ -591,28 +622,38 @@ def clone_into(parent, elem):
 
 
 def merge_xmp_packet(old, new):
-    ''' Merge metadata present in the old packet that is not present in the new
+    """
+    Merge metadata present in the old packet that is not present in the new
     one into the new one. Assumes the new packet was generated by
-    metadata_to_xmp_packet() '''
+    metadata_to_xmp_packet()
+    """
     old, new = parse_xmp_packet(old), parse_xmp_packet(new)
-    # As per the adobe spec all metadata items have to be present inside top-level rdf:Description containers
+    # As per the adobe spec all metadata items have to be present inside
+    # top-level rdf:Description containers
     item_xpath = XPath('//rdf:RDF/rdf:Description/*')
 
     # First remove all data fields that metadata_to_xmp_packet() knowns about,
     # since either they will have been set or if not present, imply they have
     # been cleared
-    defined_tags = {expand(prefix + ':' + scheme) for prefix in ('prism', 'pdfx') for scheme in KNOWN_ID_SCHEMES}
-    defined_tags |= {expand('dc:' + x) for x in ('identifier', 'title', 'creator', 'date', 'description', 'language', 'publisher', 'subject')}
-    defined_tags |= {expand('xmp:' + x) for x in ('MetadataDate', 'Identifier')}
+    defined_tags = {expand(prefix + ':' + scheme)
+                    for prefix in ('prism', 'pdfx')
+                    for scheme in KNOWN_ID_SCHEMES}
+    defined_tags |= {expand('dc:' + x)
+                     for x in ('identifier', 'title', 'creator', 'date',
+                               'description', 'language', 'publisher',
+                               'subject')}
+    defined_tags |= {expand('xmp:' + x)
+                     for x in ('MetadataDate', 'Identifier')}
     # For redundancy also remove all fields explicitly set in the new packet
     defined_tags |= {x.tag for x in item_xpath(new)}
     calibrens = '{%s}' % NS_MAP['calibre']
     for elem in item_xpath(old):
-        if elem.tag in defined_tags or (elem.tag and elem.tag.startswith(calibrens)):
+        if elem.tag in defined_tags or (elem.tag and
+                                        elem.tag.startswith(calibrens)):
             elem.getparent().remove(elem)
 
     # Group all items into groups based on their namespaces
-    groups = defaultdict(list)
+    groups = collections.defaultdict(list)
     for item in item_xpath(new):
         ns = item.nsmap[item.prefix]
         groups[ns].append(item)
@@ -626,9 +667,14 @@ def merge_xmp_packet(old, new):
     root = A.xmpmeta(R.RDF)
     rdf = root[0]
 
-    for namespace in sorted(groups, key=lambda x:{NS_MAP['dc']:'a', NS_MAP['xmp']:'b', NS_MAP['calibre']:'c'}.get(x, 'z'+x)):
+    for namespace in sorted(groups,
+                            key=lambda x: {NS_MAP['dc']: 'a',
+                                           NS_MAP['xmp']: 'b',
+                                           NS_MAP['calibre']: 'c'}.get(x,
+                                                                       'z'+x)):
         items = groups[namespace]
-        desc = rdf.makeelement(expand('rdf:Description'), nsmap=find_nsmap(items))
+        desc = rdf.makeelement(expand('rdf:Description'),
+                               nsmap=find_nsmap(items))
         desc.set(expand('rdf:about'), '')
         rdf.append(desc)
         for item in items:
diff --git a/ebook_converter/ebooks/mobi/mobiml.py b/ebook_converter/ebooks/mobi/mobiml.py
index 07cea1c..d833d88 100644
--- a/ebook_converter/ebooks/mobi/mobiml.py
+++ b/ebook_converter/ebooks/mobi/mobiml.py
@@ -5,8 +5,10 @@ import copy
 import re
 import numbers
 from lxml import etree
-from ebook_converter.ebooks.oeb.base import namespace, barename
-from ebook_converter.ebooks.oeb.base import XHTML, XHTML_NS, urlnormalize
+
+from ebook_converter import constants as const
+from ebook_converter.ebooks.oeb import base
+from ebook_converter.ebooks.oeb import parse_utils
 from ebook_converter.ebooks.oeb.stylizer import Stylizer
 from ebook_converter.ebooks.oeb.transforms.flatcss import KeyMapper
 from ebook_converter.ebooks.mobi.utils import convert_color_for_font_tag
@@ -23,7 +25,7 @@ def MBP(name):
     return '{%s}%s' % (MBP_NS, name)
 
 
-MOBI_NSMAP = {None: XHTML_NS, 'mbp': MBP_NS}
+MOBI_NSMAP = {None: const.XHTML_NS, 'mbp': const.MBP_NS}
 INLINE_TAGS = {'span', 'a', 'code', 'u', 's', 'big', 'strike', 'tt', 'font', 'q', 'i', 'b', 'em', 'strong', 'sup', 'sub'}
 HEADER_TAGS = {'h1', 'h2', 'h3', 'h4', 'h5', 'h6'}
 # GR: Added 'caption' to both sets
@@ -129,9 +131,9 @@ class MobiMLizer(object):
         'Iterate over the spine and convert it to MOBIML'
         for item in self.oeb.spine:
             stylizer = Stylizer(item.data, item.href, self.oeb, self.opts, self.profile)
-            body = item.data.find(XHTML('body'))
-            nroot = etree.Element(XHTML('html'), nsmap=MOBI_NSMAP)
-            nbody = etree.SubElement(nroot, XHTML('body'))
+            body = item.data.find(base.tag('xhtml', 'body'))
+            nroot = etree.Element(base.tag('xhtml', 'html'), nsmap=MOBI_NSMAP)
+            nbody = etree.SubElement(nroot, base.tag('xhtml', 'body'))
             self.current_spine_item = item
             self.mobimlize_elem(body, stylizer, BlockState(nbody),
                                 [FormatState()])
@@ -162,7 +164,7 @@ class MobiMLizer(object):
         lines = text.split('\n')
         result = lines[:1]
         for line in lines[1:]:
-            result.append(etree.Element(XHTML('br')))
+            result.append(etree.Element(base.tag('xhtml', 'br')))
             if line:
                 result.append(line)
         return result
@@ -194,7 +196,7 @@ class MobiMLizer(object):
                 indent = (indent / abs(indent)) * self.profile.fbase
             if tag in NESTABLE_TAGS and not istate.rendered:
                 para = wrapper = etree.SubElement(
-                    parent, XHTML(tag), attrib=istate.attrib)
+                    parent, base.tag('xhtml', tag), attrib=istate.attrib)
                 bstate.nested.append(para)
                 if tag == 'li' and len(istates) > 1:
                     istates[-2].list_num += 1
@@ -203,21 +205,21 @@ class MobiMLizer(object):
                 para = wrapper = bstate.nested[-1]
             elif not self.opts.mobi_ignore_margins and left > 0 and indent >= 0:
                 ems = self.profile.mobi_ems_per_blockquote
-                para = wrapper = etree.SubElement(parent, XHTML('blockquote'))
+                para = wrapper = etree.SubElement(parent, base.tag('xhtml', 'blockquote'))
                 para = wrapper
                 emleft = int(round(left / self.profile.fbase)) - ems
                 emleft = min((emleft, 10))
                 while emleft > ems / 2:
-                    para = etree.SubElement(para, XHTML('blockquote'))
+                    para = etree.SubElement(para, base.tag('xhtml', 'blockquote'))
                     emleft -= ems
             else:
-                para = wrapper = etree.SubElement(parent, XHTML('p'))
+                para = wrapper = etree.SubElement(parent, base.tag('xhtml', 'p'))
             bstate.inline = bstate.para = para
             vspace = bstate.vpadding + bstate.vmargin
             bstate.vpadding = bstate.vmargin = 0
             if tag not in TABLE_TAGS:
                 if tag in ('ul', 'ol') and vspace > 0:
-                    wrapper.addprevious(etree.Element(XHTML('div'),
+                    wrapper.addprevious(etree.Element(base.tag('xhtml', 'div'),
                         height=self.mobimlize_measure(vspace)))
                 else:
                     wrapper.attrib['height'] = self.mobimlize_measure(vspace)
@@ -225,7 +227,7 @@ class MobiMLizer(object):
             elif tag == 'table' and vspace > 0:
                 vspace = int(round(vspace / self.profile.fbase))
                 while vspace > 0:
-                    wrapper.addprevious(etree.Element(XHTML('br')))
+                    wrapper.addprevious(etree.Element(base.tag('xhtml', 'br')))
                     vspace -= 1
             if istate.halign != 'auto' and isinstance(istate.halign, (bytes, str)):
                 if isinstance(istate.halign, bytes):
@@ -237,7 +239,7 @@ class MobiMLizer(object):
             bstate.inline = para
             pstate = bstate.istate = None
             try:
-                etree.SubElement(para, XHTML(tag), attrib=istate.attrib)
+                etree.SubElement(para, base.tag('xhtml', tag), attrib=istate.attrib)
             except:
                 print('Invalid subelement:', para, tag, istate.attrib)
                 raise
@@ -245,7 +247,7 @@ class MobiMLizer(object):
             para.attrib['valign'] = 'top'
         if istate.ids:
             for id_ in istate.ids:
-                anchor = etree.Element(XHTML('a'), attrib={'id': id_})
+                anchor = etree.Element(base.tag('xhtml', 'a'), attrib={'id': id_})
                 if tag == 'li':
                     try:
                         last = bstate.body[-1][-1]
@@ -262,7 +264,7 @@ class MobiMLizer(object):
                     # This could potentially break if inserting an anchor at
                     # this point in the markup is illegal, but I cannot think
                     # of such a case offhand.
-                    if barename(last.tag) in LEAF_TAGS:
+                    if parse_utils.barename(last.tag) in LEAF_TAGS:
                         last.addprevious(anchor)
                     else:
                         last.append(anchor)
@@ -279,28 +281,28 @@ class MobiMLizer(object):
             elif pstate and pstate.href == href:
                 inline = bstate.anchor
             else:
-                inline = etree.SubElement(inline, XHTML('a'), href=href)
+                inline = etree.SubElement(inline, base.tag('xhtml', 'a'), href=href)
                 bstate.anchor = inline
 
             if fsize != 3:
-                inline = etree.SubElement(inline, XHTML('font'),
+                inline = etree.SubElement(inline, base.tag('xhtml', 'font'),
                                           size=str(fsize))
             if istate.family == 'monospace':
-                inline = etree.SubElement(inline, XHTML('tt'))
+                inline = etree.SubElement(inline, base.tag('xhtml', 'tt'))
             if istate.italic:
-                inline = etree.SubElement(inline, XHTML('i'))
+                inline = etree.SubElement(inline, base.tag('xhtml', 'i'))
             if istate.bold:
-                inline = etree.SubElement(inline, XHTML('b'))
+                inline = etree.SubElement(inline, base.tag('xhtml', 'b'))
             if istate.bgcolor is not None and istate.bgcolor != 'transparent' :
-                inline = etree.SubElement(inline, XHTML('span'),
+                inline = etree.SubElement(inline, base.tag('xhtml', 'span'),
                         bgcolor=convert_color_for_font_tag(istate.bgcolor))
             if istate.fgcolor != 'black':
-                inline = etree.SubElement(inline, XHTML('font'),
+                inline = etree.SubElement(inline, base.tag('xhtml', 'font'),
                         color=convert_color_for_font_tag(istate.fgcolor))
             if istate.strikethrough:
-                inline = etree.SubElement(inline, XHTML('s'))
+                inline = etree.SubElement(inline, base.tag('xhtml', 's'))
             if istate.underline:
-                inline = etree.SubElement(inline, XHTML('u'))
+                inline = etree.SubElement(inline, base.tag('xhtml', 'u'))
             bstate.inline = inline
         bstate.istate = istate
         inline = bstate.inline
@@ -318,7 +320,7 @@ class MobiMLizer(object):
     def mobimlize_elem(self, elem, stylizer, bstate, istates,
             ignore_valign=False):
         if not isinstance(elem.tag, (str, bytes)) \
-           or namespace(elem.tag) != XHTML_NS:
+           or parse_utils.namespace(elem.tag) != const.XHTML_NS:
             return
         style = stylizer.style(elem)
         # <mbp:frame-set/> does not exist lalalala
@@ -333,10 +335,10 @@ class MobiMLizer(object):
                 elem.text = None
                 elem.set('id', id_)
                 elem.tail = tail
-                elem.tag = XHTML('a')
+                elem.tag = base.tag('xhtml', 'a')
             else:
                 return
-        tag = barename(elem.tag)
+        tag = parse_utils.barename(elem.tag)
         istate = copy.copy(istates[-1])
         istate.rendered = False
         istate.list_num = 0
@@ -451,7 +453,7 @@ class MobiMLizer(object):
             if 'width' not in istate.attrib or 'height' not in istate.attrib:
                 href = self.current_spine_item.abshref(elem.attrib['src'])
                 try:
-                    item = self.oeb.manifest.hrefs[urlnormalize(href)]
+                    item = self.oeb.manifest.hrefs[base.urlnormalize(href)]
                 except:
                     self.oeb.logger.warn('Failed to find image:',
                             href)
@@ -534,9 +536,9 @@ class MobiMLizer(object):
             isinstance(valign, numbers.Number) and valign > 0)
         vtag = 'sup' if issup else 'sub'
         if not_baseline and not ignore_valign and tag not in NOT_VTAGS and not isblock:
-            nroot = etree.Element(XHTML('html'), nsmap=MOBI_NSMAP)
-            vbstate = BlockState(etree.SubElement(nroot, XHTML('body')))
-            vbstate.para = etree.SubElement(vbstate.body, XHTML('p'))
+            nroot = etree.Element(base.tag('xhtml', 'html'), nsmap=MOBI_NSMAP)
+            vbstate = BlockState(etree.SubElement(nroot, base.tag('xhtml', 'body')))
+            vbstate.para = etree.SubElement(vbstate.body, base.tag('xhtml', 'p'))
             self.mobimlize_elem(elem, stylizer, vbstate, istates,
                     ignore_valign=True)
             if len(istates) > 0:
@@ -548,8 +550,8 @@ class MobiMLizer(object):
                 self.mobimlize_content('span', '', bstate, istates)
             parent = bstate.para if bstate.inline is None else bstate.inline
             if parent is not None:
-                vtag = etree.SubElement(parent, XHTML(vtag))
-                vtag = etree.SubElement(vtag, XHTML('small'))
+                vtag = etree.SubElement(parent, base.tag('xhtml', vtag))
+                vtag = etree.SubElement(vtag, base.tag('xhtml', 'small'))
                 # Add anchors
                 for child in vbstate.body:
                     if child is not vbstate.para:
@@ -601,7 +603,7 @@ class MobiMLizer(object):
             para = bstate.para
             if para is not None and para.text == '\xa0' and len(para) < 1:
                 if style.height > 2:
-                    para.getparent().replace(para, etree.Element(XHTML('br')))
+                    para.getparent().replace(para, etree.Element(base.tag('xhtml', 'br')))
                 else:
                     # This is too small to be rendered effectively, drop it
                     para.getparent().remove(para)
diff --git a/ebook_converter/ebooks/mobi/reader/mobi8.py b/ebook_converter/ebooks/mobi/reader/mobi8.py
index a50d6bc..02b2f0e 100644
--- a/ebook_converter/ebooks/mobi/reader/mobi8.py
+++ b/ebook_converter/ebooks/mobi/reader/mobi8.py
@@ -8,6 +8,7 @@ import uuid
 
 from lxml import etree
 
+from ebook_converter import constants as const
 from ebook_converter.ebooks.mobi.reader.headers import NULL_INDEX
 from ebook_converter.ebooks.mobi.reader.index import read_index
 from ebook_converter.ebooks.mobi.reader.ncx import read_ncx, build_toc
@@ -17,7 +18,7 @@ from ebook_converter.ebooks.metadata.opf2 import Guide, OPFCreator
 from ebook_converter.ebooks.metadata.toc import TOC
 from ebook_converter.ebooks.mobi.utils import read_font_record
 from ebook_converter.ebooks.oeb.parse_utils import parse_html
-from ebook_converter.ebooks.oeb.base import XPath, XHTML, xml2text
+from ebook_converter.ebooks.oeb.base import XPath, xml2text
 from ebook_converter.polyglot.builtins import as_unicode
 
 
@@ -553,8 +554,8 @@ class Mobi8Reader(object):
         seen = set()
         links = []
         for elem in root.iterdescendants(etree.Element):
-            if reached and elem.tag == XHTML('a') and elem.get('href',
-                    False):
+            if reached and elem.tag == const.XHTML_A and elem.get('href',
+                                                                  False):
                 href = elem.get('href')
                 href, frag = urllib.parse.urldefrag(href)
                 href = base_href + '/' + href
diff --git a/ebook_converter/ebooks/mobi/writer2/serializer.py b/ebook_converter/ebooks/mobi/writer2/serializer.py
index c1f2587..6d0ae8d 100644
--- a/ebook_converter/ebooks/mobi/writer2/serializer.py
+++ b/ebook_converter/ebooks/mobi/writer2/serializer.py
@@ -4,16 +4,11 @@ import re
 import unicodedata
 import urllib.parse
 
+from ebook_converter import constants as const
 from ebook_converter.ebooks.mobi.mobiml import MBP_NS
 from ebook_converter.ebooks.mobi.utils import is_guide_ref_start
-from ebook_converter.ebooks.oeb.base import (
-    OEB_DOCS, XHTML, XHTML_NS, XML_NS, namespace, prefixname, urlnormalize
-)
-
-
-__license__ = 'GPL v3'
-__copyright__ = '2011, Kovid Goyal <kovid@kovidgoyal.net>'
-__docformat__ = 'restructuredtext en'
+from ebook_converter.ebooks.oeb import base
+from ebook_converter.ebooks.oeb import parse_utils
 
 
 class Buf(io.BytesIO):
@@ -25,9 +20,14 @@ class Buf(io.BytesIO):
 
 
 class Serializer(object):
-    NSRMAP = {'': None, XML_NS: 'xml', XHTML_NS: '', MBP_NS: 'mbp'}
+    NSRMAP = {'': None,
+              const.XML_NS: 'xml',
+              const.XHTML_NS: '',
+              MBP_NS: 'mbp'}  # TODO(gryf): check why this is different than
+                              # MBP_NS from const.
 
-    def __init__(self, oeb, images, is_periodical, write_page_breaks_after_item=True):
+    def __init__(self, oeb, images, is_periodical,
+                 write_page_breaks_after_item=True):
         '''
         Write all the HTML markup in oeb into a single in memory buffer
         containing a single html document with links replaced by offsets into
@@ -157,7 +157,8 @@ class Serializer(object):
         buf.write(b'<guide>')
         for ref in self.oeb.guide.values():
             path = urllib.parse.urldefrag(ref.href)[0]
-            if path not in hrefs or hrefs[path].media_type not in OEB_DOCS:
+            if (path not in hrefs or
+                    hrefs[path].media_type not in base.OEB_DOCS):
                 continue
 
             buf.write(b'<reference type="')
@@ -178,28 +179,28 @@ class Serializer(object):
 
         buf.write(b'</guide>')
 
-    def serialize_href(self, href, base=None):
-        '''
+    def serialize_href(self, href, _base=None):
+        """
         Serialize the href attribute of an <a> or <reference> tag. It is
         serialized as filepos="000000000" and a pointer to its location is
         stored in self.href_offsets so that the correct value can be filled in
         at the end.
-        '''
+        """
         hrefs = self.oeb.manifest.hrefs
         try:
-            path, frag = urllib.parse.urldefrag(urlnormalize(href))
+            path, frag = urllib.parse.urldefrag(base.urlnormalize(href))
         except ValueError:
             # Unparseable URL
             return False
-        if path and base:
-            path = base.abshref(path)
+        if path and _base:
+            path = _base.abshref(path)
         if path and path not in hrefs:
             return False
         buf = self.buf
         item = hrefs[path] if path else None
         if item and item.spine_position is None:
             return False
-        path = item.href if item else base.href
+        path = item.href if item else _base.href
         href = '#'.join((path, frag)) if frag else path
         buf.write(b'filepos=')
         self.href_offsets[href].append(buf.tell())
@@ -219,7 +220,7 @@ class Serializer(object):
             if href is not None:
                 # resolve the section url in id_offsets
                 buf.write(b'<mbp:pagebreak />')
-                self.id_offsets[urlnormalize(href)] = buf.tell()
+                self.id_offsets[base.urlnormalize(href)] = buf.tell()
 
             if tocref.klass == "periodical":
                 buf.write(b'<div> <div height="1em"></div>')
@@ -267,7 +268,7 @@ class Serializer(object):
 
             if self.is_periodical and item.is_section_start:
                 for section_toc in top_toc.nodes:
-                    if urlnormalize(item.href) == section_toc.href:
+                    if base.urlnormalize(item.href) == section_toc.href:
                         # create section url of the form r'feed_\d+/index.html'
                         section_url = re.sub(r'article_\d+/', '', section_toc.href)
                         serialize_toc_level(section_toc, section_url)
@@ -287,12 +288,12 @@ class Serializer(object):
         buf = self.buf
         if not item.linear:
             self.breaks.append(buf.tell() - 1)
-        self.id_offsets[urlnormalize(item.href)] = buf.tell()
+        self.id_offsets[base.urlnormalize(item.href)] = buf.tell()
         if item.is_section_start:
             buf.write(b'<a ></a> ')
         if item.is_article_start:
             buf.write(b'<a ></a> <a ></a>')
-        for elem in item.data.find(XHTML('body')):
+        for elem in item.data.find(base.tag('xhtml', 'body')):
             self.serialize_elem(elem, item)
         if self.write_page_breaks_after_item:
             buf.write(b'<mbp:pagebreak/>')
@@ -306,15 +307,15 @@ class Serializer(object):
     def serialize_elem(self, elem, item, nsrmap=NSRMAP):
         buf = self.buf
         if not isinstance(elem.tag, (str, bytes)) \
-            or namespace(elem.tag) not in nsrmap:
+            or parse_utils.namespace(elem.tag) not in nsrmap:
             return
-        tag = prefixname(elem.tag, nsrmap)
+        tag = base.prefixname(elem.tag, nsrmap)
         # Previous layers take care of @name
         id_ = elem.attrib.pop('id', None)
         if id_:
             href = '#'.join((item.href, id_))
             offset = self.anchor_offset or buf.tell()
-            key = urlnormalize(href)
+            key = base.urlnormalize(href)
             # Only set this id_offset if it wasn't previously seen
             self.id_offsets[key] = self.id_offsets.get(key, offset)
         if self.anchor_offset is not None and \
@@ -326,15 +327,15 @@ class Serializer(object):
         buf.write(tag.encode('utf-8'))
         if elem.attrib:
             for attr, val in elem.attrib.items():
-                if namespace(attr) not in nsrmap:
+                if parse_utils.namespace(attr) not in nsrmap:
                     continue
-                attr = prefixname(attr, nsrmap)
+                attr = base.prefixname(attr, nsrmap)
                 buf.write(b' ')
                 if attr == 'href':
                     if self.serialize_href(val, item):
                         continue
                 elif attr == 'src':
-                    href = urlnormalize(item.abshref(val))
+                    href = base.urlnormalize(item.abshref(val))
                     if href in self.images:
                         index = self.images[href]
                         self.used_images.add(href)
diff --git a/ebook_converter/ebooks/mobi/writer8/exth.py b/ebook_converter/ebooks/mobi/writer8/exth.py
index 435aa23..4c3ed6e 100644
--- a/ebook_converter/ebooks/mobi/writer8/exth.py
+++ b/ebook_converter/ebooks/mobi/writer8/exth.py
@@ -2,6 +2,7 @@ import re
 from struct import pack
 from io import BytesIO
 
+from ebook_converter.ebooks.oeb import base
 from ebook_converter.constants_old import iswindows, isosx
 from ebook_converter.ebooks.mobi.utils import (utf8_text, to_base)
 from ebook_converter.utils.localization import lang_as_iso639_1
@@ -95,9 +96,8 @@ def build_exth(metadata, prefer_author_sort=False, is_periodical=False,
 
     # Write UUID as ASIN
     uuid = None
-    from ebook_converter.ebooks.oeb.base import OPF
     for x in metadata['identifier']:
-        if (x.get(OPF('scheme'), None).lower() == 'uuid' or
+        if (x.get(base.tag('opf', 'scheme'), None).lower() == 'uuid' or
                 str(x).startswith('urn:uuid:')):
             uuid = str(x).split(':')[-1]
             break
diff --git a/ebook_converter/ebooks/odt/input.py b/ebook_converter/ebooks/odt/input.py
index ab3f5e2..ba6b7bf 100644
--- a/ebook_converter/ebooks/odt/input.py
+++ b/ebook_converter/ebooks/odt/input.py
@@ -1,11 +1,12 @@
 """
 Convert an ODT file into a Open Ebook
 """
-import os, logging
+import logging
+import os
 
-from lxml import etree
 from css_parser import CSSParser
 from css_parser.css import CSSRule
+from lxml import etree
 
 from odf.odf2xhtml import ODF2XHTML
 from odf.opendocument import load as odLoad
@@ -14,15 +15,9 @@ from odf.namespaces import TEXTNS as odTEXTNS
 
 from ebook_converter import CurrentDir, walk
 from ebook_converter.ebooks.oeb.base import _css_logger
-from ebook_converter.utils.xml_parse import safe_xml_fromstring
 from ebook_converter.polyglot.builtins import as_bytes
 
 
-__license__ = 'GPL v3'
-__copyright__ = '2008, Kovid Goyal kovid@kovidgoyal.net'
-__docformat__ = 'restructuredtext en'
-
-
 class Extract(ODF2XHTML):
 
     def extract_pictures(self, zf):
@@ -46,7 +41,7 @@ class Extract(ODF2XHTML):
                 ol.set('start', val)
 
     def fix_markup(self, html, log):
-        root = safe_xml_fromstring(html)
+        root = etree.fromstring(html)
         self.filter_css(root, log)
         self.extract_css(root, log)
         self.epubify_markup(root, log)
diff --git a/ebook_converter/ebooks/oeb/base.py b/ebook_converter/ebooks/oeb/base.py
index 2862fd9..769c87b 100644
--- a/ebook_converter/ebooks/oeb/base.py
+++ b/ebook_converter/ebooks/oeb/base.py
@@ -1,97 +1,45 @@
 """
 Basic support for manipulating OEB 1.x/2.0 content and metadata.
 """
-import os, re, logging, sys, numbers
-from collections import defaultdict
-from itertools import count
-from operator import attrgetter
-import urllib.parse
+import collections
+import itertools
+import logging
+import numbers
+import operator
+import os
+import re
 import string
+import sys
+import urllib.parse
 
-from lxml import etree, html
+from lxml import etree
+from lxml import html
+
+from ebook_converter import constants as const
 from ebook_converter import force_unicode
 from ebook_converter.constants_old import filesystem_encoding, __version__
-from ebook_converter.utils.xml_parse import safe_xml_fromstring
 from ebook_converter.ebooks.chardet import xml_to_unicode
 from ebook_converter.ebooks.conversion.preprocess import CSSPreProcessor
 from ebook_converter import (isbytestring, as_unicode, get_types_map)
-from ebook_converter.ebooks.oeb.parse_utils import barename, XHTML_NS, namespace, XHTML, parse_html, NotHTML
+from ebook_converter.ebooks.oeb import parse_utils
 from ebook_converter.utils.cleantext import clean_xml_chars
 from ebook_converter.utils.short_uuid import uuid4
 from ebook_converter.polyglot.urllib import unquote as urlunquote
 
 
-__license__ = 'GPL v3'
-__copyright__ = '2008, Marshall T. Vandegrift <llasram@gmail.com>'
-__docformat__ = 'restructuredtext en'
-
-XML_NS       = 'http://www.w3.org/XML/1998/namespace'
-OEB_DOC_NS   = 'http://openebook.org/namespaces/oeb-document/1.0/'
-OPF1_NS      = 'http://openebook.org/namespaces/oeb-package/1.0/'
-OPF2_NS      = 'http://www.idpf.org/2007/opf'
-OPF_NSES     = {OPF1_NS, OPF2_NS}
-DC09_NS      = 'http://purl.org/metadata/dublin_core'
-DC10_NS      = 'http://purl.org/dc/elements/1.0/'
-DC11_NS      = 'http://purl.org/dc/elements/1.1/'
-DC_NSES      = {DC09_NS, DC10_NS, DC11_NS}
-XSI_NS       = 'http://www.w3.org/2001/XMLSchema-instance'
-DCTERMS_NS   = 'http://purl.org/dc/terms/'
-NCX_NS       = 'http://www.daisy.org/z3986/2005/ncx/'
-SVG_NS       = 'http://www.w3.org/2000/svg'
-XLINK_NS     = 'http://www.w3.org/1999/xlink'
-CALIBRE_NS   = 'http://calibre.kovidgoyal.net/2009/metadata'
-RE_NS        = 'http://exslt.org/regular-expressions'
-MBP_NS       = 'http://www.mobipocket.com'
-EPUB_NS      = 'http://www.idpf.org/2007/ops'
-MATHML_NS    = 'http://www.w3.org/1998/Math/MathML'
-
-XPNSMAP      = {
-        'h': XHTML_NS, 'o1': OPF1_NS, 'o2': OPF2_NS, 'd09': DC09_NS,
-        'd10': DC10_NS, 'd11': DC11_NS, 'xsi': XSI_NS, 'dt': DCTERMS_NS,
-        'ncx': NCX_NS, 'svg': SVG_NS, 'xl': XLINK_NS, 're': RE_NS,
-        'mathml': MATHML_NS, 'mbp': MBP_NS, 'calibre': CALIBRE_NS,
-        'epub':EPUB_NS
-}
-
-OPF1_NSMAP   = {'dc': DC11_NS, 'oebpackage': OPF1_NS}
-OPF2_NSMAP   = {'opf': OPF2_NS, 'dc': DC11_NS, 'dcterms': DCTERMS_NS,
-                'xsi': XSI_NS, 'calibre': CALIBRE_NS}
-
-
-def XML(name):
-    return '{%s}%s' % (XML_NS, name)
-
-
-def OPF(name):
-    return '{%s}%s' % (OPF2_NS, name)
-
-
-def DC(name):
-    return '{%s}%s' % (DC11_NS, name)
-
-
-def XSI(name):
-    return '{%s}%s' % (XSI_NS, name)
-
-
-def DCTERMS(name):
-    return '{%s}%s' % (DCTERMS_NS, name)
-
-
-def NCX(name):
-    return '{%s}%s' % (NCX_NS, name)
-
-
-def SVG(name):
-    return '{%s}%s' % (SVG_NS, name)
-
-
-def XLINK(name):
-    return '{%s}%s' % (XLINK_NS, name)
-
-
-def CALIBRE(name):
-    return '{%s}%s' % (CALIBRE_NS, name)
+def tag(tag_ns, name):
+    tag_map = {'calibre': const.CALIBRE_NS,
+               'dc': const.DC11_NS,
+               'dcterms': const.DCTERMS_NS,
+               'epub': const.EPUB_NS,
+               'ncx': const.NCX_NS,
+               'opf': const.OPF2_NS,
+               'svg': const.SVG_NS,
+               'xhtml': const.XHTML_NS,
+               'xlink': const.XLINK_NS,
+               'xml': const.XML_NS,
+               'xsi': const.XSI_NS}
+    return '{%s}%s' % (tag_map[tag_ns], name)
 
 
 _css_url_re = re.compile(r'url\s*\([\'"]{0,1}(.*?)[\'"]{0,1}\)', re.I)
@@ -99,14 +47,18 @@ _css_import_re = re.compile(r'@import "(.*?)"')
 _archive_re = re.compile(r'[^ ]+')
 
 # Tags that should not be self closed in epub output
-self_closing_bad_tags = {'a', 'abbr', 'address', 'article', 'aside', 'audio', 'b',
-'bdo', 'blockquote', 'body', 'button', 'cite', 'code', 'dd', 'del', 'details',
-'dfn', 'div', 'dl', 'dt', 'em', 'fieldset', 'figcaption', 'figure', 'footer',
-'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'header', 'hgroup', 'i', 'iframe', 'ins', 'kbd',
-'label', 'legend', 'li', 'map', 'mark', 'meter', 'nav', 'ol', 'output', 'p',
-'pre', 'progress', 'q', 'rp', 'rt', 'samp', 'section', 'select', 'small',
-'span', 'strong', 'sub', 'summary', 'sup', 'textarea', 'time', 'ul', 'var',
-'video', 'title', 'script', 'style'}
+self_closing_bad_tags = {'a', 'abbr', 'address', 'article', 'aside', 'audio',
+                         'b', 'bdo', 'blockquote', 'body', 'button', 'cite',
+                         'code', 'dd', 'del', 'details', 'dfn', 'div', 'dl',
+                         'dt', 'em', 'fieldset', 'figcaption', 'figure',
+                         'footer', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6',
+                         'header', 'hgroup', 'i', 'iframe', 'ins', 'kbd',
+                         'label', 'legend', 'li', 'map', 'mark', 'meter',
+                         'nav', 'ol', 'output', 'p', 'pre', 'progress', 'q',
+                         'rp', 'rt', 'samp', 'section', 'select', 'small',
+                         'span', 'strong', 'sub', 'summary', 'sup',
+                         'textarea', 'time', 'ul', 'var', 'video', 'title',
+                         'script', 'style'}
 
 
 def css_text(x):
@@ -157,7 +109,7 @@ def itercsslinks(raw):
         yield match.group(1), match.start(1)
 
 
-_link_attrs = set(html.defs.link_attrs) | {XLINK('href'), 'poster'}
+_link_attrs = set(html.defs.link_attrs) | {tag('xlink', 'href'), 'poster'}
 
 
 def iterlinks(root, find_links_in_css=True):
@@ -170,7 +122,7 @@ def iterlinks(root, find_links_in_css=True):
 
     for el in root.iter('*'):
         try:
-            tag = barename(el.tag).lower()
+            tag = parse_utils.barename(el.tag).lower()
         except Exception:
             continue
         attribs = el.attrib
@@ -225,7 +177,7 @@ def make_links_absolute(root, base_url):
 def resolve_base_href(root):
     base_href = None
     basetags = root.xpath('//base[@href]|//h:base[@href]',
-            namespaces=XPNSMAP)
+                          namespaces=const.XPNSMAP)
     for b in basetags:
         base_href = b.get('href')
         b.drop_tree()
@@ -278,20 +230,20 @@ def rewrite_links(root, link_repl_func, resolve_base_href=False):
                 el.attrib[attrib] = new
 
     parser = CSSParser(raiseExceptions=False, log=_css_logger,
-            fetcher=lambda x:(None, None))
+                       fetcher=lambda x: (None, None))
     for el in root.iter(etree.Element):
         try:
             tag = el.tag
         except UnicodeDecodeError:
             continue
 
-        if tag == XHTML('style') and el.text and \
+        if tag == parse_utils.XHTML('style') and el.text and \
                 (_css_url_re.search(el.text) is not None or '@import' in
-                        el.text):
+                 el.text):
             stylesheet = parser.parseString(el.text, validate=False)
             replaceUrls(stylesheet, link_repl_func)
             repl = css_text(stylesheet)
-            el.text = '\n'+ clean_xml_chars(repl) + '\n'
+            el.text = '\n' + clean_xml_chars(repl) + '\n'
 
         text = el.get('style')
         if text and _css_url_re.search(text) is not None:
@@ -301,43 +253,41 @@ def rewrite_links(root, link_repl_func, resolve_base_href=False):
                 # Parsing errors are raised by css_parser
                 continue
             replaceUrls(stext, link_repl_func)
-            repl = css_text(stext).replace('\n', ' ').replace('\r',
-                    ' ')
+            repl = css_text(stext).replace('\n', ' ').replace('\r', ' ')
             el.set('style', repl)
 
 
 types_map = get_types_map()
-EPUB_MIME      = types_map['.epub']
-XHTML_MIME     = types_map['.xhtml']
-CSS_MIME       = types_map['.css']
-NCX_MIME       = types_map['.ncx']
-OPF_MIME       = types_map['.opf']
-PAGE_MAP_MIME  = 'application/oebps-page-map+xml'
-OEB_DOC_MIME   = 'text/x-oeb1-document'
-OEB_CSS_MIME   = 'text/x-oeb1-css'
-OPENTYPE_MIME  = types_map['.otf']
-GIF_MIME       = types_map['.gif']
-JPEG_MIME      = types_map['.jpeg']
-PNG_MIME       = types_map['.png']
-SVG_MIME       = types_map['.svg']
-BINARY_MIME    = 'application/octet-stream'
+EPUB_MIME = types_map['.epub']
+XHTML_MIME = types_map['.xhtml']
+CSS_MIME = types_map['.css']
+NCX_MIME = types_map['.ncx']
+OPF_MIME = types_map['.opf']
+PAGE_MAP_MIME = 'application/oebps-page-map+xml'
+OEB_DOC_MIME = 'text/x-oeb1-document'
+OEB_CSS_MIME = 'text/x-oeb1-css'
+OPENTYPE_MIME = types_map['.otf']
+GIF_MIME = types_map['.gif']
+JPEG_MIME = types_map['.jpeg']
+PNG_MIME = types_map['.png']
+SVG_MIME = types_map['.svg']
+BINARY_MIME = 'application/octet-stream'
 
-XHTML_CSS_NAMESPACE = '@namespace "%s";\n' % XHTML_NS
+XHTML_CSS_NAMESPACE = '@namespace "%s";\n' % const.XHTML_NS
 
-OEB_STYLES        = {CSS_MIME, OEB_CSS_MIME, 'text/x-oeb-css', 'xhtml/css'}
-OEB_DOCS          = {XHTML_MIME, 'text/html', OEB_DOC_MIME,
-                         'text/x-oeb-document'}
+OEB_STYLES = {CSS_MIME, OEB_CSS_MIME, 'text/x-oeb-css', 'xhtml/css'}
+OEB_DOCS = {XHTML_MIME, 'text/html', OEB_DOC_MIME, 'text/x-oeb-document'}
 OEB_RASTER_IMAGES = {GIF_MIME, JPEG_MIME, PNG_MIME}
-OEB_IMAGES        = {GIF_MIME, JPEG_MIME, PNG_MIME, SVG_MIME}
+OEB_IMAGES = {GIF_MIME, JPEG_MIME, PNG_MIME, SVG_MIME}
 
 MS_COVER_TYPE = 'other.ms-coverimage-standard'
 
-ENTITY_RE     = re.compile(r'&([a-zA-Z_:][a-zA-Z0-9.-_:]+);')
-COLLAPSE_RE   = re.compile(r'[ \t\r\n\v]+')
-QNAME_RE      = re.compile(r'^[{][^{}]+[}][^{}]+$')
+ENTITY_RE = re.compile(r'&([a-zA-Z_:][a-zA-Z0-9.-_:]+);')
+COLLAPSE_RE = re.compile(r'[ \t\r\n\v]+')
+QNAME_RE = re.compile(r'^[{][^{}]+[}][^{}]+$')
 PREFIXNAME_RE = re.compile(r'^[^:]+[:][^:]+')
-XMLDECL_RE    = re.compile(r'^\s*<[?]xml.*?[?]>')
-CSSURL_RE     = re.compile(r'''url[(](?P<q>["']?)(?P<url>[^)]+)(?P=q)[)]''')
+XMLDECL_RE = re.compile(r'^\s*<[?]xml.*?[?]>')
+CSSURL_RE = re.compile(r'''url[(](?P<q>["']?)(?P<url>[^)]+)(?P=q)[)]''')
 
 
 def element(parent, *args, **kwargs):
@@ -349,13 +299,13 @@ def element(parent, *args, **kwargs):
 def prefixname(name, nsrmap):
     if not isqname(name):
         return name
-    ns = namespace(name)
+    ns = parse_utils.namespace(name)
     if ns not in nsrmap:
         return name
     prefix = nsrmap[ns]
     if not prefix:
-        return barename(name)
-    return ':'.join((prefix, barename(name)))
+        return parse_utils.barename(name)
+    return ':'.join((prefix, parse_utils.barename(name)))
 
 
 def isprefixname(name):
@@ -376,11 +326,11 @@ def isqname(name):
 
 
 def XPath(expr):
-    return etree.XPath(expr, namespaces=XPNSMAP)
+    return etree.XPath(expr, namespaces=const.XPNSMAP)
 
 
 def xpath(elem, expr):
-    return elem.xpath(expr, namespaces=XPNSMAP)
+    return elem.xpath(expr, namespaces=const.XPNSMAP)
 
 
 def xml2str(root, pretty_print=False, strip_comments=False, with_tail=True):
@@ -390,7 +340,7 @@ def xml2str(root, pretty_print=False, strip_comments=False, with_tail=True):
             if x.text and '--' in x.text:
                 x.text = x.text.replace('--', '__')
     ans = etree.tostring(root, encoding='utf-8', xml_declaration=True,
-                          pretty_print=pretty_print, with_tail=with_tail)
+                         pretty_print=pretty_print, with_tail=with_tail)
 
     if strip_comments:
         ans = re.compile(br'<!--.*?-->', re.DOTALL).sub(b'', ans)
@@ -399,12 +349,14 @@ def xml2str(root, pretty_print=False, strip_comments=False, with_tail=True):
 
 
 def xml2text(elem, pretty_print=False, method='text'):
-    return etree.tostring(elem, method=method, encoding='unicode', with_tail=False, pretty_print=pretty_print)
+    return etree.tostring(elem, method=method, encoding='unicode',
+                          with_tail=False, pretty_print=pretty_print)
 
 
 def escape_cdata(root):
     pat = re.compile(r'[<>&]')
-    for elem in root.iterdescendants('{%s}style' % XHTML_NS, '{%s}script' % XHTML_NS):
+    for elem in root.iterdescendants('{%s}style' % const.XHTML_NS,
+                                     '{%s}script' % const.XHTML_NS):
         if elem.text and pat.search(elem.text) is not None:
             elem.text = etree.CDATA(elem.text.replace(']]>', r'\]\]\>'))
 
@@ -431,12 +383,12 @@ def serialize(data, media_type, pretty_print=False):
     return bytes(data)
 
 
-ASCII_CHARS   = frozenset(chr(x) for x in range(128))
+ASCII_CHARS = frozenset(chr(x) for x in range(128))
 UNIBYTE_CHARS = frozenset(x.encode('ascii') for x in ASCII_CHARS)
-USAFE         = ('ABCDEFGHIJKLMNOPQRSTUVWXYZ'
-                 'abcdefghijklmnopqrstuvwxyz'
-                 '0123456789' '_.-/~')
-URL_SAFE      = frozenset(USAFE)
+USAFE = ('ABCDEFGHIJKLMNOPQRSTUVWXYZ'
+         'abcdefghijklmnopqrstuvwxyz'
+         '0123456789' '_.-/~')
+URL_SAFE = frozenset(USAFE)
 URL_SAFE_BYTES = frozenset(USAFE.encode('ascii'))
 URL_UNSAFE = [ASCII_CHARS - URL_SAFE, UNIBYTE_CHARS - URL_SAFE_BYTES]
 del USAFE
@@ -466,7 +418,8 @@ def urlnormalize(href):
     try:
         parts = urllib.parse.urlparse(href)
     except ValueError as e:
-        raise ValueError('Failed to parse the URL: %r with underlying error: %s' % (href, as_unicode(e)))
+        raise ValueError('Failed to parse the URL: %r with underlying error: '
+                         '%s' % (href, as_unicode(e)))
     if not parts.scheme or parts.scheme == 'file':
         path, frag = urllib.parse.urldefrag(href)
         parts = ('', '', path, '', '', frag)
@@ -503,8 +456,10 @@ class DummyHandler(logging.Handler):
     def emit(self, record):
         if self.log is not None:
             msg = self.format(record)
-            f = self.log.error if record.levelno >= logging.ERROR \
-                    else self.log.warn
+            if record.levelno >= logging.ERROR:
+                f = self.log.error
+            else:
+                self.log.warn
             f(msg)
 
 
@@ -630,18 +585,21 @@ class Metadata(object):
     metadata items.
     """
 
-    DC_TERMS      = {'contributor', 'coverage', 'creator', 'date',
-                         'description', 'format', 'identifier', 'language',
-                         'publisher', 'relation', 'rights', 'source',
-                         'subject', 'title', 'type'}
+    DC_TERMS = {'contributor', 'coverage', 'creator', 'date', 'description',
+                'format', 'identifier', 'language', 'publisher', 'relation',
+                'rights', 'source', 'subject', 'title', 'type'}
     CALIBRE_TERMS = {'series', 'series_index', 'rating', 'timestamp',
-                         'publication_type', 'title_sort'}
-    OPF_ATTRS     = {'role': OPF('role'), 'file-as': OPF('file-as'),
-                     'scheme': OPF('scheme'), 'event': OPF('event'),
-                     'type': XSI('type'), 'lang': XML('lang'), 'id': 'id'}
-    OPF1_NSMAP    = {'dc': DC11_NS, 'oebpackage': OPF1_NS}
-    OPF2_NSMAP    = {'opf': OPF2_NS, 'dc': DC11_NS, 'dcterms': DCTERMS_NS,
-                     'xsi': XSI_NS, 'calibre': CALIBRE_NS}
+                     'publication_type', 'title_sort'}
+    OPF_ATTRS = {'role': tag('opf', 'role'), 'file-as': tag('opf', 'file-as'),
+                 'scheme': tag('opf', 'scheme'), 'event': tag('opf', 'event'),
+                 'type': tag('xsi', 'type'), 'lang': tag('xml', 'lang'),
+                 'id': 'id'}
+    OPF1_NSMAP = {'dc': const.DC11_NS, 'oebpackage': const.OPF1_NS}
+    OPF2_NSMAP = {'calibre': const.CALIBRE_NS,
+                  'dc': const.DC11_NS,
+                  'dcterms': const.DCTERMS_NS,
+                  'opf': const.OPF2_NS,
+                  'xsi': const.XSI_NS}
 
     class Item(object):
         """An item of OEB data model metadata.
@@ -667,13 +625,13 @@ class Metadata(object):
 
             def term_attr(self, obj):
                 term = obj.term
-                if namespace(term) != DC11_NS:
-                    term = OPF('meta')
+                if parse_utils.namespace(term) != const.DC11_NS:
+                    term = tag('opf', 'meta')
                 allowed = self.allowed
                 if allowed is not None and term not in allowed:
                     raise AttributeError(
                         'attribute %r not valid for metadata term %r' % (
-                            self.attr(term), barename(obj.term)))
+                            self.attr(term), parse_utils.barename(obj.term)))
                 return self.attr(term)
 
             def __get__(self, obj, cls):
@@ -688,23 +646,25 @@ class Metadata(object):
             self.attrib = attrib = dict(attrib)
             self.nsmap = nsmap = dict(nsmap)
             attrib.update(kwargs)
-            if namespace(term) == OPF2_NS:
-                term = barename(term)
-            ns = namespace(term)
-            local = barename(term).lower()
-            if local in Metadata.DC_TERMS and (not ns or ns in DC_NSES):
+            if parse_utils.namespace(term) == const.OPF2_NS:
+                term = parse_utils.barename(term)
+            ns = parse_utils.namespace(term)
+            local = parse_utils.barename(term).lower()
+            if local in Metadata.DC_TERMS and (not ns or ns in const.DC_NSES):
                 # Anything looking like Dublin Core is coerced
-                term = DC(local)
-            elif local in Metadata.CALIBRE_TERMS and ns in (CALIBRE_NS, ''):
+                term = tag('dc', local)
+            elif local in Metadata.CALIBRE_TERMS and ns in (const.CALIBRE_NS,
+                                                            ''):
                 # Ditto for Calibre-specific metadata
-                term = CALIBRE(local)
+                term = tag('calibre', local)
             self.term = term
             self.value = value
             for attr, value in tuple(attrib.items()):
                 if isprefixname(value):
                     attrib[attr] = qname(value, nsmap)
                 nsattr = Metadata.OPF_ATTRS.get(attr, attr)
-                if nsattr == OPF('scheme') and namespace(term) != DC11_NS:
+                if (nsattr == tag('opf', 'scheme') and
+                        parse_utils.namespace(term) != const.DC11_NS):
                     # The opf:meta element takes @scheme, not @opf:scheme
                     nsattr = 'scheme'
                 if attr != nsattr:
@@ -722,21 +682,29 @@ class Metadata(object):
         def content(self, value):
             self.value = value
 
-        scheme  = Attribute(lambda term: 'scheme' if
-                            term == OPF('meta') else OPF('scheme'),
-                            [DC('identifier'), OPF('meta')])
-        file_as = Attribute(OPF('file-as'), [DC('creator'), DC('contributor'),
-                                             DC('title')])
-        role    = Attribute(OPF('role'), [DC('creator'), DC('contributor')])
-        event   = Attribute(OPF('event'), [DC('date')])
-        id      = Attribute('id')
-        type    = Attribute(XSI('type'), [DC('date'), DC('format'),
-                                          DC('type')])
-        lang    = Attribute(XML('lang'), [DC('contributor'), DC('coverage'),
-                                          DC('creator'), DC('publisher'),
-                                          DC('relation'), DC('rights'),
-                                          DC('source'), DC('subject'),
-                                          OPF('meta')])
+        scheme = Attribute(lambda term: 'scheme' if
+                           term == tag('opf', 'meta') else
+                           tag('opf', 'scheme'),
+                           [tag('dc', 'identifier'), tag('opf', 'meta')])
+        file_as = Attribute(tag('opf', 'file-as'), [tag('dc', 'creator'),
+                                                    tag('dc', 'contributor'),
+                                                    tag('dc', 'title')])
+        role = Attribute(tag('opf', 'role'), [tag('dc', 'creator'),
+                                              tag('dc', 'contributor')])
+        event = Attribute(tag('opf', 'event'), [tag('dc', 'date')])
+        id = Attribute('id')
+        type = Attribute(tag('xsi', 'type'), [tag('dc', 'date'),
+                                              tag('dc', 'format'),
+                                              tag('dc', 'type')])
+        lang = Attribute(tag('xml', 'lang'), [tag('dc', 'contributor'),
+                                              tag('dc', 'coverage'),
+                                              tag('dc', 'creator'),
+                                              tag('dc', 'publisher'),
+                                              tag('dc', 'relation'),
+                                              tag('dc', 'rights'),
+                                              tag('dc', 'source'),
+                                              tag('dc', 'subject'),
+                                              tag('opf', 'meta')])
 
         def __getitem__(self, key):
             return self.attrib[key]
@@ -752,7 +720,7 @@ class Metadata(object):
 
         def __repr__(self):
             return 'Item(term=%r, value=%r, attrib=%r)' \
-                % (barename(self.term), self.value, self.attrib)
+                % (parse_utils.barename(self.term), self.value, self.attrib)
 
         def __str__(self):
             return as_unicode(self.value)
@@ -760,11 +728,12 @@ class Metadata(object):
         def to_opf1(self, dcmeta=None, xmeta=None, nsrmap={}):
             attrib = {}
             for key, value in self.attrib.items():
-                if namespace(key) == OPF2_NS:
-                    key = barename(key)
+                if parse_utils.namespace(key) == const.OPF2_NS:
+                    key = parse_utils.barename(key)
                 attrib[key] = prefixname(value, nsrmap)
-            if namespace(self.term) == DC11_NS:
-                name = DC(string.capwords(barename(self.term)))
+            if parse_utils.namespace(self.term) == const.DC11_NS:
+                name = tag('dc',
+                           string.capwords(parse_utils.barename(self.term)))
                 elem = element(dcmeta, name, attrib=attrib)
                 elem.text = self.value
             else:
@@ -777,27 +746,27 @@ class Metadata(object):
             attrib = {}
             for key, value in self.attrib.items():
                 attrib[key] = prefixname(value, nsrmap)
-            if namespace(self.term) == DC11_NS:
+            if parse_utils.namespace(self.term) == const.DC11_NS:
                 elem = element(parent, self.term, attrib=attrib)
                 try:
                     elem.text = self.value
-                except:
+                except Exception:
                     elem.text = repr(self.value)
             else:
-                elem = element(parent, OPF('meta'), attrib=attrib)
+                elem = element(parent, tag('opf', 'meta'), attrib=attrib)
                 elem.attrib['name'] = prefixname(self.term, nsrmap)
                 elem.attrib['content'] = prefixname(self.value, nsrmap)
             return elem
 
     def __init__(self, oeb):
         self.oeb = oeb
-        self.items = defaultdict(list)
+        self.items = collections.defaultdict(list)
         self.primary_writing_mode = None
 
     def add(self, term, value, attrib={}, nsmap={}, **kwargs):
         """Add a new metadata item."""
         item = self.Item(term, value, attrib, nsmap, **kwargs)
-        items = self.items[barename(item.term)]
+        items = self.items[parse_utils.barename(item.term)]
         items.append(item)
         return item
 
@@ -807,15 +776,15 @@ class Metadata(object):
     __iter__ = iterkeys
 
     def clear(self, key):
-        l = self.items[key]
-        for x in list(l):
-            l.remove(x)
+        val = self.items[key]
+        for x in list(val):
+            val.remove(x)
 
     def filter(self, key, predicate):
-        l = self.items[key]
-        for x in list(l):
+        val = self.items[key]
+        for x in list(val):
             if predicate(x):
-                l.remove(x)
+                val.remove(x)
 
     def __getitem__(self, key):
         return self.items[key]
@@ -838,21 +807,21 @@ class Metadata(object):
     def _opf1_nsmap(self):
         nsmap = self._nsmap
         for key, value in nsmap.items():
-            if value in OPF_NSES or value in DC_NSES:
+            if value in const.OPF_NSES or value in const.DC_NSES:
                 del nsmap[key]
         return nsmap
 
     @property
     def _opf2_nsmap(self):
         nsmap = self._nsmap
-        nsmap.update(OPF2_NSMAP)
+        nsmap.update(const.OPF2_NSMAP)
         return nsmap
 
     def to_opf1(self, parent=None):
         nsmap = self._opf1_nsmap
         nsrmap = {value: key for key, value in nsmap.items()}
         elem = element(parent, 'metadata', nsmap=nsmap)
-        dcmeta = element(elem, 'dc-metadata', nsmap=OPF1_NSMAP)
+        dcmeta = element(elem, 'dc-metadata', nsmap=const.OPF1_NSMAP)
         xmeta = element(elem, 'x-metadata')
         for term in self.items:
             for item in self.items[term]:
@@ -865,12 +834,16 @@ class Metadata(object):
     def to_opf2(self, parent=None):
         nsmap = self._opf2_nsmap
         nsrmap = {value: key for key, value in nsmap.items()}
-        elem = element(parent, OPF('metadata'), nsmap=nsmap)
+        elem = element(parent, tag('opf', 'metadata'), nsmap=nsmap)
         for term in self.items:
             for item in self.items[term]:
                 item.to_opf2(elem, nsrmap=nsrmap)
         if self.primary_writing_mode:
-            elem.append(elem.makeelement(OPF('meta'), attrib={'name':'primary-writing-mode', 'content':self.primary_writing_mode}))
+            elem.append(elem.makeelement(tag('opf', 'meta'),
+                                         attrib={'name':
+                                                 'primary-writing-mode',
+                                                 'content':
+                                                 self.primary_writing_mode}))
         return elem
 
 
@@ -936,8 +909,8 @@ class Manifest(object):
             if not data:
                 return
             data = xml_to_unicode(data, strip_encoding_pats=True,
-                    assume_utf8=True, resolve_entities=True)[0]
-            return safe_xml_fromstring(data)
+                                  assume_utf8=True, resolve_entities=True)[0]
+            return etree.fromstring(data)
 
         def _parse_xhtml(self, data):
             orig_data = data
@@ -945,11 +918,13 @@ class Manifest(object):
             self.oeb.log.debug('Parsing', fname, '...')
             self.oeb.html_preprocessor.current_href = self.href
             try:
-                data = parse_html(data, log=self.oeb.log,
-                        decoder=self.oeb.decode,
-                        preprocessor=self.oeb.html_preprocessor,
-                        filename=fname, non_html_file_tags={'ncx'})
-            except NotHTML:
+                data = parse_utils.parse_html(data, log=self.oeb.log,
+                                              decoder=self.oeb.decode,
+                                              preprocessor=self.oeb.
+                                              html_preprocessor,
+                                              filename=fname,
+                                              non_html_file_tags={'ncx'})
+            except parse_utils.NotHTML:
                 return self._parse_xml(orig_data)
             return data
 
@@ -981,7 +956,8 @@ class Manifest(object):
             data = self.oeb.decode(data)
             data = self.oeb.css_preprocessor(data, add_namespace=False)
             parser = CSSParser(loglevel=logging.WARNING,
-                               fetcher=self.override_css_fetch or self._fetch_css,
+                               fetcher=self.override_css_fetch or
+                               self._fetch_css,
                                log=_css_logger)
             data = parser.parseString(data, href=self.href, validate=False)
             data = resolveImports(data)
@@ -1013,8 +989,8 @@ class Manifest(object):
               convert and return as an lxml.etree element in the XHTML
               namespace.
             - XML content is parsed and returned as an lxml.etree element.
-            - CSS and CSS-variant content is parsed and returned as a css_parser
-              CSS DOM stylesheet.
+            - CSS and CSS-variant content is parsed and returned as a
+              css_parser CSS DOM stylesheet.
             - All other content is returned as a :class:`str` or :class:`bytes`
               object with no special parsing.
             """
@@ -1023,7 +999,7 @@ class Manifest(object):
                 if self._loader is None:
                     return None
                 data = self._loader(getattr(self, 'html_input_href',
-                    self.href))
+                                            self.href))
             try:
                 mt = self.media_type.lower()
             except Exception:
@@ -1037,8 +1013,8 @@ class Manifest(object):
             elif mt in OEB_STYLES:
                 data = self._parse_css(data)
             elif mt == 'text/plain':
-                self.oeb.log.warn('%s contains data in TXT format'%self.href,
-                        'converting to HTML')
+                self.oeb.log.warn('%s contains data in TXT format' % self.href,
+                                  'converting to HTML')
                 data = self._parse_txt(data)
                 self.media_type = XHTML_MIME
             self._data = data
@@ -1055,8 +1031,10 @@ class Manifest(object):
         def unload_data_from_memory(self, memory=None):
             if isinstance(self._data, bytes):
                 if memory is None:
-                    from ebook_converter.ptempfile import PersistentTemporaryFile
-                    pt = PersistentTemporaryFile(suffix='_oeb_base_mem_unloader.img')
+                    from ebook_converter.ptempfile import \
+                            PersistentTemporaryFile
+                    pt = PersistentTemporaryFile(suffix='_oeb_base_mem_'
+                                                 'unloader.img')
                     with pt:
                         pt.write(self._data)
                     self.oeb._temp_files.append(pt.name)
@@ -1088,7 +1066,8 @@ class Manifest(object):
 
         @property
         def bytes_representation(self):
-            return serialize(self.data, self.media_type, pretty_print=self.oeb.pretty_print)
+            return serialize(self.data, self.media_type,
+                             pretty_print=self.oeb.pretty_print)
 
         def __str__(self):
             return self.unicode_representation
@@ -1107,7 +1086,11 @@ class Manifest(object):
             href = self.href
             if isinstance(href, bytes):
                 href = force_unicode(href)
-            sp = self.spine_position if isinstance(self.spine_position, numbers.Number) else sys.maxsize
+
+            if isinstance(self.spine_position, numbers.Number):
+                sp = self.spine_position
+            else:
+                sp = sys.maxsize
 
             return sp, (self.media_type or '').lower(), href, self.id
 
@@ -1238,8 +1221,8 @@ class Manifest(object):
         return elem
 
     def to_opf2(self, parent=None):
-        elem = element(parent, OPF('manifest'))
-        for item in sorted(self.items, key=attrgetter('sort_key')):
+        elem = element(parent, tag('opf', 'manifest'))
+        for item in sorted(self.items, key=operator.attrgetter('sort_key')):
             media_type = item.media_type
             if media_type in OEB_DOCS:
                 media_type = XHTML_MIME
@@ -1249,7 +1232,7 @@ class Manifest(object):
                       'media-type': media_type}
             if item.fallback:
                 attrib['fallback'] = item.fallback
-            element(elem, OPF('item'), attrib=attrib)
+            element(elem, tag('opf', 'item'), attrib=attrib)
         return elem
 
     @property
@@ -1341,12 +1324,12 @@ class Spine(object):
         return elem
 
     def to_opf2(self, parent=None):
-        elem = element(parent, OPF('spine'))
+        elem = element(parent, tag('opf', 'spine'))
         for item in self.items:
             attrib = {'idref': item.id}
             if not item.linear:
                 attrib['linear'] = 'no'
-            element(elem, OPF('itemref'), attrib=attrib)
+            element(elem, tag('opf', 'itemref'), attrib=attrib)
         return elem
 
 
@@ -1394,8 +1377,7 @@ class Guide(object):
             self.oeb = oeb
             if type.lower() in self.TYPES:
                 type = type.lower()
-            elif type not in self.TYPES and \
-                 not type.startswith('other.'):
+            elif type not in self.TYPES and not type.startswith('other.'):
                 type = 'other.' + type
             if not title and type in self.TITLES:
                 title = oeb.translate(self.TITLES[type])
@@ -1440,7 +1422,8 @@ class Guide(object):
     __iter__ = iterkeys
 
     def values(self):
-        return sorted(self.refs.values(), key=lambda ref: ref.ORDER.get(ref.type, 10000))
+        return sorted(self.refs.values(),
+                      key=lambda ref: ref.ORDER.get(ref.type, 10000))
 
     def items(self):
         for type, ref in self.refs.items():
@@ -1473,12 +1456,12 @@ class Guide(object):
     def to_opf2(self, parent=None):
         if not len(self):
             return
-        elem = element(parent, OPF('guide'))
+        elem = element(parent, tag('opf', 'guide'))
         for ref in self.refs.values():
             attrib = {'type': ref.type, 'href': urlunquote(ref.href)}
             if ref.title:
                 attrib['title'] = ref.title
-            element(elem, OPF('reference'), attrib=attrib)
+            element(elem, tag('opf', 'reference'), attrib=attrib)
         return elem
 
 
@@ -1499,7 +1482,8 @@ class TOC(object):
     """
 
     def __init__(self, title=None, href=None, klass=None, id=None,
-            play_order=None, author=None, description=None, toc_thumbnail=None):
+                 play_order=None, author=None, description=None,
+                 toc_thumbnail=None):
         self.title = title
         self.href = urlnormalize(href) if href else href
         self.klass = klass
@@ -1513,9 +1497,11 @@ class TOC(object):
         self.description = description
         self.toc_thumbnail = toc_thumbnail
 
-    def add(self, title, href, klass=None, id=None, play_order=0, author=None, description=None, toc_thumbnail=None):
+    def add(self, title, href, klass=None, id=None, play_order=0, author=None,
+            description=None, toc_thumbnail=None):
         """Create and return a new sub-node of this node."""
-        node = TOC(title, href, klass, id, play_order, author, description, toc_thumbnail)
+        node = TOC(title, href, klass, id, play_order, author, description,
+                   toc_thumbnail)
         self.nodes.append(node)
         return node
 
@@ -1583,7 +1569,8 @@ class TOC(object):
         """
         prev = None
         for node in list(self.nodes):
-            if prev and urllib.parse.urldefrag(prev.href)[0] == urllib.parse.urldefrag(node.href)[0]:
+            if (prev and urllib.parse.urldefrag(prev.href)[0] ==
+                    urllib.parse.urldefrag(node.href)[0]):
                 self.nodes.remove(node)
                 prev.nodes.append(node)
             else:
@@ -1597,7 +1584,7 @@ class TOC(object):
             return 1
 
     def get_lines(self, lvl=0):
-        ans = [('\t'*lvl) + 'TOC: %s --> %s'%(self.title, self.href)]
+        ans = [('\t'*lvl) + 'TOC: %s --> %s' % (self.title, self.href)]
         for child in self:
             ans.extend(child.get_lines(lvl+1))
         return ans
@@ -1614,7 +1601,7 @@ class TOC(object):
 
     def to_ncx(self, parent=None):
         if parent is None:
-            parent = etree.Element(NCX('navMap'))
+            parent = etree.Element(tag('ncx', 'navMap'))
         for node in self.nodes:
             id = node.id or uuid_id()
             po = node.play_order
@@ -1623,15 +1610,15 @@ class TOC(object):
             attrib = {'id': id, 'playOrder': str(po)}
             if node.klass:
                 attrib['class'] = node.klass
-            point = element(parent, NCX('navPoint'), attrib=attrib)
-            label = etree.SubElement(point, NCX('navLabel'))
+            point = element(parent, tag('ncx', 'navPoint'), attrib=attrib)
+            label = etree.SubElement(point, tag('ncx', 'navLabel'))
             title = node.title
             if title:
                 title = re.sub(r'\s+', ' ', title)
-            element(label, NCX('text')).text = title
+            element(label, tag('ncx', 'text')).text = title
             # Do not unescape this URL as ADE requires it to be escaped to
             # handle semi colons and other special characters in the file names
-            element(point, NCX('content'), src=node.href)
+            element(point, tag('ncx', 'content'), src=node.href)
             node.to_ncx(point)
         return parent
 
@@ -1659,7 +1646,7 @@ class TOC(object):
             if y is not None:
                 if x.href != y.href:
                     x.play_order = getattr(href_node(x), 'play_order',
-                            self.next_play_order())
+                                           self.next_play_order())
             y = href_node(x)
             if y is not None:
                 x.play_order = y.play_order
@@ -1723,8 +1710,9 @@ class PageList(object):
         return self.pages.remove(page)
 
     def to_ncx(self, parent=None):
-        plist = element(parent, NCX('pageList'), id=uuid_id())
-        values = {t: count(1) for t in ('front', 'normal', 'special')}
+        plist = element(parent, tag('ncx', 'pageList'), id=uuid_id())
+        values = {t: itertools.count(1)
+                  for t in ('front', 'normal', 'special')}
         for page in self.pages:
             id = page.id or uuid_id()
             type = page.type
@@ -1732,30 +1720,31 @@ class PageList(object):
             attrib = {'id': id, 'value': value, 'type': type, 'playOrder': '0'}
             if page.klass:
                 attrib['class'] = page.klass
-            ptarget = element(plist, NCX('pageTarget'), attrib=attrib)
-            label = element(ptarget, NCX('navLabel'))
-            element(label, NCX('text')).text = page.name
-            element(ptarget, NCX('content'), src=page.href)
+            ptarget = element(plist, tag('ncx', 'pageTarget'), attrib=attrib)
+            label = element(ptarget, tag('ncx', 'navLabel'))
+            element(label, tag('ncx', 'text')).text = page.name
+            element(ptarget, tag('ncx', 'content'), src=page.href)
         return plist
 
     def to_page_map(self):
-        pmap = etree.Element(OPF('page-map'), nsmap={None: OPF2_NS})
+        pmap = etree.Element(tag('opf', 'page-map'),
+                             nsmap={None: const.OPF2_NS})
         for page in self.pages:
-            element(pmap, OPF('page'), name=page.name, href=page.href)
+            element(pmap, tag('opf', 'page'), name=page.name, href=page.href)
         return pmap
 
 
 class OEBBook(object):
     """Representation of a book in the IDPF OEB data model."""
 
-    COVER_SVG_XP    = XPath('h:body//svg:svg[position() = 1]')
+    COVER_SVG_XP = XPath('h:body//svg:svg[position() = 1]')
     COVER_OBJECT_XP = XPath('h:body//h:object[@data][position() = 1]')
 
     def __init__(self, logger,
-            html_preprocessor,
-            css_preprocessor=CSSPreProcessor(),
-            encoding='utf-8', pretty_print=False,
-            input_encoding='utf-8'):
+                 html_preprocessor,
+                 css_preprocessor=CSSPreProcessor(),
+                 encoding='utf-8', pretty_print=False,
+                 input_encoding='utf-8'):
         """Create empty book.  Arguments:
 
         :param:`encoding`: Default encoding for textual content read
@@ -1809,7 +1798,7 @@ class OEBBook(object):
         for path in self._temp_files:
             try:
                 os.remove(path)
-            except:
+            except Exception:
                 pass
 
     @classmethod
@@ -1831,11 +1820,12 @@ class OEBBook(object):
             return fix_data(data)
         bom_enc = None
         if data[:4] in (b'\0\0\xfe\xff', b'\xff\xfe\0\0'):
-            bom_enc = {b'\0\0\xfe\xff':'utf-32-be',
-                    b'\xff\xfe\0\0':'utf-32-le'}[data[:4]]
+            bom_enc = {b'\0\0\xfe\xff': 'utf-32-be',
+                       b'\xff\xfe\0\0': 'utf-32-le'}[data[:4]]
             data = data[4:]
         elif data[:2] in (b'\xff\xfe', b'\xfe\xff'):
-            bom_enc = {b'\xff\xfe':'utf-16-le', 'b\xfe\xff':'utf-16-be'}[data[:2]]
+            bom_enc = {b'\xff\xfe': 'utf-16-le',
+                       'b\xfe\xff': 'utf-16-be'}[data[:2]]
             data = data[2:]
         elif data[:3] == b'\xef\xbb\xbf':
             bom_enc = 'utf-8'
@@ -1864,13 +1854,13 @@ class OEBBook(object):
         are tuples of (default) filenames and lxml.etree element structures.
         """
         package = etree.Element('package',
-            attrib={'unique-identifier': self.uid.id})
+                                attrib={'unique-identifier': self.uid.id})
         self.metadata.to_opf1(package)
         self.manifest.to_opf1(package)
         self.spine.to_opf1(package)
         tours = element(package, 'tours')
         tour = element(tours, 'tour',
-            attrib={'id': 'chaptertour', 'title': 'Chapter Tour'})
+                       attrib={'id': 'chaptertour', 'title': 'Chapter Tour'})
         self.toc.to_opf1(tour)
         self.guide.to_opf1(package)
         return {OPF_MIME: ('content.opf', package)}
@@ -1898,7 +1888,8 @@ class OEBBook(object):
                 if added:
                     next += 1
         selector = XPath('ncx:content/@src')
-        for i, elem in enumerate(xpath(ncx, '//*[@playOrder and ./ncx:content[@src]]')):
+        for i, elem in enumerate(xpath(ncx, '//*[@playOrder and '
+                                       './ncx:content[@src]]')):
             href = urlnormalize(selector(elem)[0])
             order = playorder.get(href, i)
             elem.attrib['playOrder'] = str(order)
@@ -1907,25 +1898,26 @@ class OEBBook(object):
     def _to_ncx(self):
         lang = str(self.metadata.language[0])
         lang = lang.replace('_', '-')
-        ncx = etree.Element(NCX('ncx'),
-            attrib={'version': '2005-1', XML('lang'): lang},
-            nsmap={None: NCX_NS})
-        head = etree.SubElement(ncx, NCX('head'))
-        etree.SubElement(head, NCX('meta'),
-            name='dtb:uid', content=str(self.uid))
-        etree.SubElement(head, NCX('meta'),
-            name='dtb:depth', content=str(self.toc.depth()))
+        ncx = etree.Element(tag('ncx', 'ncx'),
+                            attrib={'version': '2005-1',
+                                    tag('xml', 'lang'): lang},
+                            nsmap={None: const.NCX_NS})
+        head = etree.SubElement(ncx, tag('ncx', 'head'))
+        etree.SubElement(head, tag('ncx', 'meta'),
+                         name='dtb:uid', content=str(self.uid))
+        etree.SubElement(head, tag('ncx', 'meta'),
+                         name='dtb:depth', content=str(self.toc.depth()))
         generator = ''.join(['calibre (', __version__, ')'])
-        etree.SubElement(head, NCX('meta'),
-            name='dtb:generator', content=generator)
-        etree.SubElement(head, NCX('meta'),
-            name='dtb:totalPageCount', content=str(len(self.pages)))
-        maxpnum = etree.SubElement(head, NCX('meta'),
-            name='dtb:maxPageNumber', content='0')
-        title = etree.SubElement(ncx, NCX('docTitle'))
-        text = etree.SubElement(title, NCX('text'))
+        etree.SubElement(head, tag('ncx', 'meta'),
+                         name='dtb:generator', content=generator)
+        etree.SubElement(head, tag('ncx', 'meta'), name='dtb:totalPageCount',
+                         content=str(len(self.pages)))
+        maxpnum = etree.SubElement(head, tag('ncx', 'meta'),
+                                   name='dtb:maxPageNumber', content='0')
+        title = etree.SubElement(ncx, tag('ncx', 'docTitle'))
+        text = etree.SubElement(title, tag('ncx', 'text'))
         text.text = str(self.metadata.title[0])
-        navmap = etree.SubElement(ncx, NCX('navMap'))
+        navmap = etree.SubElement(ncx, tag('ncx', 'navMap'))
         self.toc.to_ncx(navmap)
         if len(self.pages) > 0:
             plist = self.pages.to_ncx(ncx)
@@ -1941,27 +1933,29 @@ class OEBBook(object):
         are tuples of (default) filenames and lxml.etree element structures.
         """
         results = {}
-        package = etree.Element(OPF('package'),
-            attrib={'version': '2.0', 'unique-identifier': self.uid.id},
-            nsmap={None: OPF2_NS})
+        package = etree.Element(tag('opf', 'package'),
+                                attrib={'version': '2.0',
+                                        'unique-identifier': self.uid.id},
+                                nsmap={None: const.OPF2_NS})
         self.metadata.to_opf2(package)
         manifest = self.manifest.to_opf2(package)
         spine = self.spine.to_opf2(package)
         self.guide.to_opf2(package)
         results[OPF_MIME] = ('content.opf', package)
         id, href = self.manifest.generate('ncx', 'toc.ncx')
-        etree.SubElement(manifest, OPF('item'), id=id, href=href,
+        etree.SubElement(manifest, tag('opf', 'item'), id=id, href=href,
                          attrib={'media-type': NCX_MIME})
         spine.attrib['toc'] = id
         results[NCX_MIME] = (href, self._to_ncx())
         if page_map and len(self.pages) > 0:
             id, href = self.manifest.generate('page-map', 'page-map.xml')
-            etree.SubElement(manifest, OPF('item'), id=id, href=href,
+            etree.SubElement(manifest, tag('opf', 'item'), id=id, href=href,
                              attrib={'media-type': PAGE_MAP_MIME})
             spine.attrib['page-map'] = id
             results[PAGE_MAP_MIME] = (href, self.pages.to_page_map())
         if self.spine.page_progression_direction in {'ltr', 'rtl'}:
-            spine.attrib['page-progression-direction'] = self.spine.page_progression_direction
+            spine.attrib['page-progression-direction'] = \
+                    self.spine.page_progression_direction
         return results
 
 
@@ -1972,7 +1966,9 @@ def rel_href(base_href, href):
         return href
     if '/' not in base_href:
         return href
-    base = list(filter(lambda x: x and x != '.', os.path.dirname(os.path.normpath(base_href)).replace(os.sep, '/').split('/')))
+    base = list(filter(lambda x: x and x != '.',
+                       os.path.dirname(os.path.normpath(base_href))
+                       .replace(os.sep, '/').split('/')))
     while True:
         try:
             idx = base.index('..')
diff --git a/ebook_converter/ebooks/oeb/parse_utils.py b/ebook_converter/ebooks/oeb/parse_utils.py
index 64be9f2..ef1337b 100644
--- a/ebook_converter/ebooks/oeb/parse_utils.py
+++ b/ebook_converter/ebooks/oeb/parse_utils.py
@@ -1,20 +1,16 @@
 import re
 
-from lxml import etree, html
+from lxml import etree
+from lxml import html
 
+from ebook_converter import constants as const
 from ebook_converter import xml_replace_entities, force_unicode
-from ebook_converter.utils.xml_parse import safe_xml_fromstring
 from ebook_converter.constants_old import filesystem_encoding
 from ebook_converter.ebooks.chardet import xml_to_unicode, strip_encoding_declarations
 
 
-__license__ = 'GPL v3'
-__copyright__ = '2011, Kovid Goyal <kovid@kovidgoyal.net>'
-__docformat__ = 'restructuredtext en'
-
-RECOVER_PARSER = etree.XMLParser(recover=True, no_network=True, resolve_entities=False)
-XHTML_NS     = 'http://www.w3.org/1999/xhtml'
-XMLNS_NS     = 'http://www.w3.org/2000/xmlns/'
+RECOVER_PARSER = etree.XMLParser(recover=True, no_network=True,
+                                 resolve_entities=False)
 
 
 class NotHTML(Exception):
@@ -33,15 +29,15 @@ def namespace(name):
 
 
 def XHTML(name):
-    return '{%s}%s' % (XHTML_NS, name)
+    return '{%s}%s' % (const.XHTML_NS, name)
 
 
 def xpath(elem, expr):
-    return elem.xpath(expr, namespaces={'h':XHTML_NS})
+    return elem.xpath(expr, namespaces={'h':const.XHTML_NS})
 
 
 def XPath(expr):
-    return etree.XPath(expr, namespaces={'h':XHTML_NS})
+    return etree.XPath(expr, namespaces={'h':const.XHTML_NS})
 
 
 META_XP = XPath('/h:html/h:head/h:meta[@http-equiv="Content-Type"]')
@@ -111,7 +107,7 @@ def _html4_parse(data):
             elem.text = elem.text.strip('-')
     data = etree.tostring(data, encoding='unicode')
 
-    data = safe_xml_fromstring(data)
+    data = etree.fromstring(data)
     return data
 
 
@@ -204,14 +200,14 @@ def parse_html(data, log=None, decoder=None, preprocessor=None,
 
     # Try with more & more drastic measures to parse
     try:
-        data = safe_xml_fromstring(data, recover=False)
+        data = etree.fromstring(data)
         check_for_html5(pre, data)
     except (HTML5Doc, etree.XMLSyntaxError):
         log.debug('Initial parse failed, using more'
                 ' forgiving parsers')
         raw = data = xml_replace_entities(raw)
         try:
-            data = safe_xml_fromstring(data, recover=False)
+            data = etree.fromstring(data)
             check_for_html5(pre, data)
         except (HTML5Doc, etree.XMLSyntaxError):
             log.debug('Parsing %s as HTML' % filename)
@@ -240,7 +236,7 @@ def parse_html(data, log=None, decoder=None, preprocessor=None,
         if barename(data.tag) in non_html_file_tags:
             raise NotHTML(data.tag)
         log.warn('File %r does not appear to be (X)HTML'%filename)
-        nroot = safe_xml_fromstring('<html></html>')
+        nroot = etree.fromstring('<html></html>')
         has_body = False
         for child in list(data):
             if isinstance(child.tag, (str, bytes)) and barename(child.tag) == 'body':
@@ -249,7 +245,7 @@ def parse_html(data, log=None, decoder=None, preprocessor=None,
         parent = nroot
         if not has_body:
             log.warn('File %r appears to be a HTML fragment'%filename)
-            nroot = safe_xml_fromstring('<html><body/></html>')
+            nroot = etree.fromstring('<html><body/></html>')
             parent = nroot[0]
         for child in list(data.iter()):
             oparent = child.getparent()
@@ -261,16 +257,16 @@ def parse_html(data, log=None, decoder=None, preprocessor=None,
     # Force into the XHTML namespace
     if not namespace(data.tag):
         log.warn('Forcing', filename, 'into XHTML namespace')
-        data.attrib['xmlns'] = XHTML_NS
+        data.attrib['xmlns'] = const.XHTML_NS
         data = etree.tostring(data, encoding='unicode')
 
         try:
-            data = safe_xml_fromstring(data, recover=False)
+            data = etree.fromstring(data)
         except:
             data = data.replace(':=', '=').replace(':>', '>')
             data = data.replace('<http:/>', '')
             try:
-                data = safe_xml_fromstring(data, recover=False)
+                data = etree.fromstring(data)
             except etree.XMLSyntaxError:
                 log.warn('Stripping comments from %s'%
                         filename)
@@ -281,17 +277,17 @@ def parse_html(data, log=None, decoder=None, preprocessor=None,
                     '')
                 data = data.replace("<?xml version='1.0' encoding='utf-8'??>", '')
                 try:
-                    data = safe_xml_fromstring(data)
+                    data = etree.fromstring(data)
                 except etree.XMLSyntaxError:
                     log.warn('Stripping meta tags from %s'% filename)
                     data = re.sub(r'<meta\s+[^>]+?>', '', data)
-                    data = safe_xml_fromstring(data)
-    elif namespace(data.tag) != XHTML_NS:
+                    data = etree.fromstring(data)
+    elif namespace(data.tag) != const.XHTML_NS:
         # OEB_DOC_NS, but possibly others
         ns = namespace(data.tag)
         attrib = dict(data.attrib)
         nroot = etree.Element(XHTML('html'),
-            nsmap={None: XHTML_NS}, attrib=attrib)
+            nsmap={None: const.XHTML_NS}, attrib=attrib)
         for elem in data.iterdescendants():
             if isinstance(elem.tag, (str, bytes)) and \
                 namespace(elem.tag) == ns:
@@ -301,7 +297,7 @@ def parse_html(data, log=None, decoder=None, preprocessor=None,
         data = nroot
 
     # Remove non default prefixes referring to the XHTML namespace
-    data = ensure_namespace_prefixes(data, {None: XHTML_NS})
+    data = ensure_namespace_prefixes(data, {None: const.XHTML_NS})
 
     data = merge_multiple_html_heads_and_bodies(data, log)
     # Ensure has a <head/>
diff --git a/ebook_converter/ebooks/oeb/polish/container.py b/ebook_converter/ebooks/oeb/polish/container.py
index e3b9530..2ee2957 100644
--- a/ebook_converter/ebooks/oeb/polish/container.py
+++ b/ebook_converter/ebooks/oeb/polish/container.py
@@ -14,7 +14,9 @@ from itertools import count
 import urllib.parse
 
 from css_parser import getUrls, replaceUrls
+from lxml import etree
 
+from ebook_converter import constants as const
 from ebook_converter import CurrentDir, walk
 from ebook_converter.constants_old import iswindows
 from ebook_converter.customize.ui import plugin_for_input_format, plugin_for_output_format
@@ -34,7 +36,7 @@ from ebook_converter.ebooks.mobi import MobiError
 from ebook_converter.ebooks.mobi.reader.headers import MetadataHeader
 from ebook_converter.ebooks.mobi.tweak import set_cover
 from ebook_converter.ebooks.oeb.base import (
-    DC11_NS, OEB_DOCS, OEB_STYLES, OPF, OPF2_NS, Manifest, itercsslinks, iterlinks,
+    OEB_DOCS, OEB_STYLES, Manifest, itercsslinks, iterlinks,
     rewrite_links, serialize, urlquote, urlunquote
 )
 from ebook_converter.ebooks.oeb.parse_utils import NotHTML, parse_html
@@ -47,13 +49,11 @@ from ebook_converter.ptempfile import PersistentTemporaryDirectory, PersistentTe
 from ebook_converter.utils.filenames import hardlink_file, nlinks_file
 from ebook_converter.utils.ipc.simple_worker import WorkerError, fork_job
 from ebook_converter.utils.logging import default_log
-from ebook_converter.utils.xml_parse import safe_xml_fromstring
 from ebook_converter.utils.zipfile import ZipFile
 
 exists, join, relpath = os.path.exists, os.path.join, os.path.relpath
 
 OEB_FONTS = {guess_type('a.ttf'), guess_type('b.otf'), guess_type('a.woff'), 'application/x-font-ttf', 'application/x-font-otf', 'application/font-sfnt'}
-OPF_NAMESPACES = {'opf':OPF2_NS, 'dc':DC11_NS}
 null = object()
 
 
@@ -195,7 +195,7 @@ class ContainerBase(object):  # {{{
         data, self.used_encoding = xml_to_unicode(
             data, strip_encoding_pats=True, assume_utf8=True, resolve_entities=True)
         data = unicodedata.normalize('NFC', data)
-        return safe_xml_fromstring(data)
+        return etree.fromstring(data)
 
     def parse_xhtml(self, data, fname='<string>', force_html5_parse=False):
         if self.tweak_mode:
@@ -324,7 +324,7 @@ class Container(ContainerBase):  # {{{
             item_id = 'id' + '%d'%c
         manifest = self.opf_xpath('//opf:manifest')[0]
         href = self.name_to_href(name, self.opf_name)
-        item = manifest.makeelement(OPF('item'),
+        item = manifest.makeelement(const.OPF_ITEM,
                                     id=item_id, href=href)
         item.set('media-type', self.mime_map[name])
         self.insert_into_xml(manifest, item)
@@ -380,7 +380,7 @@ class Container(ContainerBase):  # {{{
         if mt in OEB_DOCS:
             manifest = self.opf_xpath('//opf:manifest')[0]
             spine = self.opf_xpath('//opf:spine')[0]
-            si = manifest.makeelement(OPF('itemref'), idref=item_id)
+            si = manifest.makeelement(const.OPF_ITEMREF, idref=item_id)
             self.insert_into_xml(spine, si, index=spine_index)
         return name
 
@@ -533,7 +533,7 @@ class Container(ContainerBase):  # {{{
 
     def opf_xpath(self, expr):
         ' Convenience method to evaluate an XPath expression on the OPF file, has the opf: and dc: namespace prefixes pre-defined. '
-        return self.opf.xpath(expr, namespaces=OPF_NAMESPACES)
+        return self.opf.xpath(expr, namespaces=const.OPF_NAMESPACES)
 
     def has_name(self, name):
         ''' Return True iff a file with the same canonical name as that specified exists. Unlike :meth:`exists` this method is always case-sensitive. '''
@@ -813,7 +813,8 @@ class Container(ContainerBase):  # {{{
         spine = self.opf_xpath('//opf:spine')[0]
         spine.text = tail
         for name, linear in spine_items:
-            i = spine.makeelement('{%s}itemref' % OPF_NAMESPACES['opf'], nsmap={'opf':OPF_NAMESPACES['opf']})
+            i = spine.makeelement(const.OPF_ITEMREF,
+                                  nsmap={'opf': const.OPF2_NS})
             i.tail = tail
             i.set('idref', imap[name])
             spine.append(i)
@@ -944,7 +945,7 @@ class Container(ContainerBase):  # {{{
             item_id = id_prefix + '%d'%c
 
         manifest = self.opf_xpath('//opf:manifest')[0]
-        item = manifest.makeelement(OPF('item'),
+        item = manifest.makeelement(const.OPF_ITEM,
                                     id=item_id, href=href)
         item.set('media-type', media_type)
         self.insert_into_xml(manifest, item)
@@ -993,7 +994,7 @@ class Container(ContainerBase):  # {{{
             self.format_opf()
         data = serialize(data, self.mime_map[name], pretty_print=name in
                          self.pretty_print)
-        if name == self.opf_name and root.nsmap.get(None) == OPF2_NS:
+        if name == self.opf_name and root.nsmap.get(None) == const.OPF2_NS:
             # Needed as I can't get lxml to output opf:role and
             # not output <opf:metadata> as well
             data = re.sub(br'(<[/]{0,1})opf:', r'\1', data)
@@ -1172,7 +1173,7 @@ class EpubContainer(Container):
         container_path = join(self.root, 'META-INF', 'container.xml')
         if not exists(container_path):
             raise InvalidEpub('No META-INF/container.xml in epub')
-        container = safe_xml_fromstring(open(container_path, 'rb').read())
+        container = etree.fromstring(open(container_path, 'rb').read())
         opf_files = container.xpath((
             r'child::ocf:rootfiles/ocf:rootfile'
             '[@media-type="%s" and @full-path]'%guess_type('a.opf')
diff --git a/ebook_converter/ebooks/oeb/polish/css.py b/ebook_converter/ebooks/oeb/polish/css.py
index ef7a5b7..43b209c 100644
--- a/ebook_converter/ebooks/oeb/polish/css.py
+++ b/ebook_converter/ebooks/oeb/polish/css.py
@@ -2,10 +2,11 @@ from collections import defaultdict
 from functools import partial
 
 from css_parser.css import CSSRule, CSSStyleDeclaration
-from ebook_converter.css_selectors import parse, SelectorSyntaxError
 
+from ebook_converter import constants as const
 from ebook_converter import force_unicode
-from ebook_converter.ebooks.oeb.base import OEB_STYLES, OEB_DOCS, XHTML, css_text
+from ebook_converter.css_selectors import parse, SelectorSyntaxError
+from ebook_converter.ebooks.oeb.base import OEB_STYLES, OEB_DOCS, css_text
 from ebook_converter.ebooks.oeb.normalize_css import normalize_filter_css, normalizers
 from ebook_converter.ebooks.oeb.polish.pretty import pretty_script_or_style, pretty_xml_tree, serialize
 from ebook_converter.utils.icu import numeric_sort_key
@@ -382,7 +383,7 @@ def add_stylesheet_links(container, name, text):
     if not sheets:
         return
     for sname in sheets:
-        link = head.makeelement(XHTML('link'), type='text/css', rel='stylesheet', href=container.name_to_href(sname, name))
+        link = head.makeelement(const.XHTML_LINK, type='text/css', rel='stylesheet', href=container.name_to_href(sname, name))
         head.append(link)
     pretty_xml_tree(head)
     return serialize(root, 'text/html')
diff --git a/ebook_converter/ebooks/oeb/polish/opf.py b/ebook_converter/ebooks/oeb/polish/opf.py
index b66adcb..77cd848 100644
--- a/ebook_converter/ebooks/oeb/polish/opf.py
+++ b/ebook_converter/ebooks/oeb/polish/opf.py
@@ -1,13 +1,9 @@
 from lxml import etree
 
-from ebook_converter.ebooks.oeb.polish.container import OPF_NAMESPACES
+from ebook_converter import constants as const
 from ebook_converter.utils.localization import canonicalize_lang
 
 
-__license__ = 'GPL v3'
-__copyright__ = '2014, Kovid Goyal <kovid at kovidgoyal.net>'
-
-
 def get_book_language(container):
     for lang in container.opf_xpath('//dc:language'):
         raw = lang.text
@@ -18,7 +14,7 @@ def get_book_language(container):
 
 
 def set_guide_item(container, item_type, title, name, frag=None):
-    ref_tag = '{%s}reference' % OPF_NAMESPACES['opf']
+    ref_tag = const.OPF_REFERENCE
     href = None
     if name:
         href = container.name_to_href(name, container.opf_name)
@@ -27,23 +23,27 @@ def set_guide_item(container, item_type, title, name, frag=None):
 
     guides = container.opf_xpath('//opf:guide')
     if not guides and href:
-        g = container.opf.makeelement('{%s}guide' % OPF_NAMESPACES['opf'], nsmap={'opf':OPF_NAMESPACES['opf']})
+        g = container.opf.makeelement(const.OPF_GUIDE,
+                                      nsmap={'opf': const.OPF2_NS})
         container.insert_into_xml(container.opf, g)
         guides = [g]
 
     for guide in guides:
         matches = []
         for child in guide.iterchildren(etree.Element):
-            if child.tag == ref_tag and child.get('type', '').lower() == item_type.lower():
+            if (child.tag == ref_tag and
+                    child.get('type', '').lower() == item_type.lower()):
                 matches.append(child)
         if not matches and href:
-            r = guide.makeelement(ref_tag, type=item_type, nsmap={'opf':OPF_NAMESPACES['opf']})
+            r = guide.makeelement(ref_tag, type=item_type,
+                                  nsmap={'opf': const.OPF2_NS})
             container.insert_into_xml(guide, r)
             matches.append(r)
         for m in matches:
             if href:
-                m.set('title', title), m.set('href', href), m.set('type', item_type)
+                m.set('title', title)
+                m.set('href', href)
+                m.set('type', item_type)
             else:
                 container.remove_from_xml(m)
     container.dirty(container.opf_name)
-
diff --git a/ebook_converter/ebooks/oeb/polish/parsing.py b/ebook_converter/ebooks/oeb/polish/parsing.py
index d847094..99cbd5e 100644
--- a/ebook_converter/ebooks/oeb/polish/parsing.py
+++ b/ebook_converter/ebooks/oeb/polish/parsing.py
@@ -1,21 +1,18 @@
 import re
 
-from lxml.etree import Element as LxmlElement
+from lxml import etree
 import html5_parser
 
+from ebook_converter import constants as const
 from ebook_converter import xml_replace_entities
-from ebook_converter.utils.xml_parse import safe_xml_fromstring
-from ebook_converter.ebooks.chardet import xml_to_unicode, strip_encoding_declarations
+from ebook_converter.ebooks.chardet import strip_encoding_declarations
+from ebook_converter.ebooks.chardet import xml_to_unicode
 from ebook_converter.utils.cleantext import clean_xml_chars
 
 
-__license__ = 'GPL v3'
-__copyright__ = '2013, Kovid Goyal <kovid at kovidgoyal.net>'
-
-XHTML_NS     = 'http://www.w3.org/1999/xhtml'
-
-
-def parse_html5(raw, decoder=None, log=None, discard_namespaces=False, line_numbers=True, linenumber_attribute=None, replace_entities=True, fix_newlines=True):
+def parse_html5(raw, decoder=None, log=None, discard_namespaces=False,
+                line_numbers=True, linenumber_attribute=None,
+                replace_entities=True, fix_newlines=True):
     if isinstance(raw, bytes):
         raw = xml_to_unicode(raw)[0] if decoder is None else decoder(raw)
     if replace_entities:
@@ -23,10 +20,14 @@ def parse_html5(raw, decoder=None, log=None, discard_namespaces=False, line_numb
     if fix_newlines:
         raw = raw.replace('\r\n', '\n').replace('\r', '\n')
     raw = clean_xml_chars(raw)
-    root = html5_parser.parse(raw, maybe_xhtml=not discard_namespaces, line_number_attr=linenumber_attribute, keep_doctype=False, sanitize_names=True)
-    if (discard_namespaces and root.tag != 'html') or (
-        not discard_namespaces and (root.tag != '{%s}%s' % (XHTML_NS, 'html') or root.prefix)):
-        raise ValueError('Failed to parse correctly, root has tag: %s and prefix: %s' % (root.tag, root.prefix))
+    root = html5_parser.parse(raw, maybe_xhtml=not discard_namespaces,
+                              line_number_attr=linenumber_attribute,
+                              keep_doctype=False, sanitize_names=True)
+    if ((discard_namespaces and root.tag != 'html') or
+        (not discard_namespaces and
+         (root.tag != '{%s}%s' % (const.XHTML_NS, 'html') or root.prefix))):
+        raise ValueError('Failed to parse correctly, root has tag: %s and '
+                         'prefix: %s' % (root.tag, root.prefix))
     return root
 
 
@@ -48,12 +49,14 @@ def handle_private_entities(data):
                 user_entities[match.group(1)] = val
             if user_entities:
                 data = ('\n' * num_of_nl_in_pre) + data[idx:]
-                pat = re.compile(r'&(%s);'%('|'.join(user_entities.keys())))
-                data = pat.sub(lambda m:user_entities[m.group(1)], data)
+                pat = re.compile(r'&(%s);' % ('|'.join(user_entities.keys())))
+                data = pat.sub(lambda m: user_entities[m.group(1)], data)
     return data
 
 
-def parse(raw, decoder=None, log=None, line_numbers=True, linenumber_attribute=None, replace_entities=True, force_html5_parse=False):
+def parse(raw, decoder=None, log=None, line_numbers=True,
+          linenumber_attribute=None, replace_entities=True,
+          force_html5_parse=False):
     if isinstance(raw, bytes):
         raw = xml_to_unicode(raw)[0] if decoder is None else decoder(raw)
     raw = handle_private_entities(raw)
@@ -70,26 +73,32 @@ def parse(raw, decoder=None, log=None, line_numbers=True, linenumber_attribute=N
         raw = ('\n' * newlines) + raw[match.start():]
         break
 
-    raw = strip_encoding_declarations(raw, limit=10*1024, preserve_newlines=True)
+    raw = strip_encoding_declarations(raw, limit=10*1024,
+                                      preserve_newlines=True)
     if force_html5_parse:
-        return parse_html5(raw, log=log, line_numbers=line_numbers, linenumber_attribute=linenumber_attribute, replace_entities=False, fix_newlines=False)
+        return parse_html5(raw, log=log, line_numbers=line_numbers,
+                           linenumber_attribute=linenumber_attribute,
+                           replace_entities=False, fix_newlines=False)
     try:
-        ans = safe_xml_fromstring(raw, recover=False)
-        if ans.tag != '{%s}html' % XHTML_NS:
+        ans = etree.fromstring(raw)
+        if ans.tag != '{%s}html' % const.XHTML_NS:
             raise ValueError('Root tag is not <html> in the XHTML namespace')
         if linenumber_attribute:
-            for elem in ans.iter(LxmlElement):
+            for elem in ans.iter(etree.element):
                 if elem.sourceline is not None:
                     elem.set(linenumber_attribute, str(elem.sourceline))
         return ans
     except Exception:
         if log is not None:
             log.exception('Failed to parse as XML, parsing as tag soup')
-        return parse_html5(raw, log=log, line_numbers=line_numbers, linenumber_attribute=linenumber_attribute, replace_entities=False, fix_newlines=False)
+        return parse_html5(raw, log=log, line_numbers=line_numbers,
+                           linenumber_attribute=linenumber_attribute,
+                           replace_entities=False, fix_newlines=False)
 
 
 if __name__ == '__main__':
-    from lxml import etree
-    root = parse_html5('\n<html><head><title>a\n
 \nb', discard_namespaces=False)
+    root = parse_html5('\na\n
 '
+                       '\nb',
+                       discard_namespaces=False)
     print(etree.tostring(root, encoding='utf-8'))
     print()
diff --git a/ebook_converter/ebooks/oeb/polish/pretty.py b/ebook_converter/ebooks/oeb/polish/pretty.py
index 402a546..8620f9e 100644
--- a/ebook_converter/ebooks/oeb/polish/pretty.py
+++ b/ebook_converter/ebooks/oeb/polish/pretty.py
@@ -2,10 +2,10 @@ import textwrap
 
 # from lxml.etree import Element
 
+from ebook_converter import constants as const
 from ebook_converter import force_unicode
-from ebook_converter.ebooks.oeb.base import (
-    serialize, OEB_DOCS, barename, OEB_STYLES, XPNSMAP, XHTML, SVG)
-from ebook_converter.ebooks.oeb.polish.container import OPF_NAMESPACES
+from ebook_converter.ebooks.oeb import parse_utils
+from ebook_converter.ebooks.oeb.base import serialize, OEB_DOCS, OEB_STYLES
 from ebook_converter.ebooks.oeb.polish.utils import guess_type
 from ebook_converter.utils.icu import sort_key
 
@@ -38,15 +38,15 @@ def pretty_opf(root):
     # Put all dc: tags first starting with title and author. Preserve order for
     # the rest.
     def dckey(x):
-        return {'title':0, 'creator':1}.get(barename(x.tag), 2)
-    for metadata in root.xpath('//opf:metadata', namespaces=OPF_NAMESPACES):
-        dc_tags = metadata.xpath('./*[namespace-uri()="%s"]' % OPF_NAMESPACES['dc'])
+        return {'title':0, 'creator':1}.get(parse_utils.barename(x.tag), 2)
+    for metadata in root.xpath('//opf:metadata', namespaces=const.OPF_NAMESPACES):
+        dc_tags = metadata.xpath('./*[namespace-uri()="%s"]' % const.DC11_NS)
         dc_tags.sort(key=dckey)
         for x in reversed(dc_tags):
             metadata.insert(0, x)
 
     # Group items in the manifest
-    spine_ids = root.xpath('//opf:spine/opf:itemref/@idref', namespaces=OPF_NAMESPACES)
+    spine_ids = root.xpath('//opf:spine/opf:itemref/@idref', namespaces=const.OPF_NAMESPACES)
     spine_ids = {x:i for i, x in enumerate(spine_ids)}
 
     def manifest_key(x):
@@ -75,7 +75,7 @@ def pretty_opf(root):
             i = sort_key(href)
         return (cat, i)
 
-    for manifest in root.xpath('//opf:manifest', namespaces=OPF_NAMESPACES):
+    for manifest in root.xpath('//opf:manifest', namespaces=const.OPF_NAMESPACES):
         try:
             children = sorted(manifest, key=manifest_key)
         except AttributeError:
@@ -84,19 +84,11 @@ def pretty_opf(root):
             manifest.insert(0, x)
 
 
-SVG_TAG = SVG('svg')
-BLOCK_TAGS = frozenset(map(XHTML, (
-    'address', 'article', 'aside', 'audio', 'blockquote', 'body', 'canvas', 'col', 'colgroup', 'dd',
-    'div', 'dl', 'dt', 'fieldset', 'figcaption', 'figure', 'footer', 'form',
-    'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'header', 'hgroup', 'hr', 'li',
-    'noscript', 'ol', 'output', 'p', 'pre', 'script', 'section', 'style', 'table', 'tbody', 'td',
-    'tfoot', 'th', 'thead', 'tr', 'ul', 'video', 'img'))) | {SVG_TAG}
-
 
 def isblock(x):
     if callable(x.tag) or not x.tag:
         return True
-    if x.tag in BLOCK_TAGS:
+    if x.tag in const.XHTML_BLOCK_TAGS | {const.SVG_SVG}:
         return True
     return False
 
@@ -141,12 +133,12 @@ def pretty_block(parent, level=1, indent='  '):
     that contain only other block tags '''
     if not parent.text or isspace(parent.text):
         parent.text = ''
-    nn = '\n' if hasattr(parent.tag, 'strip') and barename(parent.tag) in {'tr', 'td', 'th'} else '\n\n'
+    nn = '\n' if hasattr(parent.tag, 'strip') and parse_utils.barename(parent.tag) in {'tr', 'td', 'th'} else '\n\n'
     parent.text = parent.text + nn + (indent * level)
     for i, child in enumerate(parent):
         if isblock(child) and has_only_blocks(child):
             pretty_block(child, level=level+1, indent=indent)
-        elif child.tag == SVG_TAG:
+        elif child.tag == const.SVG_SVG:
             pretty_xml_tree(child, level=level, indent=indent)
         l = level
         if i == len(parent) - 1:
@@ -172,13 +164,13 @@ def pretty_html_tree(container, root):
         child.tail = '\n\n'
         if hasattr(child.tag, 'endswith') and child.tag.endswith('}head'):
             pretty_xml_tree(child)
-    for body in root.findall('h:body', namespaces=XPNSMAP):
+    for body in root.findall('h:body', namespaces=const.XPNSMAP):
         pretty_block(body)
         # Special case the handling of a body that contains a single block tag
         # with all content. In this case we prettify the containing block tag
         # even if it has non block children.
         if (len(body) == 1 and not callable(body[0].tag) and isblock(body[0]) and not has_only_blocks(
-            body[0]) and barename(body[0].tag) not in (
+            body[0]) and parse_utils.barename(body[0].tag) not in (
                     'pre', 'p', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6') and len(body[0]) > 0):
             pretty_block(body[0], level=2)
 
diff --git a/ebook_converter/ebooks/oeb/polish/split.py b/ebook_converter/ebooks/oeb/polish/split.py
index 568aefa..36b8a67 100644
--- a/ebook_converter/ebooks/oeb/polish/split.py
+++ b/ebook_converter/ebooks/oeb/polish/split.py
@@ -1,7 +1,11 @@
-import copy, os, re
+import copy
+import os
+import re
 import urllib.parse
 
-from ebook_converter.ebooks.oeb.base import barename, XPNSMAP, XPath, OPF, XHTML, OEB_DOCS
+from ebook_converter import constants as const
+from ebook_converter.ebooks.oeb import base
+from ebook_converter.ebooks.oeb import parse_utils
 from ebook_converter.ebooks.oeb.polish.errors import MalformedMarkup
 from ebook_converter.ebooks.oeb.polish.toc import node_from_loc
 from ebook_converter.ebooks.oeb.polish.replace import LinkRebaser
@@ -35,7 +39,7 @@ def adjust_split_point(split_point, log):
         parent = sp.getparent()
         if (
             parent is None or
-            barename(parent.tag) in {'body', 'html'} or
+            parse_utils.barename(parent.tag) in {'body', 'html'} or
             (parent.text and parent.text.strip()) or
             parent.index(sp) > 0
         ):
@@ -49,7 +53,7 @@ def adjust_split_point(split_point, log):
 
 
 def get_body(root):
-    return root.find('h:body', namespaces=XPNSMAP)
+    return root.find('h:body', namespaces=const.XPNSMAP)
 
 
 def do_split(split_point, log, before=True):
@@ -113,7 +117,7 @@ def do_split(split_point, log, before=True):
             nix_element(elem)
 
     # Tree 2
-    ancestors = frozenset(XPath('ancestor::*')(split_point2))
+    ancestors = frozenset(base.XPath('ancestor::*')(split_point2))
     for elem in tuple(body2.iterdescendants()):
         if elem is split_point2:
             if not before:
@@ -251,7 +255,7 @@ def split(container, name, loc_or_xpath, before=True, totals=None):
             break
     index = spine.index(spine_item) + 1
 
-    si = spine.makeelement(OPF('itemref'), idref=manifest_item.get('id'))
+    si = spine.makeelement(base.tag('opf', 'itemref'), idref=manifest_item.get('id'))
     if not linear:
         si.set('linear', 'no')
     container.insert_into_xml(spine, si, index=index)
@@ -268,7 +272,7 @@ def multisplit(container, name, xpath, before=True):
     :param before: If True the splits occur before the identified element otherwise after it.
     '''
     root = container.parsed(name)
-    nodes = root.xpath(xpath, namespaces=XPNSMAP)
+    nodes = root.xpath(xpath, namespaces=const.XPNSMAP)
     if not nodes:
         raise AbortError('The expression %s did not match any nodes' % xpath)
     for split_point in nodes:
@@ -329,7 +333,7 @@ def all_anchors(root):
 
 
 def all_stylesheets(container, name):
-    for link in XPath('//h:head/h:link[@href]')(container.parsed(name)):
+    for link in base.XPath('//h:head/h:link[@href]')(container.parsed(name)):
         name = container.href_to_name(link.get('href'), name)
         typ = link.get('type', 'text/css')
         if typ == 'text/css':
@@ -358,14 +362,14 @@ def merge_html(container, names, master, insert_page_breaks=False):
     root = p(master)
 
     # Ensure master has a 
-    head = root.find('h:head', namespaces=XPNSMAP)
+    head = root.find('h:head', namespaces=const.XPNSMAP)
     if head is None:
-        head = root.makeelement(XHTML('head'))
+        head = root.makeelement(base.tag('xhtml', 'head'))
         container.insert_into_xml(root, head, 0)
 
     seen_anchors = all_anchors(root)
     seen_stylesheets = set(all_stylesheets(container, master))
-    master_body = p(master).findall('h:body', namespaces=XPNSMAP)[-1]
+    master_body = p(master).findall('h:body', namespaces=const.XPNSMAP)[-1]
     master_base = os.path.dirname(master)
     anchor_map = {n:{} for n in names if n != master}
     first_anchor_map = {}
@@ -377,7 +381,7 @@ def merge_html(container, names, master, insert_page_breaks=False):
         for sheet in all_stylesheets(container, name):
             if sheet not in seen_stylesheets:
                 seen_stylesheets.add(sheet)
-                link = head.makeelement(XHTML('link'), rel='stylesheet', type='text/css', href=container.name_to_href(sheet, master))
+                link = head.makeelement(base.tag('xhtml', 'link'), rel='stylesheet', type='text/css', href=container.name_to_href(sheet, master))
                 container.insert_into_xml(head, link)
 
         # Rebase links if master is in a different directory
@@ -386,7 +390,7 @@ def merge_html(container, names, master, insert_page_breaks=False):
 
         root = p(name)
         children = []
-        for body in p(name).findall('h:body', namespaces=XPNSMAP):
+        for body in p(name).findall('h:body', namespaces=const.XPNSMAP):
             children.append(body.text if body.text and body.text.strip() else '\n\n')
             children.extend(body)
 
@@ -396,7 +400,7 @@ def merge_html(container, names, master, insert_page_breaks=False):
                 break
         if isinstance(first_child, (str, bytes)):
             # body contained only text, no tags
-            first_child = body.makeelement(XHTML('p'))
+            first_child = body.makeelement(base.tag('xhtml', 'p'))
             first_child.text, children[0] = children[0], first_child
 
         amap = anchor_map[name]
@@ -424,7 +428,7 @@ def merge_html(container, names, master, insert_page_breaks=False):
         amap[''] = first_child.get('id')
 
         # Fix links that point to local changed anchors
-        for a in XPath('//h:a[starts-with(@href, "#")]')(root):
+        for a in base.XPath('//h:a[starts-with(@href, "#")]')(root):
             q = a.get('href')[1:]
             if q in amap:
                 a.set('href', '#' + amap[q])
@@ -472,10 +476,10 @@ def merge_css(container, names, master):
     # Remove links to merged stylesheets in the html files, replacing with a
     # link to the master sheet
     for name, mt in container.mime_map.items():
-        if mt in OEB_DOCS:
+        if mt in base.OEB_DOCS:
             removed = False
             root = p(name)
-            for link in XPath('//h:link[@href]')(root):
+            for link in base.XPath('//h:link[@href]')(root):
                 q = container.href_to_name(link.get('href'), name)
                 if q in merged:
                     container.remove_from_xml(link)
@@ -483,9 +487,9 @@ def merge_css(container, names, master):
             if removed:
                 container.dirty(name)
             if removed and master not in set(all_stylesheets(container, name)):
-                head = root.find('h:head', namespaces=XPNSMAP)
+                head = root.find('h:head', namespaces=const.XPNSMAP)
                 if head is not None:
-                    link = head.makeelement(XHTML('link'), type='text/css', rel='stylesheet', href=container.name_to_href(master, name))
+                    link = head.makeelement(base.tag('xhtml', 'link'), type='text/css', rel='stylesheet', href=container.name_to_href(master, name))
                     container.insert_into_xml(head, link)
 
 
diff --git a/ebook_converter/ebooks/oeb/polish/toc.py b/ebook_converter/ebooks/oeb/polish/toc.py
index 15f94b4..42af230 100644
--- a/ebook_converter/ebooks/oeb/polish/toc.py
+++ b/ebook_converter/ebooks/oeb/polish/toc.py
@@ -1,16 +1,16 @@
-import re
-from collections import Counter, OrderedDict
-from functools import partial
-from operator import itemgetter
+import collections
+import functools
+import operator
 import pkg_resources
+import re
 import urllib.parse
 
 from lxml import etree
 from lxml.builder import ElementMaker
 
 from ebook_converter import __version__
-from ebook_converter.ebooks.oeb.base import (
-    XPath, uuid_id, xml2text, NCX, NCX_NS, XML, XHTML, XHTML_NS, serialize, EPUB_NS, XML_NS, OEB_DOCS)
+from ebook_converter import constants as const
+from ebook_converter.ebooks.oeb import base
 from ebook_converter.ebooks.oeb.polish.errors import MalformedMarkup
 from ebook_converter.ebooks.oeb.polish.utils import guess_type, extract
 from ebook_converter.ebooks.oeb.polish.opf import set_guide_item, get_book_language
@@ -18,10 +18,6 @@ from ebook_converter.ebooks.oeb.polish.pretty import pretty_html_tree
 from ebook_converter.utils.localization import get_lang, canonicalize_lang, lang_as_iso639_1
 
 
-__license__ = 'GPL v3'
-__copyright__ = '2013, Kovid Goyal '
-__docformat__ = 'restructuredtext en'
-
 ns = etree.FunctionNamespace('calibre_xpath_extensions')
 ns.prefix = 'calibre'
 ns['lower-case'] = lambda c, x: x.lower() if hasattr(x, 'lower') else x
@@ -81,7 +77,8 @@ class TOC(object):
         seen = set()
         remove = []
         for child in self:
-            key = child.title if only_text else (child.title, child.dest, (child.frag or None))
+            key = child.title if only_text else (child.title, child.dest,
+                                                 (child.frag or None))
             if key in seen:
                 remove.append(child)
             else:
@@ -104,7 +101,7 @@ class TOC(object):
 
     def get_lines(self, lvl=0):
         frag = ('#'+self.frag) if self.frag else ''
-        ans = [('\t'*lvl) + 'TOC: %s --> %s%s'%(self.title, self.dest, frag)]
+        ans = [('\t'*lvl) + 'TOC: %s --> %s%s' % (self.title, self.dest, frag)]
         for child in self:
             ans.extend(child.get_lines(lvl+1))
         return ans
@@ -113,10 +110,8 @@ class TOC(object):
         return '\n'.join(self.get_lines())
 
     def to_dict(self, node_counter=None):
-        ans = {
-            'title':self.title, 'dest':self.dest, 'frag':self.frag,
-            'children':[c.to_dict(node_counter) for c in self.children]
-        }
+        ans = {'title': self.title, 'dest': self.dest, 'frag': self.frag,
+               'children': [c.to_dict(node_counter) for c in self.children]}
         if self.dest_exists is not None:
             ans['dest_exists'] = self.dest_exists
         if self.dest_error is not None:
@@ -131,7 +126,7 @@ class TOC(object):
 
 
 def child_xpath(tag, name):
-    return tag.xpath('./*[calibre:lower-case(local-name()) = "%s"]'%name)
+    return tag.xpath('./*[calibre:lower-case(local-name()) = "%s"]' % name)
 
 
 def add_from_navpoint(container, navpoint, parent, ncx_name):
@@ -142,7 +137,7 @@ def add_from_navpoint(container, navpoint, parent, ncx_name):
         text = ''
         for txt in child_xpath(nl, 'text'):
             text += etree.tostring(txt, method='text',
-                    encoding='unicode', with_tail=False)
+                                   encoding='unicode', with_tail=False)
     content = child_xpath(navpoint, 'content')
     if content:
         content = content[0]
@@ -154,7 +149,8 @@ def add_from_navpoint(container, navpoint, parent, ncx_name):
 
 
 def process_ncx_node(container, node, toc_parent, ncx_name):
-    for navpoint in node.xpath('./*[calibre:lower-case(local-name()) = "navpoint"]'):
+    for navpoint in node.xpath('./*[calibre:lower-case(local-name()) '
+                               '= "navpoint"]'):
         child = add_from_navpoint(container, navpoint, toc_parent, ncx_name)
         if child is not None:
             process_ncx_node(container, navpoint, child, ncx_name)
@@ -171,29 +167,38 @@ def parse_ncx(container, ncx_name):
         if attr.endswith('lang'):
             toc_root.lang = str(val)
             break
-    for uid in root.xpath('//*[calibre:lower-case(local-name()) = "meta" and @name="dtb:uid"]/@content'):
+    for uid in root.xpath('//*[calibre:lower-case(local-name()) = "meta" and '
+                          '@name="dtb:uid"]/@content'):
         if uid:
             toc_root.uid = str(uid)
             break
     for pl in root.xpath('//*[calibre:lower-case(local-name()) = "pagelist"]'):
-        for pt in pl.xpath('descendant::*[calibre:lower-case(local-name()) = "pagetarget"]'):
+        for pt in pl.xpath('descendant::*[calibre:lower-case(local-name()) = '
+                           '"pagetarget"]'):
             pagenum = pt.get('value')
             if pagenum:
-                href = pt.xpath('descendant::*[calibre:lower-case(local-name()) = "content"]/@src')
+                href = pt.xpath('descendant::*[calibre:lower-case(local-name()'
+                                ') = "content"]/@src')
                 if href:
                     dest = container.href_to_name(href[0], base=ncx_name)
                     frag = urllib.parse.urlparse(href[0]).fragment or None
-                    toc_root.page_list.append({'dest': dest, 'pagenum': pagenum, 'frag': frag})
+                    toc_root.page_list.append({'dest': dest,
+                                               'pagenum': pagenum,
+                                               'frag': frag})
     return toc_root
 
 
 def add_from_li(container, li, parent, nav_name):
     dest = frag = text = None
-    for x in li.iterchildren(XHTML('a'), XHTML('span')):
-        text = etree.tostring(x, method='text', encoding='unicode', with_tail=False).strip() or ' '.join(x.xpath('descendant-or-self::*/@title')).strip()
+    for x in li.iterchildren(base.tag('xhtml', 'a'),
+                             base.tag('xhtml', 'span')):
+        text = (etree.tostring(x, method='text', encoding='unicode',
+                               with_tail=False).strip() or
+                ' '.join(x.xpath('descendant-or-self::*/@title')).strip())
         href = x.get('href')
         if href:
-            dest = nav_name if href.startswith('#') else container.href_to_name(href, base=nav_name)
+            dest = (nav_name if href.startswith('#') else
+                    container.href_to_name(href, base=nav_name))
             frag = urllib.parse.urlparse(href).fragment or None
         break
     return parent.add(text or None, dest or None, frag or None)
@@ -207,9 +212,9 @@ def first_child(parent, tagname):
 
 
 def process_nav_node(container, node, toc_parent, nav_name):
-    for li in node.iterchildren(XHTML('li')):
+    for li in node.iterchildren(base.tag('xhtml', 'li')):
         child = add_from_li(container, li, toc_parent, nav_name)
-        ol = first_child(li, XHTML('ol'))
+        ol = first_child(li, base.tag('xhtml', 'ol'))
         if child is not None and ol is not None:
             process_nav_node(container, ol, child, nav_name)
 
@@ -218,14 +223,16 @@ def parse_nav(container, nav_name):
     root = container.parsed(nav_name)
     toc_root = TOC()
     toc_root.lang = toc_root.uid = None
-    et = '{%s}type' % EPUB_NS
-    for nav in root.iterdescendants(XHTML('nav')):
-        if nav.get(et) == 'toc':
-            ol = first_child(nav, XHTML('ol'))
+    xhtml = functools.partial(base.tag, 'xhtml')
+    for nav in root.iterdescendants(base.tag('xhtml', 'nav')):
+        if nav.get(base.tag('epub', 'type')) == 'toc':
+            ol = first_child(nav, base.tag('xhtml', 'ol'))
             if ol is not None:
                 process_nav_node(container, ol, toc_root, nav_name)
-                for h in nav.iterchildren(*map(XHTML, 'h1 h2 h3 h4 h5 h6'.split())):
-                    text = etree.tostring(h, method='text', encoding='unicode', with_tail=False) or h.get('title')
+                for h in nav.iterchildren(*map(xhtml,
+                                               'h1 h2 h3 h4 h5 h6'.split())):
+                    text = etree.tostring(h, method='text', encoding='unicode',
+                                          with_tail=False) or h.get('title')
                     if text:
                         toc_root.toc_title = text
                         break
@@ -235,7 +242,7 @@ def parse_nav(container, nav_name):
 
 def verify_toc_destinations(container, toc):
     anchor_map = {}
-    anchor_xpath = XPath('//*/@id|//h:a/@name')
+    anchor_xpath = base.XPath('//*/@id|//h:a/@name')
     for item in toc.iterdescendants():
         name = item.dest
         if not name:
@@ -284,7 +291,8 @@ def get_x_toc(container, find_toc, parse_toc, verify_destinations=True):
         ans.lang = ans.uid = None
         return ans
     toc = find_toc(container)
-    ans = empty_toc() if toc is None or not container.has_name(toc) else parse_toc(container, toc)
+    ans = (empty_toc() if toc is None or not container.has_name(toc) else
+           parse_toc(container, toc))
     ans.toc_file_name = toc if toc and container.has_name(toc) else None
     if verify_destinations:
         verify_toc_destinations(container, ans)
@@ -294,11 +302,14 @@ def get_x_toc(container, find_toc, parse_toc, verify_destinations=True):
 def get_toc(container, verify_destinations=True):
     ver = container.opf_version_parsed
     if ver.major < 3:
-        return get_x_toc(container, find_existing_ncx_toc, parse_ncx, verify_destinations=verify_destinations)
+        return get_x_toc(container, find_existing_ncx_toc, parse_ncx,
+                         verify_destinations=verify_destinations)
     else:
-        ans = get_x_toc(container, find_existing_nav_toc, parse_nav, verify_destinations=verify_destinations)
+        ans = get_x_toc(container, find_existing_nav_toc, parse_nav,
+                        verify_destinations=verify_destinations)
         if len(ans) == 0:
-            ans = get_x_toc(container, find_existing_ncx_toc, parse_ncx, verify_destinations=verify_destinations)
+            ans = get_x_toc(container, find_existing_ncx_toc, parse_ncx,
+                            verify_destinations=verify_destinations)
         return ans
 
 
@@ -308,25 +319,33 @@ def get_guide_landmarks(container):
         href, frag = href.partition('#')[::2]
         name = container.href_to_name(href, container.opf_name)
         if container.has_name(name):
-            yield {'dest':name, 'frag':frag, 'title':title or '', 'type':rtype or ''}
+            yield {'dest': name,
+                   'frag': frag,
+                   'title': title or '',
+                   'type': rtype or ''}
 
 
 def get_nav_landmarks(container):
     nav = find_existing_nav_toc(container)
     if nav and container.has_name(nav):
         root = container.parsed(nav)
-        et = '{%s}type' % EPUB_NS
-        for elem in root.iterdescendants(XHTML('nav')):
+        et = base('epub', 'type')
+        for elem in root.iterdescendants(base.tag('xhtml', 'nav')):
             if elem.get(et) == 'landmarks':
-                for li in elem.iterdescendants(XHTML('li')):
-                    for a in li.iterdescendants(XHTML('a')):
+                for li in elem.iterdescendants(base.tag('xhtml', 'li')):
+                    for a in li.iterdescendants(base.tag('xhtml', 'a')):
                         href, rtype = a.get('href'), a.get(et)
                         if href:
-                            title = etree.tostring(a, method='text', encoding='unicode', with_tail=False).strip()
+                            title = etree.tostring(a, method='text',
+                                                   encoding='unicode',
+                                                   with_tail=False).strip()
                             href, frag = href.partition('#')[::2]
                             name = container.href_to_name(href, nav)
                             if container.has_name(name):
-                                yield {'dest':name, 'frag':frag, 'title':title or '', 'type':rtype or ''}
+                                yield {'dest': name,
+                                       'frag': frag,
+                                       'title': title or '',
+                                       'type': rtype or ''}
                             break
 
 
@@ -344,7 +363,7 @@ def ensure_id(elem, all_ids):
     elem_id = elem.get('id')
     if elem_id:
         return False, elem_id
-    if elem.tag == XHTML('a'):
+    if elem.tag == base.tag('xhtml', 'a'):
         anchor = elem.get('name', None)
         if anchor:
             elem.set('id', anchor)
@@ -361,7 +380,7 @@ def ensure_id(elem, all_ids):
 
 
 def elem_to_toc_text(elem):
-    text = xml2text(elem).strip()
+    text = base.xml2text(elem).strip()
     if not text:
         text = elem.get('title', '')
     if not text:
@@ -375,7 +394,7 @@ def elem_to_toc_text(elem):
 
 def item_at_top(elem):
     try:
-        body = XPath('//h:body')(elem.getroottree().getroot())[0]
+        body = base.XPath('//h:body')(elem.getroottree().getroot())[0]
     except (TypeError, IndexError, KeyError, AttributeError):
         return False
     tree = body.getroottree()
@@ -387,7 +406,7 @@ def item_at_top(elem):
         try:
             if el.tag.endswith('}img') or (el.text and el.text.strip()):
                 return False
-        except:
+        except Exception:
             return False
         if not path.startswith(epath):
             # Only check tail of non-parent elements
@@ -404,24 +423,26 @@ def from_xpaths(container, xpaths):
     Table of Contents from the ``
``, ``
`` and ``
`` tags.
     '''
     tocroot = TOC()
-    xpaths = [XPath(xp) for xp in xpaths]
+    xpaths = [base.XPath(xp) for xp in xpaths]
 
     # Find those levels that have no elements in all spine items
-    maps = OrderedDict()
+    maps = collections.OrderedDict()
     empty_levels = {i+1 for i, xp in enumerate(xpaths)}
     for spinepath in container.spine_items:
         name = container.abspath_to_name(spinepath)
         root = container.parsed(name)
-        level_item_map = maps[name] = {i+1:frozenset(xp(root)) for i, xp in enumerate(xpaths)}
+        level_item_map = maps[name] = {i + 1: frozenset(xp(root))
+                                       for i, xp in enumerate(xpaths)}
         for lvl, elems in level_item_map.items():
             if elems:
                 empty_levels.discard(lvl)
     # Remove empty levels from all level_maps
     if empty_levels:
         for name, lmap in tuple(maps.items()):
-            lmap = {lvl:items for lvl, items in lmap.items() if lvl not in empty_levels}
-            lmap = sorted(lmap.items(), key=itemgetter(0))
-            lmap = {i+1:items for i, (l, items) in enumerate(lmap)}
+            lmap = {lvl: items for lvl, items in lmap.items()
+                    if lvl not in empty_levels}
+            lmap = sorted(lmap.items(), key=operator.itemgetter(0))
+            lmap = {i + 1: items for i, (l, items) in enumerate(lmap)}
             maps[name] = lmap
 
     node_level_map = {tocroot: 0}
@@ -434,13 +455,15 @@ def from_xpaths(container, xpaths):
             if child is None:
                 return node
             lvl = node_level_map[child]
-            return node if lvl > limit else child if lvl == limit else process_node(child)
+            return (node if lvl > limit else
+                    child if lvl == limit else process_node(child))
 
         return process_node(tocroot)
 
     for name, level_item_map in maps.items():
         root = container.parsed(name)
-        item_level_map = {e:i for i, elems in level_item_map.items() for e in elems}
+        item_level_map = {e: i for i, elems in level_item_map.items()
+                          for e in elems}
         item_dirtied = False
         all_ids = set(root.xpath('//*/@id'))
 
@@ -470,7 +493,7 @@ def from_links(container):
     Generate a Table of Contents from links in the book.
     '''
     toc = TOC()
-    link_path = XPath('//h:a[@href]')
+    link_path = base.XPath('//h:a[@href]')
     seen_titles, seen_dests = set(), set()
     for name, is_linear in container.spine_names:
         root = container.parsed(name)
@@ -506,7 +529,7 @@ def find_text(node):
     pat = re.compile(r'\s+')
     for child in node:
         if isinstance(child, etree._Element):
-            text = xml2text(child).strip()
+            text = base.xml2text(child).strip()
             text = pat.sub(' ', text)
             if len(text) < 1:
                 continue
@@ -526,7 +549,7 @@ def from_files(container):
     for i, spinepath in enumerate(container.spine_items):
         name = container.abspath_to_name(spinepath)
         root = container.parsed(name)
-        body = XPath('//h:body')(root)
+        body = base.XPath('//h:body')(root)
         if not body:
             continue
         text = find_text(body[0])
@@ -576,42 +599,46 @@ def add_id(container, name, loc, totals=None):
 
 def create_ncx(toc, to_href, btitle, lang, uid):
     lang = lang.replace('_', '-')
-    ncx = etree.Element(NCX('ncx'),
-        attrib={'version': '2005-1', XML('lang'): lang},
-        nsmap={None: NCX_NS})
-    head = etree.SubElement(ncx, NCX('head'))
-    etree.SubElement(head, NCX('meta'),
-        name='dtb:uid', content=str(uid))
-    etree.SubElement(head, NCX('meta'),
-        name='dtb:depth', content=str(toc.depth))
+    ncx = etree.Element(base.tag('ncx', 'ncx'),
+                        attrib={'version': '2005-1',
+                                base.tag('xml', 'lang'): lang},
+                        nsmap={None: const.NCX_NS})
+    head = etree.SubElement(ncx, base.tag('ncx', 'head'))
+    etree.SubElement(head, base.tag('ncx', 'meta'),
+                     name='dtb:uid', content=str(uid))
+    etree.SubElement(head, base.tag('ncx', 'meta'),
+                     name='dtb:depth', content=str(toc.depth))
     generator = ''.join(['calibre (', __version__, ')'])
-    etree.SubElement(head, NCX('meta'),
-        name='dtb:generator', content=generator)
-    etree.SubElement(head, NCX('meta'), name='dtb:totalPageCount', content='0')
-    etree.SubElement(head, NCX('meta'), name='dtb:maxPageNumber', content='0')
-    title = etree.SubElement(ncx, NCX('docTitle'))
-    text = etree.SubElement(title, NCX('text'))
+    etree.SubElement(head, base.tag('ncx', 'meta'),
+                     name='dtb:generator', content=generator)
+    etree.SubElement(head, base.tag('ncx', 'meta'), name='dtb:totalPageCount',
+                     content='0')
+    etree.SubElement(head, base.tag('ncx', 'meta'), name='dtb:maxPageNumber',
+                     content='0')
+    title = etree.SubElement(ncx, base.tag('ncx', 'docTitle'))
+    text = etree.SubElement(title, base.tag('ncx', 'text'))
     text.text = btitle
-    navmap = etree.SubElement(ncx, NCX('navMap'))
+    navmap = etree.SubElement(ncx, base.tag('ncx', 'navMap'))
     spat = re.compile(r'\s+')
 
-    play_order = Counter()
+    play_order = collections.Counter()
 
     def process_node(xml_parent, toc_parent):
         for child in toc_parent:
             play_order['c'] += 1
-            point = etree.SubElement(xml_parent, NCX('navPoint'), id='num_%d' % play_order['c'],
-                            playOrder=str(play_order['c']))
-            label = etree.SubElement(point, NCX('navLabel'))
+            point = etree.SubElement(xml_parent, base.tag('ncx', 'navPoint'),
+                                     id='num_%d' % play_order['c'],
+                                     playOrder=str(play_order['c']))
+            label = etree.SubElement(point, base.tag('ncx', 'navLabel'))
             title = child.title
             if title:
                 title = spat.sub(' ', title)
-            etree.SubElement(label, NCX('text')).text = title
+            etree.SubElement(label, base.tag('ncx', 'text')).text = title
             if child.dest:
                 href = to_href(child.dest)
                 if child.frag:
                     href += '#'+child.frag
-                etree.SubElement(point, NCX('content'), src=href)
+                etree.SubElement(point, base.tag('ncx', 'content'), src=href)
             process_node(point, child)
 
     process_node(navmap, toc)
@@ -622,41 +649,43 @@ def commit_ncx_toc(container, toc, lang=None, uid=None):
     tocname = find_existing_ncx_toc(container)
     if tocname is None:
         item = container.generate_item('toc.ncx', id_prefix='toc')
-        tocname = container.href_to_name(item.get('href'), base=container.opf_name)
+        tocname = container.href_to_name(item.get('href'),
+                                         base=container.opf_name)
         ncx_id = item.get('id')
         [s.set('toc', ncx_id) for s in container.opf_xpath('//opf:spine')]
     if not lang:
         lang = get_lang()
-        for l in container.opf_xpath('//dc:language'):
-            l = canonicalize_lang(xml2text(l).strip())
-            if l:
-                lang = l
-                lang = lang_as_iso639_1(l) or l
+        for _l in container.opf_xpath('//dc:language'):
+            _l = canonicalize_lang(base.xml2text(_l).strip())
+            if _l:
+                lang = _l
+                lang = lang_as_iso639_1(_l) or _l
                 break
     lang = lang_as_iso639_1(lang) or lang
     if not uid:
-        uid = uuid_id()
+        uid = base.uuid_id()
         eid = container.opf.get('unique-identifier', None)
         if eid:
-            m = container.opf_xpath('//*[@id="%s"]'%eid)
+            m = container.opf_xpath('//*[@id="%s"]' % eid)
             if m:
-                uid = xml2text(m[0])
+                uid = base.xml2text(m[0])
 
     title = 'Table of Contents'
     m = container.opf_xpath('//dc:title')
     if m:
-        x = xml2text(m[0]).strip()
+        x = base.xml2text(m[0]).strip()
         title = x or title
 
-    to_href = partial(container.name_to_href, base=tocname)
+    to_href = functools.partial(container.name_to_href, base=tocname)
     root = create_ncx(toc, to_href, title, lang, uid)
     container.replace(tocname, root)
     container.pretty_print.add(tocname)
 
 
 def ensure_single_nav_of_type(root, ntype='toc'):
-    et = '{%s}type' % EPUB_NS
-    navs = [n for n in root.iterdescendants(XHTML('nav')) if n.get(et) == ntype]
+    et = base('epub', 'type')
+    navs = [n for n in root.iterdescendants(base.tag('xhtml', 'nav'))
+            if n.get(et) == ntype]
     for x in navs[1:]:
         extract(x)
     if navs:
@@ -667,13 +696,14 @@ def ensure_single_nav_of_type(root, ntype='toc'):
         nav.attrib.update(attrib)
         nav.tail = tail
     else:
-        nav = root.makeelement(XHTML('nav'))
-        first_child(root, XHTML('body')).append(nav)
-    nav.set('{%s}type' % EPUB_NS, ntype)
+        nav = root.makeelement(base.tag('xhtml', 'nav'))
+        first_child(root, base.tag('xhtml', 'body')).append(nav)
+    nav.set(et, ntype)
     return nav
 
 
-def commit_nav_toc(container, toc, lang=None, landmarks=None, previous_nav=None):
+def commit_nav_toc(container, toc, lang=None, landmarks=None,
+                   previous_nav=None):
     from ebook_converter.ebooks.oeb.polish.pretty import pretty_xml_tree
     tocname = find_existing_nav_toc(container)
     if previous_nav is not None:
@@ -684,7 +714,8 @@ def commit_nav_toc(container, toc, lang=None, landmarks=None, previous_nav=None)
     if tocname is None:
         item = container.generate_item('nav.xhtml', id_prefix='nav')
         item.set('properties', 'nav')
-        tocname = container.href_to_name(item.get('href'), base=container.opf_name)
+        tocname = container.href_to_name(item.get('href'),
+                                         base=container.opf_name)
         if previous_nav is not None:
             root = previous_nav[1]
         else:
@@ -698,24 +729,25 @@ def commit_nav_toc(container, toc, lang=None, landmarks=None, previous_nav=None)
     if lang:
         lang = lang_as_iso639_1(lang) or lang
         root.set('lang', lang)
-        root.set('{%s}lang' % XML_NS, lang)
+        root.set(base.tag('xml', 'lang'), lang)
     nav = ensure_single_nav_of_type(root, 'toc')
     if toc.toc_title:
-        nav.append(nav.makeelement(XHTML('h1')))
+        nav.append(nav.makeelement(base.tag('xhtml', 'h1')))
         nav[-1].text = toc.toc_title
 
-    rnode = nav.makeelement(XHTML('ol'))
+    rnode = nav.makeelement(base.tag('xhtml', 'ol'))
     nav.append(rnode)
-    to_href = partial(container.name_to_href, base=tocname)
+    to_href = functools.partial(container.name_to_href, base=tocname)
     spat = re.compile(r'\s+')
 
     def process_node(xml_parent, toc_parent):
         for child in toc_parent:
-            li = xml_parent.makeelement(XHTML('li'))
+            li = xml_parent.makeelement(base.tag('xhtml', 'li'))
             xml_parent.append(li)
             title = child.title or ''
             title = spat.sub(' ', title).strip()
-            a = li.makeelement(XHTML('a' if child.dest else 'span'))
+            a = li.makeelement(base.tag('xhtml', 'a'
+                                        if child.dest else 'span'))
             a.text = title
             li.append(a)
             if child.dest:
@@ -724,14 +756,14 @@ def commit_nav_toc(container, toc, lang=None, landmarks=None, previous_nav=None)
                     href += '#'+child.frag
                 a.set('href', href)
             if len(child):
-                ol = li.makeelement(XHTML('ol'))
+                ol = li.makeelement(base.tag('xhtml', 'ol'))
                 li.append(ol)
                 process_node(ol, child)
     process_node(rnode, toc)
     pretty_xml_tree(nav)
 
     def collapse_li(parent):
-        for li in parent.iterdescendants(XHTML('li')):
+        for li in parent.iterdescendants(base.tag('xhtml', 'li')):
             if len(li) == 1:
                 li.text = None
                 li[0].tail = None
@@ -739,9 +771,9 @@ def commit_nav_toc(container, toc, lang=None, landmarks=None, previous_nav=None)
     nav.tail = '\n'
 
     def create_li(ol, entry):
-        li = ol.makeelement(XHTML('li'))
+        li = ol.makeelement(base.tag('xhtml', 'li'))
         ol.append(li)
-        a = li.makeelement(XHTML('a'))
+        a = li.makeelement(base.tag('xhtml', 'a'))
         li.append(a)
         href = container.name_to_href(entry['dest'], tocname)
         if entry['frag']:
@@ -752,12 +784,13 @@ def commit_nav_toc(container, toc, lang=None, landmarks=None, previous_nav=None)
     if landmarks is not None:
         nav = ensure_single_nav_of_type(root, 'landmarks')
         nav.set('hidden', '')
-        ol = nav.makeelement(XHTML('ol'))
+        ol = nav.makeelement(base.tag('xhtml', 'ol'))
         nav.append(ol)
         for entry in landmarks:
-            if entry['type'] and container.has_name(entry['dest']) and container.mime_map[entry['dest']] in OEB_DOCS:
+            if (entry['type'] and container.has_name(entry['dest']) and
+                    container.mime_map[entry['dest']] in base.OEB_DOCS):
                 a = create_li(ol, entry)
-                a.set('{%s}type' % EPUB_NS, entry['type'])
+                a.set(base.tag('epub', 'type'), entry['type'])
                 a.text = entry['title'] or None
         pretty_xml_tree(nav)
         collapse_li(nav)
@@ -765,10 +798,11 @@ def commit_nav_toc(container, toc, lang=None, landmarks=None, previous_nav=None)
     if toc.page_list:
         nav = ensure_single_nav_of_type(root, 'page-list')
         nav.set('hidden', '')
-        ol = nav.makeelement(XHTML('ol'))
+        ol = nav.makeelement(base.tag('xhtml', 'ol'))
         nav.append(ol)
         for entry in toc.page_list:
-            if container.has_name(entry['dest']) and container.mime_map[entry['dest']] in OEB_DOCS:
+            if (container.has_name(entry['dest']) and
+                    container.mime_map[entry['dest']] in base.OEB_DOCS):
                 a = create_li(ol, entry)
                 a.text = str(entry['pagenum'])
         pretty_xml_tree(nav)
@@ -785,11 +819,12 @@ def commit_toc(container, toc, lang=None, uid=None):
 def remove_names_from_toc(container, names):
     changed = []
     names = frozenset(names)
-    for find_toc, parse_toc, commit_toc in (
-            (find_existing_ncx_toc, parse_ncx, commit_ncx_toc),
-            (find_existing_nav_toc, parse_nav, commit_nav_toc),
-    ):
-        toc = get_x_toc(container, find_toc, parse_toc, verify_destinations=False)
+    for find_toc, parse_toc, commit_toc in ((find_existing_ncx_toc,
+                                             parse_ncx, commit_ncx_toc),
+                                            (find_existing_nav_toc,
+                                             parse_nav, commit_nav_toc)):
+        toc = get_x_toc(container, find_toc, parse_toc,
+                        verify_destinations=False)
         if len(toc) > 0:
             remove = []
             for node in toc.iterdescendants():
@@ -805,15 +840,16 @@ def remove_names_from_toc(container, names):
 
 def find_inline_toc(container):
     for name, linear in container.spine_names:
-        if container.parsed(name).xpath('//*[local-name()="body" and @id="calibre_generated_inline_toc"]'):
+        if container.parsed(name).xpath('//*[local-name()="body" and @id='
+                                        '"calibre_generated_inline_toc"]'):
             return name
 
 
 def toc_to_html(toc, container, toc_name, title, lang=None):
 
     def process_node(html_parent, toc, level=1, indent='  ', style_level=2):
-        li = html_parent.makeelement(XHTML('li'))
-        li.tail = '\n'+ (indent*level)
+        li = html_parent.makeelement(base.tag('xhtml', 'li'))
+        li.tail = '\n' + (indent * level)
         html_parent.append(li)
         name, frag = toc.dest, toc.frag
         href = '#'
@@ -821,32 +857,29 @@ def toc_to_html(toc, container, toc_name, title, lang=None):
             href = container.name_to_href(name, toc_name)
             if frag:
                 href += '#' + frag
-        a = li.makeelement(XHTML('a'), href=href)
+        a = li.makeelement(base.tag('xhtml', 'a'), href=href)
         a.text = toc.title
         li.append(a)
         if len(toc) > 0:
-            parent = li.makeelement(XHTML('ul'))
+            parent = li.makeelement(base.tag('xhtml', 'ul'))
             parent.set('class', 'level%d' % (style_level))
             li.append(parent)
             a.tail = '\n\n' + (indent*(level+2))
             parent.text = '\n'+(indent*(level+3))
             parent.tail = '\n\n' + (indent*(level+1))
             for child in toc:
-                process_node(parent, child, level+3, style_level=style_level + 1)
+                process_node(parent, child, level+3,
+                             style_level=style_level + 1)
             parent[-1].tail = '\n' + (indent*(level+2))
 
-    E = ElementMaker(namespace=XHTML_NS, nsmap={None:XHTML_NS})
-    html = E.html(
-        E.head(
-            E.title(title),
-            E.style(P('templates/inline_toc_styles.css', data=True), type='text/css'),
-        ),
-        E.body(
-            E.h2(title),
-            E.ul(),
-            id="calibre_generated_inline_toc",
-        )
-    )
+    E = ElementMaker(namespace=const.XHTML_NS, nsmap={None: const.XHTML_NS})
+    # TODO(gryf): revisit lack of css.
+    css_f = pkg_resources.resource_filename('ebook_converter',
+                                            'data/inline_toc_styles.css')
+    html = E.html(E.head(E.title(title),
+                         E.style(css_f, type='text/css')),
+                  E.body(E.h2(title), E.ul(),
+                         id="calibre_generated_inline_toc"))
 
     ul = html[1][1]
     ul.set('class', 'level1')
@@ -859,11 +892,12 @@ def toc_to_html(toc, container, toc_name, title, lang=None):
 
 
 def create_inline_toc(container, title=None):
-    '''
-    Create an inline (HTML) Table of Contents from an existing NCX Table of Contents.
+    """
+    Create an inline (HTML) Table of Contents from an existing NCX Table of
+    Contents.
 
     :param title: The title for this table of contents.
-    '''
+    """
     lang = get_book_language(container)
     default_title = 'Table of Contents'
     title = title or default_title
@@ -874,7 +908,7 @@ def create_inline_toc(container, title=None):
 
     name = toc_name
     html = toc_to_html(toc, container, name, title, lang)
-    raw = serialize(html, 'text/html')
+    raw = base.serialize(html, 'text/html')
     if name is None:
         name, c = 'toc.xhtml', 0
         while container.has_name(name):
@@ -884,5 +918,6 @@ def create_inline_toc(container, title=None):
     else:
         with container.open(name, 'wb') as f:
             f.write(raw)
-    set_guide_item(container, 'toc', title, name, frag='calibre_generated_inline_toc')
+    set_guide_item(container, 'toc', title, name,
+                   frag='calibre_generated_inline_toc')
     return name
diff --git a/ebook_converter/ebooks/oeb/reader.py b/ebook_converter/ebooks/oeb/reader.py
index 4139b1b..f5a02f7 100644
--- a/ebook_converter/ebooks/oeb/reader.py
+++ b/ebook_converter/ebooks/oeb/reader.py
@@ -1,21 +1,21 @@
 """
 Container-/OPF-based input OEBBook reader.
 """
-import sys, os, uuid, copy, re, io
-from collections import defaultdict
+import collections
+import copy
+import io
+import os
+import re
+import sys
 import urllib.parse
+import uuid
 
 from lxml import etree
 
-from ebook_converter.ebooks.oeb.base import OPF1_NS, OPF2_NS, OPF2_NSMAP, DC11_NS, \
-    DC_NSES, OPF, xml2text, XHTML_MIME
-from ebook_converter.ebooks.oeb.base import OEB_DOCS, OEB_STYLES, OEB_IMAGES, \
-    PAGE_MAP_MIME, JPEG_MIME, NCX_MIME, SVG_MIME
-from ebook_converter.ebooks.oeb.base import XMLDECL_RE, COLLAPSE_RE, \
-    MS_COVER_TYPE, iterlinks
-from ebook_converter.ebooks.oeb.base import namespace, barename, XPath, xpath, \
-                                    urlnormalize, BINARY_MIME, \
-                                    OEBError, OEBBook, DirContainer
+from ebook_converter import constants as const
+from ebook_converter.ebooks.oeb import base
+from ebook_converter.ebooks.oeb import parse_utils
+from ebook_converter.ebooks.metadata import opf2 as opf_meta
 from ebook_converter.ebooks.oeb.writer import OEBWriter
 from ebook_converter.utils.xml_parse import safe_xml_fromstring
 from ebook_converter.utils.cleantext import clean_xml_chars
@@ -26,18 +26,13 @@ from ebook_converter import guess_type, xml_replace_entities
 from ebook_converter.polyglot.urllib import unquote
 
 
-__all__ = ['OEBReader']
-__license__ = 'GPL v3'
-__copyright__ = '2008, Marshall T. Vandegrift '
-
-
 class OEBReader(object):
     """Read an OEBPS 1.x or OPF/OPS 2.0 file collection."""
 
-    COVER_SVG_XP    = XPath('h:body//svg:svg[position() = 1]')
-    COVER_OBJECT_XP = XPath('h:body//h:object[@data][position() = 1]')
+    COVER_SVG_XP = base.XPath('h:body//svg:svg[position() = 1]')
+    COVER_OBJECT_XP = base.XPath('h:body//h:object[@data][position() = 1]')
 
-    Container = DirContainer
+    Container = base.DirContainer
     """Container type used to access book files.  Override in sub-classes."""
 
     DEFAULT_PROFILE = 'PRS505'
@@ -75,61 +70,67 @@ class OEBReader(object):
         for elem in opf.iter(tag=etree.Element):
             nsmap.update(elem.nsmap)
         for elem in opf.iter(tag=etree.Element):
-            if namespace(elem.tag) in ('', OPF1_NS) and ':' not in barename(elem.tag):
-                elem.tag = OPF(barename(elem.tag))
-        nsmap.update(OPF2_NSMAP)
+            if (parse_utils.namespace(elem.tag) in ('', const.OPF1_NS) and
+                    ':' not in parse_utils.barename(elem.tag)):
+                elem.tag = base.tag('opf', parse_utils.barename(elem.tag))
+        nsmap.update(const.OPF2_NSMAP)
         attrib = dict(opf.attrib)
-        nroot = etree.Element(OPF('package'),
-            nsmap={None: OPF2_NS}, attrib=attrib)
-        metadata = etree.SubElement(nroot, OPF('metadata'), nsmap=nsmap)
-        ignored = (OPF('dc-metadata'), OPF('x-metadata'))
-        for elem in xpath(opf, 'o2:metadata//*'):
+        nroot = etree.Element(base.tag('opf', 'package'),
+                              nsmap={None: const.OPF2_NS}, attrib=attrib)
+        metadata = etree.SubElement(nroot, base.tag('opf', 'metadata'),
+                                    nsmap=nsmap)
+        ignored = (base.tag('opf', 'dc-metadata'), base.tag('opf', 'x-metadata'))
+        for elem in base.xpath(opf, 'o2:metadata//*'):
             if elem.tag in ignored:
                 continue
-            if namespace(elem.tag) in DC_NSES:
-                tag = barename(elem.tag).lower()
-                elem.tag = '{%s}%s' % (DC11_NS, tag)
+            if parse_utils.namespace(elem.tag) in const.DC_NSES:
+                tag = parse_utils.barename(elem.tag).lower()
+                elem.tag = '{%s}%s' % (const.DC11_NS, tag)
             if elem.tag.startswith('dc:'):
                 tag = elem.tag.partition(':')[-1].lower()
-                elem.tag = '{%s}%s' % (DC11_NS, tag)
+                elem.tag = '{%s}%s' % (const.DC11_NS, tag)
             metadata.append(elem)
-        for element in xpath(opf, 'o2:metadata//o2:meta'):
+        for element in base.xpath(opf, 'o2:metadata//o2:meta'):
             metadata.append(element)
         for tag in ('o2:manifest', 'o2:spine', 'o2:tours', 'o2:guide'):
-            for element in xpath(opf, tag):
+            for element in base.xpath(opf, tag):
                 nroot.append(element)
         return nroot
 
     def _read_opf(self):
         data = self.oeb.container.read(None)
         data = self.oeb.decode(data)
-        data = XMLDECL_RE.sub('', data)
+        data = base.XMLDECL_RE.sub('', data)
         data = re.sub(r'http://openebook.org/namespaces/oeb-package/1.0(/*)',
-                OPF1_NS, data)
+                      const.OPF1_NS, data)
         try:
-            opf = safe_xml_fromstring(data)
+            opf = etree.fromstring(data)
         except etree.XMLSyntaxError:
             data = xml_replace_entities(clean_xml_chars(data), encoding=None)
             try:
-                opf = safe_xml_fromstring(data)
+                opf = etree.fromstring(data)
                 self.logger.warn('OPF contains invalid HTML named entities')
             except etree.XMLSyntaxError:
                 data = re.sub(r'(?is).+', '', data)
                 data = data.replace('',
-                    '')
-                opf = safe_xml_fromstring(data)
+                                    '')
+                opf = etree.fromstring(data)
                 self.logger.warn('OPF contains invalid tours section')
 
-        ns = namespace(opf.tag)
-        if ns not in ('', OPF1_NS, OPF2_NS):
-            raise OEBError('Invalid namespace %r for OPF document' % ns)
+        ns = parse_utils.namespace(opf.tag)
+        if ns not in ('', const.OPF1_NS, const.OPF2_NS):
+            raise base.OEBError('Invalid namespace %r for OPF document' % ns)
         opf = self._clean_opf(opf)
         return opf
 
     def _metadata_from_opf(self, opf):
         from ebook_converter.ebooks.metadata.opf2 import OPF
-        from ebook_converter.ebooks.oeb.transforms.metadata import meta_info_to_oeb_metadata
-        stream = io.BytesIO(etree.tostring(opf, xml_declaration=True, encoding='utf-8'))
+        from ebook_converter.ebooks.oeb.transforms.metadata import \
+            meta_info_to_oeb_metadata
+        stream = io.BytesIO(etree.tostring(opf, xml_declaration=True,
+                                           encoding='utf-8'))
+        # o = opf_meta.OPF(stream)
         o = OPF(stream)
         pwm = o.primary_writing_mode
         if pwm:
@@ -139,8 +140,8 @@ class OEBReader(object):
             mi.language = get_lang().replace('_', '-')
         self.oeb.metadata.add('language', mi.language)
         if not mi.book_producer:
-            mi.book_producer = '%(a)s (%(v)s) [http://%(a)s-ebook.com]'%\
-                dict(a=__appname__, v=__version__)
+            mi.book_producer = ('%(a)s (%(v)s) [http://%(a)s-ebook.com]' %
+                                dict(a=__appname__, v=__version__))
         meta_info_to_oeb_metadata(mi, self.oeb.metadata, self.logger)
         m = self.oeb.metadata
         m.add('identifier', str(uuid.uuid4()), id='uuid_id', scheme='uuid')
@@ -162,16 +163,16 @@ class OEBReader(object):
         data.
         '''
         bad = []
-        check = OEB_DOCS.union(OEB_STYLES)
+        check = base.OEB_DOCS.union(base.OEB_STYLES)
         for item in list(self.oeb.manifest.values()):
             if item.media_type in check:
                 try:
                     item.data
                 except KeyboardInterrupt:
                     raise
-                except:
-                    self.logger.exception('Failed to parse content in %s'%
-                            item.href)
+                except Exception:
+                    self.logger.exception('Failed to parse content in %s' %
+                                          item.href)
                     bad.append(item)
                     self.oeb.manifest.remove(item)
         return bad
@@ -181,25 +182,28 @@ class OEBReader(object):
         manifest = self.oeb.manifest
         known = set(manifest.hrefs)
         unchecked = set(manifest.values())
-        cdoc = OEB_DOCS|OEB_STYLES
+        cdoc = base.OEB_DOCS | base.OEB_STYLES
         invalid = set()
         while unchecked:
             new = set()
             for item in unchecked:
                 data = None
-                if (item.media_type in cdoc or item.media_type[-4:] in ('/xml', '+xml')):
+                if (item.media_type in cdoc or
+                        item.media_type[-4:] in ('/xml', '+xml')):
                     try:
                         data = item.data
-                    except:
+                    except Exception:
                         self.oeb.log.exception('Failed to read from manifest '
-                                'entry with id: %s, ignoring'%item.id)
+                                               'entry with id: %s, ignoring' %
+                                               item.id)
                         invalid.add(item)
                         continue
                 if data is None:
                     continue
 
-                if (item.media_type in OEB_DOCS or item.media_type[-4:] in ('/xml', '+xml')):
-                    hrefs = [r[2] for r in iterlinks(data)]
+                if (item.media_type in base.OEB_DOCS or
+                        item.media_type[-4:] in ('/xml', '+xml')):
+                    hrefs = [r[2] for r in base.iterlinks(data)]
                     for href in hrefs:
                         if isinstance(href, bytes):
                             href = href.decode('utf-8')
@@ -207,22 +211,22 @@ class OEBReader(object):
                         if not href:
                             continue
                         try:
-                            href = item.abshref(urlnormalize(href))
+                            href = item.abshref(base.urlnormalize(href))
                             scheme = urllib.parse.urlparse(href).scheme
-                        except:
-                            self.oeb.log.exception(
-                                'Skipping invalid href: %r'%href)
+                        except Exception:
+                            self.oeb.log.exception('Skipping invalid href: '
+                                                   '%r' % href)
                             continue
                         if not scheme and href not in known:
                             new.add(href)
-                elif item.media_type in OEB_STYLES:
+                elif item.media_type in base.OEB_STYLES:
                     try:
                         urls = list(css_parser.getUrls(data))
-                    except:
+                    except Exception:
                         urls = []
                     for url in urls:
                         href, _ = urllib.parse.urldefrag(url)
-                        href = item.abshref(urlnormalize(href))
+                        href = item.abshref(base.urlnormalize(href))
                         scheme = urllib.parse.urlparse(href).scheme
                         if not scheme and href not in known:
                             new.add(href)
@@ -232,7 +236,7 @@ class OEBReader(object):
                 known.add(href)
                 is_invalid = False
                 for item in invalid:
-                    if href == item.abshref(urlnormalize(href)):
+                    if href == item.abshref(base.urlnormalize(href)):
                         is_invalid = True
                         break
                 if is_invalid:
@@ -243,11 +247,12 @@ class OEBReader(object):
                         warned.add(href)
                     continue
                 if href not in warned:
-                    self.logger.warn('Referenced file %r not in manifest' % href)
+                    self.logger.warn('Referenced file %r not in manifest' %
+                                     href)
                     warned.add(href)
                 id, _ = manifest.generate(id='added')
                 guessed = guess_type(href)[0]
-                media_type = guessed or BINARY_MIME
+                media_type = guessed or base.BINARY_MIME
                 added = manifest.add(id, href, media_type)
                 unchecked.add(added)
 
@@ -256,7 +261,7 @@ class OEBReader(object):
 
     def _manifest_from_opf(self, opf):
         manifest = self.oeb.manifest
-        for elem in xpath(opf, '/o2:package/o2:manifest/o2:item'):
+        for elem in base.xpath(opf, '/o2:package/o2:manifest/o2:item'):
             id = elem.get('id')
             href = elem.get('href')
             media_type = elem.get('media-type', None)
@@ -264,7 +269,7 @@ class OEBReader(object):
                 media_type = elem.get('mediatype', None)
             if not media_type or media_type == 'text/xml':
                 guessed = guess_type(href)[0]
-                media_type = guessed or media_type or BINARY_MIME
+                media_type = guessed or media_type or base.BINARY_MIME
             if hasattr(media_type, 'lower'):
                 media_type = media_type.lower()
             fallback = elem.get('fallback')
@@ -285,12 +290,12 @@ class OEBReader(object):
         manifest = self.oeb.manifest
         spine = self.oeb.spine
         unchecked = set(spine)
-        selector = XPath('h:body//h:a/@href')
+        selector = base.XPath('h:body//h:a/@href')
         extras = set()
         while unchecked:
             new = set()
             for item in unchecked:
-                if item.media_type not in OEB_DOCS:
+                if item.media_type not in base.OEB_DOCS:
                     # TODO: handle fallback chains
                     continue
                 for href in selector(item.data):
@@ -298,20 +303,21 @@ class OEBReader(object):
                     if not href:
                         continue
                     try:
-                        href = item.abshref(urlnormalize(href))
+                        href = item.abshref(base.urlnormalize(href))
                     except ValueError:  # Malformed URL
                         continue
                     if href not in manifest.hrefs:
                         continue
                     found = manifest.hrefs[href]
-                    if found.media_type not in OEB_DOCS or \
+                    if found.media_type not in base.OEB_DOCS or \
                        found in spine or found in extras:
                         continue
                     new.add(found)
             extras.update(new)
             unchecked = new
         version = int(self.oeb.version[0])
-        removed_items_to_ignore = getattr(self.oeb, 'removed_items_to_ignore', ())
+        removed_items_to_ignore = getattr(self.oeb, 'removed_items_to_ignore',
+                                          ())
         for item in sorted(extras):
             if item.href in removed_items_to_ignore:
                 continue
@@ -323,34 +329,38 @@ class OEBReader(object):
     def _spine_from_opf(self, opf):
         spine = self.oeb.spine
         manifest = self.oeb.manifest
-        for elem in xpath(opf, '/o2:package/o2:spine/o2:itemref'):
+        for elem in base.xpath(opf, '/o2:package/o2:spine/o2:itemref'):
             idref = elem.get('idref')
             if idref not in manifest.ids:
                 self.logger.warn('Spine item %r not found' % idref)
                 continue
             item = manifest.ids[idref]
-            if item.media_type.lower() in OEB_DOCS and hasattr(item.data, 'xpath') and not getattr(item.data, 'tag', '').endswith('}ncx'):
+            if (item.media_type.lower() in base.OEB_DOCS and
+                    hasattr(item.data, 'xpath') and not
+                    getattr(item.data, 'tag', '').endswith('}ncx')):
                 spine.add(item, elem.get('linear'))
             else:
-                if hasattr(item.data, 'tag') and item.data.tag and item.data.tag.endswith('}html'):
-                    item.media_type = XHTML_MIME
+                if (hasattr(item.data, 'tag') and
+                        item.data.tag and item.data.tag.endswith('}html')):
+                    item.media_type = base.XHTML_MIME
                     spine.add(item, elem.get('linear'))
                 else:
                     self.oeb.log.warn('The item %s is not a XML document.'
-                        ' Removing it from spine.'%item.href)
+                                      ' Removing it from spine.' % item.href)
         if len(spine) == 0:
-            raise OEBError("Spine is empty")
+            raise base.OEBError("Spine is empty")
         self._spine_add_extra()
-        for val in xpath(opf, '/o2:package/o2:spine/@page-progression-direction'):
+        for val in base.xpath(opf,
+                         '/o2:package/o2:spine/@page-progression-direction'):
             if val in {'ltr', 'rtl'}:
                 spine.page_progression_direction = val
 
     def _guide_from_opf(self, opf):
         guide = self.oeb.guide
         manifest = self.oeb.manifest
-        for elem in xpath(opf, '/o2:package/o2:guide/o2:reference'):
+        for elem in base.xpath(opf, '/o2:package/o2:guide/o2:reference'):
             ref_href = elem.get('href')
-            path = urlnormalize(urllib.parse.urldefrag(ref_href)[0])
+            path = base.urlnormalize(urllib.parse.urldefrag(ref_href)[0])
             if path not in manifest.hrefs:
                 corrected_href = None
                 for href in manifest.hrefs:
@@ -366,7 +376,7 @@ class OEBReader(object):
                 guide.add(typ, elem.get('title'), ref_href)
 
     def _find_ncx(self, opf):
-        result = xpath(opf, '/o2:package/o2:spine/@toc')
+        result = base.xpath(opf, '/o2:package/o2:spine/@toc')
         if result:
             id = result[0]
             if id not in self.oeb.manifest.ids:
@@ -375,30 +385,33 @@ class OEBReader(object):
             self.oeb.manifest.remove(item)
             return item
         for item in self.oeb.manifest.values():
-            if item.media_type == NCX_MIME:
+            if item.media_type == base.NCX_MIME:
                 self.oeb.manifest.remove(item)
                 return item
         return None
 
     def _toc_from_navpoint(self, item, toc, navpoint):
-        children = xpath(navpoint, 'ncx:navPoint')
+        children = base.xpath(navpoint, 'ncx:navPoint')
         for child in children:
-            title = ''.join(xpath(child, 'ncx:navLabel/ncx:text/text()'))
-            title = COLLAPSE_RE.sub(' ', title.strip())
-            href = xpath(child, 'ncx:content/@src')
+            title = ''.join(base.xpath(child, 'ncx:navLabel/ncx:text/text()'))
+            title = base.COLLAPSE_RE.sub(' ', title.strip())
+            href = base.xpath(child, 'ncx:content/@src')
             if not title:
                 self._toc_from_navpoint(item, toc, child)
                 continue
-            if (not href or not href[0]) and not xpath(child, 'ncx:navPoint'):
+            if (not href or not href[0]) and not base.xpath(child, 'ncx:navPoint'):
                 # This node is useless
                 continue
-            href = item.abshref(urlnormalize(href[0])) if href and href[0] else ''
+            if href and href[0]:
+                href = item.abshref(base.urlnormalize(href[0]))
+            else:
+                href = ''
             path, _ = urllib.parse.urldefrag(href)
             if path and path not in self.oeb.manifest.hrefs:
-                path = urlnormalize(path)
+                path = base.urlnormalize(path)
             if href and path not in self.oeb.manifest.hrefs:
                 self.logger.warn('TOC reference %r not found' % href)
-                gc = xpath(child, 'ncx:navPoint')
+                gc = base.xpath(child, 'ncx:navPoint')
                 if not gc:
                     # This node is useless
                     continue
@@ -406,36 +419,40 @@ class OEBReader(object):
             klass = child.get('class', 'chapter')
 
             try:
-                po = int(child.get('playOrder', self.oeb.toc.next_play_order()))
-            except:
+                po = int(child.get('playOrder',
+                                   self.oeb.toc.next_play_order()))
+            except Exception:
                 po = self.oeb.toc.next_play_order()
 
-            authorElement = xpath(child,
-                    'descendant::calibre:meta[@name = "author"]')
+            authorElement = base.xpath(child,
+                                  'descendant::calibre:meta[@name = "author"]')
             if authorElement:
                 author = authorElement[0].text
             else:
                 author = None
 
-            descriptionElement = xpath(child,
-                    'descendant::calibre:meta[@name = "description"]')
+            descriptionElement = base.xpath(child,
+                                       'descendant::calibre:meta[@name = '
+                                       '"description"]')
             if descriptionElement:
                 description = etree.tostring(descriptionElement[0],
-                method='text', encoding='unicode').strip()
+                                             method='text',
+                                             encoding='unicode').strip()
                 if not description:
                     description = None
             else:
                 description = None
 
-            index_image = xpath(child,
-                    'descendant::calibre:meta[@name = "toc_thumbnail"]')
+            index_image = base.xpath(child,
+                                'descendant::calibre:meta[@name = '
+                                '"toc_thumbnail"]')
             toc_thumbnail = (index_image[0].text if index_image else None)
             if not toc_thumbnail or not toc_thumbnail.strip():
                 toc_thumbnail = None
 
             node = toc.add(title, href, id=id, klass=klass,
-                    play_order=po, description=description, author=author,
-                           toc_thumbnail=toc_thumbnail)
+                           play_order=po, description=description,
+                           author=author, toc_thumbnail=toc_thumbnail)
 
             self._toc_from_navpoint(item, node, child)
 
@@ -444,31 +461,31 @@ class OEBReader(object):
             return False
         self.log.debug('Reading TOC from NCX...')
         ncx = item.data
-        title = ''.join(xpath(ncx, 'ncx:docTitle/ncx:text/text()'))
-        title = COLLAPSE_RE.sub(' ', title.strip())
+        title = ''.join(base.xpath(ncx, 'ncx:docTitle/ncx:text/text()'))
+        title = base.COLLAPSE_RE.sub(' ', title.strip())
         title = title or str(self.oeb.metadata.title[0])
         toc = self.oeb.toc
         toc.title = title
-        navmaps = xpath(ncx, 'ncx:navMap')
+        navmaps = base.xpath(ncx, 'ncx:navMap')
         for navmap in navmaps:
             self._toc_from_navpoint(item, toc, navmap)
         return True
 
     def _toc_from_tour(self, opf):
-        result = xpath(opf, 'o2:tours/o2:tour')
+        result = base.xpath(opf, 'o2:tours/o2:tour')
         if not result:
             return False
         self.log.debug('Reading TOC from tour...')
         tour = result[0]
         toc = self.oeb.toc
         toc.title = tour.get('title')
-        sites = xpath(tour, 'o2:site')
+        sites = base.xpath(tour, 'o2:site')
         for site in sites:
             title = site.get('title')
             href = site.get('href')
             if not title or not href:
                 continue
-            path, _ = urllib.parse.urldefrag(urlnormalize(href))
+            path, _ = urllib.parse.urldefrag(base.urlnormalize(href))
             if path not in self.oeb.manifest.hrefs:
                 self.logger.warn('TOC reference %r not found' % href)
                 continue
@@ -484,23 +501,23 @@ class OEBReader(object):
         item = self.oeb.manifest.hrefs[itempath]
         html = item.data
         if frag:
-            elems = xpath(html, './/*[@id="%s"]' % frag)
+            elems = base.xpath(html, './/*[@id="%s"]' % frag)
             if not elems:
-                elems = xpath(html, './/*[@name="%s"]' % frag)
+                elems = base.xpath(html, './/*[@name="%s"]' % frag)
             elem = elems[0] if elems else html
-            while elem != html and not xpath(elem, './/h:a[@href]'):
+            while elem != html and not base.xpath(elem, './/h:a[@href]'):
                 elem = elem.getparent()
             html = elem
-        titles = defaultdict(list)
+        titles = collections.defaultdict(list)
         order = []
-        for anchor in xpath(html, './/h:a[@href]'):
+        for anchor in base.xpath(html, './/h:a[@href]'):
             href = anchor.attrib['href']
-            href = item.abshref(urlnormalize(href))
+            href = item.abshref(base.urlnormalize(href))
             path, frag = urllib.parse.urldefrag(href)
             if path not in self.oeb.manifest.hrefs:
                 continue
-            title = xml2text(anchor)
-            title = COLLAPSE_RE.sub(' ', title.strip())
+            title = base.xml2text(anchor)
+            title = base.COLLAPSE_RE.sub(' ', title.strip())
             if href not in titles:
                 order.append(href)
             titles[href].append(title)
@@ -518,15 +535,15 @@ class OEBReader(object):
             if not item.linear:
                 continue
             html = item.data
-            title = ''.join(xpath(html, '/h:html/h:head/h:title/text()'))
-            title = COLLAPSE_RE.sub(' ', title.strip())
+            title = ''.join(base.xpath(html, '/h:html/h:head/h:title/text()'))
+            title = base.COLLAPSE_RE.sub(' ', title.strip())
             if title:
                 titles.append(title)
             headers.append('(unlabled)')
             for tag in ('h1', 'h2', 'h3', 'h4', 'h5', 'strong'):
                 expr = '/h:html/h:body//h:%s[position()=1]/text()'
-                header = ''.join(xpath(html, expr % tag))
-                header = COLLAPSE_RE.sub(' ', header.strip())
+                header = ''.join(base.xpath(html, expr % tag))
+                header = base.COLLAPSE_RE.sub(' ', header.strip())
                 if header:
                     headers[-1] = header
                     break
@@ -558,17 +575,17 @@ class OEBReader(object):
         ncx = item.data
         if ncx is None:
             return False
-        ptargets = xpath(ncx, 'ncx:pageList/ncx:pageTarget')
+        ptargets = base.xpath(ncx, 'ncx:pageList/ncx:pageTarget')
         if not ptargets:
             return False
         pages = self.oeb.pages
         for ptarget in ptargets:
-            name = ''.join(xpath(ptarget, 'ncx:navLabel/ncx:text/text()'))
-            name = COLLAPSE_RE.sub(' ', name.strip())
-            href = xpath(ptarget, 'ncx:content/@src')
+            name = ''.join(base.xpath(ptarget, 'ncx:navLabel/ncx:text/text()'))
+            name = base.COLLAPSE_RE.sub(' ', name.strip())
+            href = base.xpath(ptarget, 'ncx:content/@src')
             if not href:
                 continue
-            href = item.abshref(urlnormalize(href[0]))
+            href = item.abshref(base.urlnormalize(href[0]))
             id = ptarget.get('id')
             type = ptarget.get('type', 'normal')
             klass = ptarget.get('class')
@@ -576,7 +593,7 @@ class OEBReader(object):
         return True
 
     def _find_page_map(self, opf):
-        result = xpath(opf, '/o2:package/o2:spine/@page-map')
+        result = base.xpath(opf, '/o2:package/o2:spine/@page-map')
         if result:
             id = result[0]
             if id not in self.oeb.manifest.ids:
@@ -585,7 +602,7 @@ class OEBReader(object):
             self.oeb.manifest.remove(item)
             return item
         for item in self.oeb.manifest.values():
-            if item.media_type == PAGE_MAP_MIME:
+            if item.media_type == base.PAGE_MAP_MIME:
                 self.oeb.manifest.remove(item)
                 return item
         return None
@@ -596,13 +613,13 @@ class OEBReader(object):
             return False
         pmap = item.data
         pages = self.oeb.pages
-        for page in xpath(pmap, 'o2:page'):
+        for page in base.xpath(pmap, 'o2:page'):
             name = page.get('name', '')
             href = page.get('href')
             if not href:
                 continue
-            name = COLLAPSE_RE.sub(' ', name.strip())
-            href = item.abshref(urlnormalize(href))
+            name = base.COLLAPSE_RE.sub(' ', name.strip())
+            href = item.abshref(base.urlnormalize(href))
             type = 'normal'
             if not name:
                 type = 'special'
@@ -628,14 +645,14 @@ class OEBReader(object):
             if not data:
                 data = b''
         id, href = self.oeb.manifest.generate('cover', 'cover.jpg')
-        item = self.oeb.manifest.add(id, href, JPEG_MIME, data=data)
+        item = self.oeb.manifest.add(id, href, base.JPEG_MIME, data=data)
         return item
 
     def _locate_cover_image(self):
         if self.oeb.metadata.cover:
             id = str(self.oeb.metadata.cover[0])
             item = self.oeb.manifest.ids.get(id, None)
-            if item is not None and item.media_type in OEB_IMAGES:
+            if item is not None and item.media_type in base.OEB_IMAGES:
                 return item
             else:
                 self.logger.warn('Invalid cover image @id %r' % id)
@@ -644,27 +661,27 @@ class OEBReader(object):
             href = self.oeb.guide['cover'].href
             item = self.oeb.manifest.hrefs[href]
             media_type = item.media_type
-            if media_type in OEB_IMAGES:
+            if media_type in base.OEB_IMAGES:
                 return item
-            elif media_type in OEB_DOCS:
+            elif media_type in base.OEB_DOCS:
                 hcover = item
         html = hcover.data
-        if MS_COVER_TYPE in self.oeb.guide:
-            href = self.oeb.guide[MS_COVER_TYPE].href
+        if base.MS_COVER_TYPE in self.oeb.guide:
+            href = self.oeb.guide[base.MS_COVER_TYPE].href
             item = self.oeb.manifest.hrefs.get(href, None)
-            if item is not None and item.media_type in OEB_IMAGES:
+            if item is not None and item.media_type in base.OEB_IMAGES:
                 return item
         if self.COVER_SVG_XP(html):
             svg = copy.deepcopy(self.COVER_SVG_XP(html)[0])
             href = os.path.splitext(hcover.href)[0] + '.svg'
             id, href = self.oeb.manifest.generate(hcover.id, href)
-            item = self.oeb.manifest.add(id, href, SVG_MIME, data=svg)
+            item = self.oeb.manifest.add(id, href, base.SVG_MIME, data=svg)
             return item
         if self.COVER_OBJECT_XP(html):
             object = self.COVER_OBJECT_XP(html)[0]
             href = hcover.abshref(object.get('data'))
             item = self.oeb.manifest.hrefs.get(href, None)
-            if item is not None and item.media_type in OEB_IMAGES:
+            if item is not None and item.media_type in base.OEB_IMAGES:
                 return item
         return self._cover_from_html(hcover)
 
@@ -687,7 +704,8 @@ class OEBReader(object):
             items = [x for x in self.oeb.manifest if x.href == href]
             for x in items:
                 if x not in self.oeb.spine:
-                    self.oeb.log.warn('Removing duplicate manifest item with id:', x.id)
+                    self.oeb.log.warn('Removing duplicate manifest item with '
+                                      'id:', x.id)
                     self.oeb.manifest.remove_duplicate_item(x)
 
     def _all_from_opf(self, opf):
@@ -706,7 +724,7 @@ class OEBReader(object):
 def main(argv=sys.argv):
     reader = OEBReader()
     for arg in argv[1:]:
-        oeb = reader(OEBBook(), arg)
+        oeb = reader(base.OEBBook(), arg)
         for name, doc in oeb.to_opf1().values():
             print(etree.tostring(doc, pretty_print=True))
         for name, doc in oeb.to_opf2(page_map=True).values():
diff --git a/ebook_converter/ebooks/oeb/stylizer.py b/ebook_converter/ebooks/oeb/stylizer.py
index 178579e..cfef868 100644
--- a/ebook_converter/ebooks/oeb/stylizer.py
+++ b/ebook_converter/ebooks/oeb/stylizer.py
@@ -10,17 +10,16 @@ from css_parser.css import (CSSStyleRule, CSSPageRule, CSSFontFaceRule,
         cssproperties)
 from css_parser import (profile as cssprofiles, parseString, parseStyle, log as
         css_parser_log, CSSParser, profiles, replaceUrls)
+
+from ebook_converter import constants as const
 from ebook_converter import force_unicode, as_unicode
 from ebook_converter.ebooks import unit_convert
-from ebook_converter.ebooks.oeb.base import XHTML, XHTML_NS, CSS_MIME, OEB_STYLES, xpath, urlnormalize
+from ebook_converter.ebooks.oeb import base
 from ebook_converter.ebooks.oeb.normalize_css import DEFAULTS, normalizers
 from ebook_converter.css_selectors import Select, SelectorError, INAPPROPRIATE_PSEUDO_CLASSES
 from ebook_converter.tinycss.media3 import CSSMedia3Parser
 
 
-__license__ = 'GPL v3'
-__copyright__ = '2008, Marshall T. Vandegrift '
-
 css_parser_log.setLevel(logging.WARN)
 
 _html_css_stylesheet = None
@@ -208,7 +207,7 @@ class Stylizer(object):
         stylesheets = [html_css_stylesheet()]
         if base_css:
             stylesheets.append(parseString(base_css, validate=False))
-        style_tags = xpath(tree, '//*[local-name()="style" or local-name()="link"]')
+        style_tags = base.xpath(tree, '//*[local-name()="style" or local-name()="link"]')
 
         # Add css_parser parsing profiles from output_profile
         for profile in self.opts.output_profile.extra_css_modules:
@@ -219,7 +218,7 @@ class Stylizer(object):
         parser = CSSParser(fetcher=self._fetch_css_file,
                 log=logging.getLogger('calibre.css'))
         for elem in style_tags:
-            if (elem.tag == XHTML('style') and elem.get('type', CSS_MIME) in OEB_STYLES and media_ok(elem.get('media'))):
+            if (elem.tag == base.tag('xhtml', 'style') and elem.get('type', base.CSS_MIME) in base.OEB_STYLES and media_ok(elem.get('media'))):
                 text = elem.text if elem.text else ''
                 for x in elem:
                     t = getattr(x, 'text', None)
@@ -245,7 +244,7 @@ class Stylizer(object):
                                 self.logger.warn('Ignoring missing stylesheet in @import rule:', rule.href)
                                 continue
                             sitem = hrefs[ihref]
-                            if sitem.media_type not in OEB_STYLES:
+                            if sitem.media_type not in base.OEB_STYLES:
                                 self.logger.warn('CSS @import of non-CSS file %r' % rule.href)
                                 continue
                             stylesheets.append(sitem.data)
@@ -254,11 +253,11 @@ class Stylizer(object):
                     replaceUrls(stylesheet, item.abshref,
                             ignoreImportRules=True)
                     stylesheets.append(stylesheet)
-            elif (elem.tag == XHTML('link') and elem.get('href') and elem.get(
+            elif (elem.tag == base.tag('xhtml', 'link') and elem.get('href') and elem.get(
                     'rel', 'stylesheet').lower() == 'stylesheet' and elem.get(
-                    'type', CSS_MIME).lower() in OEB_STYLES and media_ok(elem.get('media'))
+                    'type', base.CSS_MIME).lower() in base.OEB_STYLES and media_ok(elem.get('media'))
                 ):
-                href = urlnormalize(elem.attrib['href'])
+                href = base.urlnormalize(elem.attrib['href'])
                 path = item.abshref(href)
                 sitem = oeb.manifest.hrefs.get(path, None)
                 if sitem is None:
@@ -326,7 +325,8 @@ class Stylizer(object):
 
                                 special_text = ''.join(punctuation_chars) + \
                                         (text[0] if text else '')
-                                span = x.makeelement('{%s}span' % XHTML_NS)
+                                span = x.makeelement('{%s}span' %
+                                                     const.XHTML_NS)
                                 span.text = special_text
                                 span.set('data-fake-first-letter', '1')
                                 span.tail = text[1:]
@@ -340,10 +340,10 @@ class Stylizer(object):
             else:
                 for elem in matches:
                     self.style(elem)._update_cssdict(cssdict)
-        for elem in xpath(tree, '//h:*[@style]'):
+        for elem in base.xpath(tree, '//h:*[@style]'):
             self.style(elem)._apply_style_attr(url_replacer=item.abshref)
         num_pat = re.compile(r'[0-9.]+$')
-        for elem in xpath(tree, '//h:img[@width or @height]'):
+        for elem in base.xpath(tree, '//h:img[@width or @height]'):
             style = self.style(elem)
             # Check if either height or width is not default
             is_styled = style._style.get('width', 'auto') != 'auto' or \
@@ -370,7 +370,7 @@ class Stylizer(object):
             self.logger.warn('CSS import of missing file %r' % path)
             return (None, None)
         item = hrefs[path]
-        if item.media_type not in OEB_STYLES:
+        if item.media_type not in base.OEB_STYLES:
             self.logger.warn('CSS import of non-CSS file %r' % path)
             return (None, None)
         data = item.data.cssText
diff --git a/ebook_converter/ebooks/oeb/transforms/cover.py b/ebook_converter/ebooks/oeb/transforms/cover.py
index bbc23be..f03e52c 100644
--- a/ebook_converter/ebooks/oeb/transforms/cover.py
+++ b/ebook_converter/ebooks/oeb/transforms/cover.py
@@ -1,66 +1,61 @@
 import textwrap
 import urllib.parse
 
+from lxml import etree
+
 from ebook_converter import guess_type
 from ebook_converter.utils.imghdr import identify
-from ebook_converter.utils.xml_parse import safe_xml_fromstring
 from ebook_converter.polyglot.urllib import unquote
 
 
-__license__ = 'GPL v3'
-__copyright__ = '2010, Kovid Goyal '
-__docformat__ = 'restructuredtext en'
-
-
 class CoverManager(object):
 
     SVG_TEMPLATE = textwrap.dedent('''\
-        
-            
-                
-                
-                Cover
-                
-            
-            
-                

-                    
-                        
-                    
-                

-            
-        
-        ''')
+
+    
+        
+        
+        Cover
+        
+    
+    
+        

+            
+                
+            
+        

+    
+''')
 
     NONSVG_TEMPLATE = textwrap.dedent('''\
-        
-            
-                
-                
-                Cover
-                
-            
-            
-                

-                    cover
-                

-            
-        
+
+    
+        
+        
+        Cover
+        
+    
+    
+        

+            cover
+        

+    
+
     ''')
 
     def __init__(self, no_default_cover=False, no_svg_cover=False,
-            preserve_aspect_ratio=False, fixed_size=None):
+                 preserve_aspect_ratio=False, fixed_size=None):
         self.no_default_cover = no_default_cover
         self.no_svg_cover = no_svg_cover
         self.preserve_aspect_ratio = preserve_aspect_ratio
@@ -72,9 +67,9 @@ class CoverManager(object):
             style = 'style="height: 100%%"'
         else:
             width, height = fixed_size
-            style = 'style="height: %s; width: %s"'%(height, width)
+            style = 'style="height: %s; width: %s"' % (height, width)
         self.non_svg_template = self.NONSVG_TEMPLATE.replace('__style__',
-                style)
+                                                             style)
 
     def __call__(self, oeb, opts, log):
         self.oeb = oeb
@@ -108,22 +103,23 @@ class CoverManager(object):
             # if self.preserve_aspect_ratio:
             #    width, height = 600, 800
             self.svg_template = self.svg_template.replace('__viewbox__',
-                    '0 0 %d %d'%(width, height))
+                                                          '0 0 %d %d' %
+                                                          (width, height))
             self.svg_template = self.svg_template.replace('__width__',
-                    str(width))
+                                                          str(width))
             self.svg_template = self.svg_template.replace('__height__',
-                    str(height))
+                                                          str(height))
 
             if href is not None:
                 templ = self.non_svg_template if self.no_svg_cover \
                         else self.svg_template
-                tp = templ%unquote(href)
+                tp = templ % unquote(href)
                 id, href = m.generate('titlepage', 'titlepage.xhtml')
                 item = m.add(id, href, guess_type('t.xhtml')[0],
-                        data=safe_xml_fromstring(tp))
+                             data=etree.fromstring(tp))
         else:
-            item = self.oeb.manifest.hrefs[
-                    urllib.parse.urldefrag(self.oeb.guide['titlepage'].href)[0]]
+            key = urllib.parse.urldefrag(self.oeb.guide['titlepage'].href)[0]
+            item = self.oeb.manifest.hrefs[key]
         if item is not None:
             self.oeb.spine.insert(0, item, True)
             if 'cover' not in self.oeb.guide.refs:
diff --git a/ebook_converter/ebooks/oeb/transforms/flatcss.py b/ebook_converter/ebooks/oeb/transforms/flatcss.py
index 3696e0a..6d29be3 100644
--- a/ebook_converter/ebooks/oeb/transforms/flatcss.py
+++ b/ebook_converter/ebooks/oeb/transforms/flatcss.py
@@ -1,26 +1,27 @@
 """
 CSS flattening transform.
 """
-import re, operator, math, numbers
-from collections import defaultdict
-from xml.dom import SyntaxErr
+import collections
+import math
+import numbers
+import operator
+import re
+from xml import dom
 
 from lxml import etree
 import css_parser
-from css_parser.css import Property
+from css_parser import css as cp_css
 
+from ebook_converter import constants as const
 from ebook_converter import guess_type
 from ebook_converter.ebooks import unit_convert
-from ebook_converter.ebooks.oeb.base import (XHTML, XHTML_NS, CSS_MIME, OEB_STYLES,
-        namespace, barename, XPath, css_text)
+from ebook_converter.ebooks.oeb import base
+from ebook_converter.ebooks.oeb import parse_utils
+
 from ebook_converter.ebooks.oeb.stylizer import Stylizer
 from ebook_converter.utils.filenames import ascii_filename, ascii_text
-from ebook_converter.utils.icu import numeric_sort_key
 
 
-__license__ = 'GPL v3'
-__copyright__ = '2008, Marshall T. Vandegrift '
-
 COLLAPSE = re.compile(r'[ \t\r\n\v]+')
 STRIPNUM = re.compile(r'[-0-9]+$')
 
@@ -121,7 +122,7 @@ class EmbedFontsCSSRules(object):
             return None
         if not self.href:
             iid, href = oeb.manifest.generate('page_styles', 'page_styles.css')
-            rules = [css_text(x) for x in self.rules]
+            rules = [base.css_text(x) for x in self.rules]
             rules = '\n\n'.join(rules)
             sheet = css_parser.parseString(rules, validate=False)
             self.href = oeb.manifest.add(iid, href, guess_type(href)[0],
@@ -186,7 +187,7 @@ class CSSFlattener(object):
         for item in oeb.manifest.values():
             # Make all links to resources absolute, as these sheets will be
             # consolidated into a single stylesheet at the root of the document
-            if item.media_type in OEB_STYLES:
+            if item.media_type in base.OEB_STYLES:
                 css_parser.replaceUrls(item.data, item.abshref,
                         ignoreImportRules=True)
 
@@ -273,7 +274,7 @@ class CSSFlattener(object):
         css = ''
         for item in self.items:
             html = item.data
-            body = html.find(XHTML('body'))
+            body = html.find(base.tag('xhtml', 'body'))
             if 'style' in html.attrib:
                 b = body.attrib.get('style', '')
                 body.set('style',  html.get('style') + ';' + b)
@@ -310,11 +311,11 @@ class CSSFlattener(object):
                 sizes[csize] += len(COLLAPSE.sub(' ', child.tail))
 
     def baseline_spine(self):
-        sizes = defaultdict(float)
+        sizes = collections.defaultdict(float)
         for item in self.items:
             html = item.data
             stylizer = self.stylizers[item]
-            body = html.find(XHTML('body'))
+            body = html.find(base.tag('xhtml', 'body'))
             fsize = self.context.source.fbase
             self.baseline_node(body, stylizer, sizes, fsize)
         try:
@@ -351,9 +352,9 @@ class CSSFlattener(object):
 
     def flatten_node(self, node, stylizer, names, styles, pseudo_styles, psize, item_id, recurse=True):
         if not isinstance(node.tag, (str, bytes)) \
-           or namespace(node.tag) != XHTML_NS:
+           or parse_utils.namespace(node.tag) != const.XHTML_NS:
             return
-        tag = barename(node.tag)
+        tag = parse_utils.barename(node.tag)
         style = stylizer.style(node)
         cssdict = style.cssdict()
         try:
@@ -375,7 +376,7 @@ class CSSFlattener(object):
                         if 'margin-left' not in cssdict and 'margin-right' not in cssdict:
                             cssdict['margin-left'] = cssdict['margin-right'] = 'auto'
                     else:
-                        for table in node.iterchildren(XHTML("table")):
+                        for table in node.iterchildren(base.tag('xhtml', "table")):
                             ts = stylizer.style(table)
                             if ts.get('margin-left') is None and ts.get('margin-right') is None:
                                 ts.set('margin-left', 'auto')
@@ -391,11 +392,12 @@ class CSSFlattener(object):
             if cssdict.get('vertical-align') == 'inherit':
                 cssdict['vertical-align'] = node.attrib['valign']
             del node.attrib['valign']
-        if node.tag == XHTML('font'):
+        if node.tag == base.tag('xhtml', 'font'):
             tags = ['descendant::h:%s'%x for x in ('p', 'div', 'table', 'h1',
                 'h2', 'h3', 'h4', 'h5', 'h6', 'ol', 'ul', 'dl', 'blockquote')]
-            tag = 'div' if XPath('|'.join(tags))(node) else 'span'
-            node.tag = XHTML(tag)
+            # TODO(gryf): this will override tag from line 355. On purpose?
+            tag = 'div' if base.XPath('|'.join(tags))(node) else 'span'
+            node.tag = base.tag('xhtml', tag)
             if 'size' in node.attrib:
                 def force_int(raw):
                     return int(re.search(r'([0-9+-]+)', raw).group(1))
@@ -425,14 +427,14 @@ class CSSFlattener(object):
                 del node.attrib['face']
         if 'color' in node.attrib:
             try:
-                cssdict['color'] = Property('color', node.attrib['color']).value
-            except (ValueError, SyntaxErr):
+                cssdict['color'] = cp_css.Property('color', node.attrib['color']).value
+            except (ValueError, dom.SyntaxErr):
                 pass
             del node.attrib['color']
         if 'bgcolor' in node.attrib:
             try:
-                cssdict['background-color'] = Property('background-color', node.attrib['bgcolor']).value
-            except (ValueError, SyntaxErr):
+                cssdict['background-color'] = cp_css.Property('background-color', node.attrib['bgcolor']).value
+            except (ValueError, dom.SyntaxErr):
                 pass
             del node.attrib['bgcolor']
         if tag == 'ol' and 'type' in node.attrib:
@@ -573,7 +575,7 @@ class CSSFlattener(object):
 
     def flatten_head(self, item, href, global_href):
         html = item.data
-        head = html.find(XHTML('head'))
+        head = html.find(base.tag('xhtml', 'head'))
 
         def safe_lower(x):
             try:
@@ -583,39 +585,39 @@ class CSSFlattener(object):
             return x
 
         for node in html.xpath('//*[local-name()="style" or local-name()="link"]'):
-            if node.tag == XHTML('link') \
+            if node.tag == base.tag('xhtml', 'link') \
                and safe_lower(node.get('rel', 'stylesheet')) == 'stylesheet' \
-               and safe_lower(node.get('type', CSS_MIME)) in OEB_STYLES:
+               and safe_lower(node.get('type', base.CSS_MIME)) in base.OEB_STYLES:
                 node.getparent().remove(node)
-            elif node.tag == XHTML('style') \
-                 and node.get('type', CSS_MIME) in OEB_STYLES:
+            elif node.tag == base.tag('xhtml', 'style') \
+                 and node.get('type', base.CSS_MIME) in base.OEB_STYLES:
                 node.getparent().remove(node)
         href = item.relhref(href)
-        l = etree.SubElement(head, XHTML('link'),
-            rel='stylesheet', type=CSS_MIME, href=href)
+        l = etree.SubElement(head, base.tag('xhtml', 'link'),
+            rel='stylesheet', type=base.CSS_MIME, href=href)
         l.tail='\n'
         if global_href:
             href = item.relhref(global_href)
-            l = etree.SubElement(head, XHTML('link'),
-                rel='stylesheet', type=CSS_MIME, href=href)
+            l = etree.SubElement(head, base.tag('xhtml', 'link'),
+                rel='stylesheet', type=base.CSS_MIME, href=href)
             l.tail = '\n'
 
     def replace_css(self, css):
         manifest = self.oeb.manifest
         for item in manifest.values():
-            if item.media_type in OEB_STYLES:
+            if item.media_type in base.OEB_STYLES:
                 manifest.remove(item)
         id, href = manifest.generate('css', 'stylesheet.css')
         sheet = css_parser.parseString(css, validate=False)
         if self.transform_css_rules:
             from ebook_converter.ebooks.css_transform_rules import transform_sheet
             transform_sheet(self.transform_css_rules, sheet)
-        item = manifest.add(id, href, CSS_MIME, data=sheet)
+        item = manifest.add(id, href, base.CSS_MIME, data=sheet)
         self.oeb.manifest.main_stylesheet = item
         return href
 
     def collect_global_css(self):
-        global_css = defaultdict(list)
+        global_css = collections.defaultdict(list)
         for item in self.items:
             stylizer = self.stylizers[item]
             if float(self.context.margin_top) >= 0:
@@ -627,7 +629,7 @@ class CSSFlattener(object):
             items = sorted(stylizer.page_rule.items())
             css = ';\n'.join("%s: %s" % (key, val) for key, val in items)
             css = ('@page {\n%s\n}\n'%css) if items else ''
-            rules = [css_text(r) for r in stylizer.font_face_rules + self.embed_font_rules]
+            rules = [base.css_text(r) for r in stylizer.font_face_rules + self.embed_font_rules]
             raw = '\n\n'.join(rules)
             css += '\n\n' + raw
             global_css[css].append(item)
@@ -642,7 +644,7 @@ class CSSFlattener(object):
                 if self.transform_css_rules:
                     from ebook_converter.ebooks.css_transform_rules import transform_sheet
                     transform_sheet(self.transform_css_rules, sheet)
-                manifest.add(id_, href, CSS_MIME, data=sheet)
+                manifest.add(id_, href, base.CSS_MIME, data=sheet)
             gc_map[css] = href
 
         ans = {}
@@ -652,8 +654,8 @@ class CSSFlattener(object):
         return ans
 
     def flatten_spine(self):
-        names = defaultdict(int)
-        styles, pseudo_styles = {}, defaultdict(dict)
+        names = collections.defaultdict(int)
+        styles, pseudo_styles = {}, collections.defaultdict(dict)
         for item in self.items:
             html = item.data
             stylizer = self.stylizers[item]
@@ -661,7 +663,7 @@ class CSSFlattener(object):
                 self.specializer(item, stylizer)
             fsize = self.context.dest.fbase
             self.flatten_node(html, stylizer, names, styles, pseudo_styles, fsize, item.id, recurse=False)
-            self.flatten_node(html.find(XHTML('body')), stylizer, names, styles, pseudo_styles, fsize, item.id)
+            self.flatten_node(html.find(base.tag('xhtml', 'body')), stylizer, names, styles, pseudo_styles, fsize, item.id)
         items = sorted(((key, val) for (val, key) in styles.items()))
         # :hover must come after link and :active must come after :hover
         psels = sorted(pseudo_styles, key=lambda x :
diff --git a/ebook_converter/ebooks/oeb/transforms/htmltoc.py b/ebook_converter/ebooks/oeb/transforms/htmltoc.py
index 8c555f6..1f65c56 100644
--- a/ebook_converter/ebooks/oeb/transforms/htmltoc.py
+++ b/ebook_converter/ebooks/oeb/transforms/htmltoc.py
@@ -1,46 +1,20 @@
 """
 HTML-TOC-adding transform.
 """
-from ebook_converter.ebooks.oeb.base import XML, XHTML, XHTML_NS
-from ebook_converter.ebooks.oeb.base import XHTML_MIME, CSS_MIME
-from ebook_converter.ebooks.oeb.base import element, XPath
+from ebook_converter import constants as const
+from ebook_converter.ebooks.oeb import base
 
 
-__all__ = ['HTMLTOCAdder']
-__license__ = 'GPL v3'
-__copyright__ = '2008, Marshall T. Vandegrift '
-
 DEFAULT_TITLE = 'Table of Contents'
+STYLE_CSS = {'nested': '.calibre_toc_header {\n  text-align: center;\n}\n'
+             '.calibre_toc_block {\n  margin-left: 1.2em;\n  text-indent: '
+             '-1.2em;\n}\n.calibre_toc_block .calibre_toc_block {\n  '
+             'margin-left: 2.4em;\n}\n.calibre_toc_block .calibre_toc_block '
+             '.calibre_toc_block {\n  margin-left: 3.6em;\n}\n',
 
-STYLE_CSS = {
-    'nested': """
-.calibre_toc_header {
-  text-align: center;
-}
-.calibre_toc_block {
-  margin-left: 1.2em;
-  text-indent: -1.2em;
-}
-.calibre_toc_block .calibre_toc_block {
-  margin-left: 2.4em;
-}
-.calibre_toc_block .calibre_toc_block .calibre_toc_block {
-  margin-left: 3.6em;
-}
-""",
-
-    'centered': """
-.calibre_toc_header {
-  text-align: center;
-}
-.calibre_toc_block {
-  text-align: center;
-}
-body > .calibre_toc_block {
-  margin-top: 1.2em;
-}
-"""
-    }
+             'centered': '.calibre_toc_header {\n  text-align: center;\n}\n'
+             '.calibre_toc_block {\n  text-align: center;\n}\nbody > '
+             '.calibre_toc_block {\n  margin-top: 1.2em;\n}\n'}
 
 
 class HTMLTOCAdder(object):
@@ -71,7 +45,7 @@ class HTMLTOCAdder(object):
             if href in oeb.manifest.hrefs:
                 item = oeb.manifest.hrefs[href]
                 if (hasattr(item.data, 'xpath') and
-                    XPath('//h:a[@href]')(item.data)):
+                        base.XPath('//h:a[@href]')(item.data)):
                     if oeb.spine.index(item) < 0:
                         if self.position == 'end':
                             oeb.spine.add(item, linear=False)
@@ -91,23 +65,24 @@ class HTMLTOCAdder(object):
             oeb.logger.error('Unknown TOC style %r' % style)
             style = 'nested'
         id, css_href = oeb.manifest.generate('tocstyle', 'tocstyle.css')
-        oeb.manifest.add(id, css_href, CSS_MIME, data=STYLE_CSS[style])
+        oeb.manifest.add(id, css_href, base.CSS_MIME, data=STYLE_CSS[style])
         language = str(oeb.metadata.language[0])
-        contents = element(None, XHTML('html'), nsmap={None: XHTML_NS},
-                           attrib={XML('lang'): language})
-        head = element(contents, XHTML('head'))
-        htitle = element(head, XHTML('title'))
+        contents = base.element(None, base.tag('xhtml', 'html'),
+                                nsmap={None: const.XHTML_NS},
+                                attrib={base.tag('xml', 'lang'): language})
+        head = base.element(contents, base.tag('xhtml', 'head'))
+        htitle = base.element(head, base.tag('xhtml', 'title'))
         htitle.text = title
-        element(head, XHTML('link'), rel='stylesheet', type=CSS_MIME,
-                href=css_href)
-        body = element(contents, XHTML('body'),
-                       attrib={'class': 'calibre_toc'})
-        h1 = element(body, XHTML('h2'),
-                     attrib={'class': 'calibre_toc_header'})
+        base.element(head, base.tag('xhtml', 'link'), rel='stylesheet',
+                     type=base.CSS_MIME, href=css_href)
+        body = base.element(contents, base.tag('xhtml', 'body'),
+                            attrib={'class': 'calibre_toc'})
+        h1 = base.element(body, base.tag('xhtml', 'h2'),
+                          attrib={'class': 'calibre_toc_header'})
         h1.text = title
         self.add_toc_level(body, oeb.toc)
         id, href = oeb.manifest.generate('contents', 'contents.xhtml')
-        item = oeb.manifest.add(id, href, XHTML_MIME, data=contents)
+        item = oeb.manifest.add(id, href, base.XHTML_MIME, data=contents)
         if self.position == 'end':
             oeb.spine.add(item, linear=False)
         else:
@@ -116,10 +91,10 @@ class HTMLTOCAdder(object):
 
     def add_toc_level(self, elem, toc):
         for node in toc:
-            block = element(elem, XHTML('div'),
-                            attrib={'class': 'calibre_toc_block'})
-            line = element(block, XHTML('a'),
-                           attrib={'href': node.href,
-                                   'class': 'calibre_toc_line'})
+            block = base.element(elem, base.tag('xhtml', 'div'),
+                                 attrib={'class': 'calibre_toc_block'})
+            line = base.element(block, base.tag('xhtml', 'a'),
+                                attrib={'href': node.href,
+                                        'class': 'calibre_toc_line'})
             line.text = node.title
             self.add_toc_level(block, node)
diff --git a/ebook_converter/ebooks/oeb/transforms/jacket.py b/ebook_converter/ebooks/oeb/transforms/jacket.py
index 0d1a88b..61a0dbf 100644
--- a/ebook_converter/ebooks/oeb/transforms/jacket.py
+++ b/ebook_converter/ebooks/oeb/transforms/jacket.py
@@ -4,9 +4,10 @@ from string import Formatter
 import pkg_resources
 import urllib.parse
 
+from ebook_converter import constants as const
 from ebook_converter import guess_type, strftime
 from ebook_converter.constants_old import iswindows
-from ebook_converter.ebooks.oeb.base import XPath, XHTML_NS, XHTML, xml2text, urlnormalize
+from ebook_converter.ebooks.oeb.base import XPath, xml2text, urlnormalize
 from ebook_converter.library.comments import comments_to_html, markdown
 from ebook_converter.utils.date import is_date_undefined, as_local_time
 from ebook_converter.ebooks.chardet import strip_encoding_declarations
@@ -303,7 +304,7 @@ def render_jacket(mi, output_profile,
                 'tags_label': 'Tags',
                 'title': title,
                 'title_str': title_str,
-                'xmlns': XHTML_NS}
+                'xmlns': const.XHTML_NS}
 
         for key in mi.custom_field_keys():
             m = mi.get_user_metadata(key, False) or {}
@@ -370,7 +371,7 @@ def render_jacket(mi, output_profile,
         # We cannot use data-calibre-rescale 100 on the body tag as that will just
         # give the body tag a font size of 1em, which is useless.
         for body in root.xpath('//*[local-name()="body"]'):
-            fw = body.makeelement(XHTML('div'))
+            fw = body.makeelement(const.XHTML_DIV)
             fw.set('data-calibre-rescale', '100')
             for child in body:
                 fw.append(child)
@@ -387,9 +388,9 @@ def linearize_jacket(oeb):
     for x in oeb.spine[:4]:
         if XPath(JACKET_XPATH)(x.data):
             for e in XPath('//h:table|//h:tr|//h:th')(x.data):
-                e.tag = XHTML('div')
+                e.tag = const.XHTML_DIV
             for e in XPath('//h:td')(x.data):
-                e.tag = XHTML('span')
+                e.tag = const.XHTML_SPAN
             break
 
 
diff --git a/ebook_converter/ebooks/oeb/transforms/manglecase.py b/ebook_converter/ebooks/oeb/transforms/manglecase.py
index 67b6493..2edf957 100644
--- a/ebook_converter/ebooks/oeb/transforms/manglecase.py
+++ b/ebook_converter/ebooks/oeb/transforms/manglecase.py
@@ -5,9 +5,9 @@ import string
 
 from lxml import etree
 
-from ebook_converter.ebooks.oeb.base import XHTML, XHTML_NS
-from ebook_converter.ebooks.oeb.base import CSS_MIME
-from ebook_converter.ebooks.oeb.base import namespace
+from ebook_converter import constants as const
+from ebook_converter.ebooks.oeb import base
+from ebook_converter.ebooks.oeb import parse_utils
 from ebook_converter.ebooks.oeb.stylizer import Stylizer
 
 
@@ -43,15 +43,16 @@ class CaseMangler(object):
 
     def mangle_spine(self):
         id, href = self.oeb.manifest.generate('manglecase', 'manglecase.css')
-        self.oeb.manifest.add(id, href, CSS_MIME, data=CASE_MANGLER_CSS)
+        self.oeb.manifest.add(id, href, base.CSS_MIME, data=CASE_MANGLER_CSS)
         for item in self.oeb.spine:
             html = item.data
             relhref = item.relhref(href)
-            etree.SubElement(html.find(XHTML('head')), XHTML('link'),
-                             rel='stylesheet', href=relhref, type=CSS_MIME)
+            etree.SubElement(html.find(base.tag('xhtml', 'head')),
+                             base.tag('xhtml', 'link'), rel='stylesheet',
+                             href=relhref, type=base.CSS_MIME)
             stylizer = Stylizer(html, item.href, self.oeb, self.opts,
                                 self.profile)
-            self.mangle_elem(html.find(XHTML('body')), stylizer)
+            self.mangle_elem(html.find(base.tag('xhtml', 'body')), stylizer)
 
     def text_transform(self, transform, text):
         if transform == 'capitalize':
@@ -85,7 +86,8 @@ class CaseMangler(object):
                 else:
                     last.tail = text
             else:
-                child = elem.makeelement(XHTML('span'), attrib=attrib)
+                child = elem.makeelement(base.tag('xhtml', 'span'),
+                                         attrib=attrib)
                 child.text = text.upper()
                 if last is None:
                     elem.insert(0, child)
@@ -99,7 +101,7 @@ class CaseMangler(object):
 
     def mangle_elem(self, elem, stylizer):
         if not isinstance(elem.tag, (str, bytes)) or \
-           namespace(elem.tag) != XHTML_NS:
+           parse_utils.namespace(elem.tag) != const.XHTML_NS:
             return
         children = list(elem)
         style = stylizer.style(elem)
diff --git a/ebook_converter/ebooks/oeb/transforms/metadata.py b/ebook_converter/ebooks/oeb/transforms/metadata.py
index c08a840..be550b6 100644
--- a/ebook_converter/ebooks/oeb/transforms/metadata.py
+++ b/ebook_converter/ebooks/oeb/transforms/metadata.py
@@ -1,15 +1,12 @@
-import os, re
+import os
+import re
+
+from ebook_converter.ebooks.oeb import base
 from ebook_converter.utils.date import isoformat, now
 from ebook_converter import guess_type
 
 
-__license__ = 'GPL v3'
-__copyright__ = '2009, Kovid Goyal '
-__docformat__ = 'restructuredtext en'
-
-
 def meta_info_to_oeb_metadata(mi, m, log, override_input_metadata=False):
-    from ebook_converter.ebooks.oeb.base import OPF
     if not mi.is_null('title'):
         m.clear('title')
         m.add('title', mi.title)
@@ -19,17 +16,17 @@ def meta_info_to_oeb_metadata(mi, m, log, override_input_metadata=False):
         m.clear('title_sort')
         m.add('title_sort', mi.title_sort)
     if not mi.is_null('authors'):
-        m.filter('creator', lambda x : x.role.lower() in ['aut', ''])
+        m.filter('creator', lambda x: x.role.lower() in ['aut', ''])
         for a in mi.authors:
-            attrib = {'role':'aut'}
+            attrib = {'role': 'aut'}
             if mi.author_sort:
-                attrib[OPF('file-as')] = mi.author_sort
+                attrib[base.tag('opf', 'file-as')] = mi.author_sort
             m.add('creator', a, attrib=attrib)
     if not mi.is_null('book_producer'):
-        m.filter('contributor', lambda x : x.role.lower() == 'bkp')
+        m.filter('contributor', lambda x: x.role.lower() == 'bkp')
         m.add('contributor', mi.book_producer, role='bkp')
     elif override_input_metadata:
-        m.filter('contributor', lambda x : x.role.lower() == 'bkp')
+        m.filter('contributor', lambda x: x.role.lower() == 'bkp')
     if not mi.is_null('comments'):
         m.clear('description')
         m.add('description', mi.comments)
@@ -71,7 +68,7 @@ def meta_info_to_oeb_metadata(mi, m, log, override_input_metadata=False):
         m.clear('series_index')
     if not mi.is_null('rating'):
         m.clear('rating')
-        m.add('rating', '%.2f'%mi.rating)
+        m.add('rating', '%.2f' % mi.rating)
     elif override_input_metadata:
         m.clear('rating')
     if not mi.is_null('tags'):
@@ -101,23 +98,25 @@ class MergeMetadata(object):
     'Merge in user metadata, including cover'
 
     def __call__(self, oeb, mi, opts, override_input_metadata=False):
+        _oim = override_input_metadata
         self.oeb, self.log = oeb, oeb.log
         m = self.oeb.metadata
         self.log('Merging user specified metadata...')
         meta_info_to_oeb_metadata(mi, m, oeb.log,
-                override_input_metadata=override_input_metadata)
+                                  override_input_metadata=_oim)
         cover_id = self.set_cover(mi, opts.prefer_metadata_cover)
         m.clear('cover')
         if cover_id is not None:
             m.add('cover', cover_id)
         if mi.uuid is not None:
-            m.filter('identifier', lambda x:x.id=='uuid_id')
+            m.filter('identifier', lambda x: x.id == 'uuid_id')
             self.oeb.metadata.add('identifier', mi.uuid, id='uuid_id',
                                   scheme='uuid')
             self.oeb.uid = self.oeb.metadata.identifier[-1]
         if mi.application_id is not None:
-            m.filter('identifier', lambda x:x.scheme=='calibre')
-            self.oeb.metadata.add('identifier', mi.application_id, scheme='calibre')
+            m.filter('identifier', lambda x: x.scheme == 'calibre')
+            self.oeb.metadata.add('identifier', mi.application_id,
+                                  scheme='calibre')
 
     def set_cover(self, mi, prefer_metadata_cover):
         cdata, ext = b'', 'jpg'
@@ -138,7 +137,8 @@ class MergeMetadata(object):
         if cdata:
             self.oeb.guide.remove('cover')
             self.oeb.guide.remove('titlepage')
-        elif self.oeb.plumber_output_format in {'mobi', 'azw3'} and old_cover is not None:
+        elif (self.oeb.plumber_output_format in {'mobi', 'azw3'} and
+                old_cover is not None):
             # The amazon formats dont support html cover pages, so remove them
             # even if no cover was specified.
             self.oeb.guide.remove('titlepage')
@@ -156,7 +156,9 @@ class MergeMetadata(object):
         new_cover_item = None
         if cdata:
             id, href = self.oeb.manifest.generate('cover', 'cover.'+ext)
-            new_cover_item = self.oeb.manifest.add(id, href, guess_type('cover.'+ext)[0], data=cdata)
+            new_cover_item = self.oeb.manifest.add(id, href,
+                                                   guess_type('cover.'+ext)[0],
+                                                   data=cdata)
             self.oeb.guide.add('cover', 'Cover', href)
         if do_remove_old_cover:
             self.remove_old_cover(item, new_cover_item.href)
@@ -186,7 +188,8 @@ class MergeMetadata(object):
                 if href == cover_item.href:
                     if new_cover_href is not None:
                         replacement_href = item.relhref(new_cover_href)
-                        attr = 'src' if img.tag.endswith('img') else XLINK('href')
+                        attr = ('src' if img.tag.endswith('img')
+                                else XLINK('href'))
                         img.set(attr, replacement_href)
                     else:
                         p = img.getparent()
@@ -202,13 +205,14 @@ class MergeMetadata(object):
         for item in affected_items:
             body = XPath('//h:body')(item.data)
             if body:
-                text = etree.tostring(body[0], method='text', encoding='unicode')
+                text = etree.tostring(body[0], method='text',
+                                      encoding='unicode')
             else:
                 text = ''
             text = re.sub(r'\s+', '', text)
             if not text and not XPath('//h:img|//svg:svg')(item.data):
-                self.log('Removing %s as it is a wrapper around'
-                        ' the cover image'%item.href)
+                self.log('Removing %s as it is a wrapper around the cover '
+                         'image' % item.href)
                 self.oeb.spine.remove(item)
                 self.oeb.manifest.remove(item)
                 self.oeb.guide.remove_by_href(item.href)
diff --git a/ebook_converter/ebooks/oeb/transforms/page_margin.py b/ebook_converter/ebooks/oeb/transforms/page_margin.py
index 8c4e232..bfd81c3 100644
--- a/ebook_converter/ebooks/oeb/transforms/page_margin.py
+++ b/ebook_converter/ebooks/oeb/transforms/page_margin.py
@@ -1,7 +1,8 @@
 import numbers
 from collections import Counter
 
-from ebook_converter.ebooks.oeb.base import barename, XPath
+from ebook_converter.ebooks.oeb import parse_utils
+from ebook_converter.ebooks.oeb.base import XPath
 
 
 __license__ = 'GPL v3'
@@ -142,7 +143,7 @@ class RemoveFakeMargins(object):
 
             for p in paras(body):
                 level = level_of(p, body)
-                level = '%s_%d'%(barename(p.tag), level)
+                level = '%s_%d' % (parse_utils.barename(p.tag), level)
                 if level not in self.levels:
                     self.levels[level] = []
                 self.levels[level].append(p)
diff --git a/ebook_converter/ebooks/oeb/transforms/rasterize.py b/ebook_converter/ebooks/oeb/transforms/rasterize.py
index 074a255..df90c2d 100644
--- a/ebook_converter/ebooks/oeb/transforms/rasterize.py
+++ b/ebook_converter/ebooks/oeb/transforms/rasterize.py
@@ -5,10 +5,8 @@ import os
 import re
 import urllib.parse
 
-# from PyQt5.Qt import (
-    # Qt, QByteArray, QBuffer, QIODevice, QColor, QImage, QPainter, QSvgRenderer)
-
-from ebook_converter.ebooks.oeb.base import XHTML, XLINK
+from ebook_converter import constants as const
+from ebook_converter.ebooks.oeb import base
 from ebook_converter.ebooks.oeb.base import SVG_MIME, PNG_MIME
 from ebook_converter.ebooks.oeb.base import xml2str, xpath
 from ebook_converter.ebooks.oeb.base import urlnormalize
@@ -17,10 +15,7 @@ from ebook_converter.ptempfile import PersistentTemporaryFile
 from ebook_converter.utils.imghdr import what
 
 
-__license__ = 'GPL v3'
-__copyright__ = '2008, Marshall T. Vandegrift '
-
-IMAGE_TAGS = {XHTML('img'), XHTML('object')}
+IMAGE_TAGS = {base.tag('xhtml', 'img'), base.tag('xhtml', 'object')}
 KEEP_ATTRS = {'class', 'style', 'width', 'height', 'align'}
 
 
@@ -113,7 +108,7 @@ class SVGRasterizer(object):
             svg = item.data
         hrefs = self.oeb.manifest.hrefs
         for elem in xpath(svg, '//svg:*[@xl:href]'):
-            href = urlnormalize(elem.attrib[XLINK('href')])
+            href = urlnormalize(elem.attrib[base.tag('xlink', 'href')])
             path = urllib.parse.urldefrag(href)[0]
             if not path:
                 continue
@@ -126,7 +121,7 @@ class SVGRasterizer(object):
             with PersistentTemporaryFile(suffix='.'+ext) as pt:
                 pt.write(data)
                 self.temp_files.append(pt.name)
-            elem.attrib[XLINK('href')] = pt.name
+            elem.attrib[base.tag('xlink', 'href')] = pt.name
         return svg
 
     def stylizer(self, item):
@@ -171,7 +166,7 @@ class SVGRasterizer(object):
         href = os.path.splitext(item.href)[0] + '.png'
         id, href = manifest.generate(item.id, href)
         manifest.add(id, href, PNG_MIME, data=data)
-        img = elem.makeelement(XHTML('img'), src=item.relhref(href))
+        img = elem.makeelement(base.tag('xhtml', 'img'), src=item.relhref(href))
         elem.getparent().replace(elem, img)
         for prop in ('width', 'height'):
             if prop in elem.attrib:
@@ -208,7 +203,7 @@ class SVGRasterizer(object):
             id, href = manifest.generate(svgitem.id, href)
             manifest.add(id, href, PNG_MIME, data=data)
             self.images[key] = href
-        elem.tag = XHTML('img')
+        elem.tag = base.tag('xhtml', 'img')
         for attr in elem.attrib:
             if attr not in KEEP_ATTRS:
                 del elem.attrib[attr]
diff --git a/ebook_converter/ebooks/oeb/transforms/split.py b/ebook_converter/ebooks/oeb/transforms/split.py
index 6cce60b..b20f32b 100644
--- a/ebook_converter/ebooks/oeb/transforms/split.py
+++ b/ebook_converter/ebooks/oeb/transforms/split.py
@@ -10,10 +10,11 @@ import urllib.parse
 from lxml.etree import XPath as _XPath
 from lxml import etree
 
+from ebook_converter import constants as const
 from ebook_converter import as_unicode, force_unicode
 from ebook_converter.ebooks.epub import rules
-from ebook_converter.ebooks.oeb.base import (OEB_STYLES, XPNSMAP as NAMESPACES,
-        rewrite_links, XHTML, urlnormalize)
+from ebook_converter.ebooks.oeb.base import \
+        OEB_STYLES, rewrite_links, urlnormalize
 from ebook_converter.ebooks.oeb.polish.split import do_split
 from ebook_converter.polyglot.urllib import unquote
 from ebook_converter.css_selectors import Select, SelectorError
@@ -22,7 +23,7 @@ from ebook_converter.css_selectors import Select, SelectorError
 __license__ = 'GPL v3'
 __copyright__ = '2008, Kovid Goyal kovid@kovidgoyal.net'
 __docformat__ = 'restructuredtext en'
-XPath = functools.partial(_XPath, namespaces=NAMESPACES)
+XPath = functools.partial(_XPath, namespaces=const.XPNSMAP)
 
 SPLIT_POINT_ATTR = 'csp'
 
@@ -104,7 +105,7 @@ class Split(object):
         select = Select(item.data)
         if not self.page_break_selectors:
             return [], []
-        body = item.data.xpath('//h:body', namespaces=NAMESPACES)
+        body = item.data.xpath('//h:body', namespaces=const.XPNSMAP)
         if not body:
             return [], []
         descendants = frozenset(body[0].iterdescendants('*'))
@@ -268,13 +269,13 @@ class FlowSplitter(object):
                     if body is not None:
                         existing_ids = frozenset(body.xpath('//*/@id'))
                         for x in ids - existing_ids:
-                            body.insert(0, body.makeelement(XHTML('div'), id=x, style='height:0pt'))
+                            body.insert(0, body.makeelement(const.XHTML_div, id=x, style='height:0pt'))
                 ids = set()
                 trees.append(tree)
         self.trees = trees
 
     def get_body(self, root):
-        body = root.xpath('//h:body', namespaces=NAMESPACES)
+        body = root.xpath('//h:body', namespaces=const.XPNSMAP)
         if not body:
             return None
         return body[0]
@@ -296,7 +297,7 @@ class FlowSplitter(object):
                 etree.tostring(body, method='text', encoding='unicode'))
         if len(txt) > 1:
             return False
-        for img in root.xpath('//h:img', namespaces=NAMESPACES):
+        for img in root.xpath('//h:img', namespaces=const.XPNSMAP):
             if img.get('style', '') != 'display:none':
                 return False
         if root.xpath('//*[local-name() = "svg"]'):
@@ -401,7 +402,7 @@ class FlowSplitter(object):
                      '//h:br',
                      '//h:li',
                      ):
-            elems = root.xpath(path, namespaces=NAMESPACES)
+            elems = root.xpath(path, namespaces=const.XPNSMAP)
             elem = pick_elem(elems)
             if elem is not None:
                 try:
@@ -436,7 +437,7 @@ class FlowSplitter(object):
         spine_pos = self.item.spine_position
 
         for current, tree in zip(*map(reversed, (self.files, self.trees))):
-            for a in tree.getroot().xpath('//h:a[@href]', namespaces=NAMESPACES):
+            for a in tree.getroot().xpath('//h:a[@href]', namespaces=const.XPNSMAP):
                 href = a.get('href').strip()
                 if href.startswith('#'):
                     anchor = href[1:]
diff --git a/ebook_converter/ebooks/oeb/transforms/structure.py b/ebook_converter/ebooks/oeb/transforms/structure.py
index d3049f2..9b042fd 100644
--- a/ebook_converter/ebooks/oeb/transforms/structure.py
+++ b/ebook_converter/ebooks/oeb/transforms/structure.py
@@ -1,22 +1,19 @@
+import collections
 import re
-import uuid
 import urllib.parse
+import uuid
 
 from lxml import etree
-from collections import OrderedDict, Counter
 
-from ebook_converter.ebooks.oeb.base import XPNSMAP, TOC, XHTML, xml2text, barename
+from ebook_converter import constants as const
+from ebook_converter.ebooks.oeb import parse_utils
+from ebook_converter.ebooks.oeb.base import TOC, xml2text
 from ebook_converter.ebooks import ConversionError
 
 
-__license__ = 'GPL v3'
-__copyright__ = '2009, Kovid Goyal '
-__docformat__ = 'restructuredtext en'
-
-
 def XPath(x):
     try:
-        return etree.XPath(x, namespaces=XPNSMAP)
+        return etree.XPath(x, namespaces=const.XPNSMAP)
     except etree.XPathSyntaxError:
         raise ConversionError(
         'The syntax of the XPath expression %s is invalid.' % repr(x))
@@ -84,7 +81,7 @@ class DetectStructure(object):
                     try:
                         prev = next(elem.itersiblings(tag=etree.Element,
                                 preceding=True))
-                        if (barename(elem.tag) in {'h1', 'h2'} and barename(
+                        if (parse_utils.barename(elem.tag) in {'h1', 'h2'} and parse_utils.barename(
                                 prev.tag) in {'h1', 'h2'} and (not prev.tail or
                                     not prev.tail.split())):
                             # We have two adjacent headings, do not put a page
@@ -165,7 +162,7 @@ class DetectStructure(object):
             chapter_mark = self.opts.chapter_mark
             page_break_before = 'display: block; page-break-before: always'
             page_break_after = 'display: block; page-break-after: always'
-            c = Counter()
+            c = collections.Counter()
             for item, elem in self.detected_chapters:
                 c[item] += 1
                 text = xml2text(elem).strip()
@@ -174,7 +171,7 @@ class DetectStructure(object):
                 if chapter_mark == 'none':
                     continue
                 if chapter_mark == 'rule':
-                    mark = elem.makeelement(XHTML('hr'))
+                    mark = elem.makeelement(const.XHTML_HR)
                 elif chapter_mark == 'pagebreak':
                     if c[item] < 3 and at_start(elem):
                         # For the first two elements in this item, check if they
@@ -184,9 +181,9 @@ class DetectStructure(object):
                         # feedbooks epubs match both a heading tag and its
                         # containing div with the default chapter expression.
                         continue
-                    mark = elem.makeelement(XHTML('div'), style=page_break_after)
+                    mark = elem.makeelement(const.XHTML_DIV, style=page_break_after)
                 else:  # chapter_mark == 'both':
-                    mark = elem.makeelement(XHTML('hr'), style=page_break_before)
+                    mark = elem.makeelement(const.XHTML_HR, style=page_break_before)
                 try:
                     elem.addprevious(mark)
                 except TypeError:
@@ -254,8 +251,8 @@ class DetectStructure(object):
         return text, href
 
     def add_leveled_toc_items(self):
-        added = OrderedDict()
-        added2 = OrderedDict()
+        added = collections.OrderedDict()
+        added2 = collections.OrderedDict()
         counter = 1
 
         def find_matches(expr, doc):
diff --git a/ebook_converter/ebooks/pdf/pdftohtml.py b/ebook_converter/ebooks/pdf/pdftohtml.py
index d10cd82..554dc37 100644
--- a/ebook_converter/ebooks/pdf/pdftohtml.py
+++ b/ebook_converter/ebooks/pdf/pdftohtml.py
@@ -5,10 +5,10 @@ import shutil
 import subprocess
 import sys
 
+from lxml import etree
+
 from ebook_converter import CurrentDir, xml_replace_entities, prints
-from ebook_converter.constants_old import (
-    filesystem_encoding, isbsd, islinux, isosx, iswindows
-)
+from ebook_converter.constants_old import isbsd, islinux, isosx, iswindows
 from ebook_converter.ebooks import ConversionError, DRMError
 from ebook_converter.ebooks.chardet import xml_to_unicode
 from ebook_converter.ptempfile import PersistentTemporaryFile
@@ -26,10 +26,13 @@ def popen(cmd, **kw):
 
 
 if isosx and hasattr(sys, 'frameworks_dir'):
-    base = os.path.join(os.path.dirname(sys.frameworks_dir), 'utils.app', 'Contents', 'MacOS')
+    base = os.path.join(os.path.dirname(sys.frameworks_dir), 'utils.app',
+                        'Contents', 'MacOS')
     PDFTOHTML = os.path.join(base, PDFTOHTML)
 if iswindows and hasattr(sys, 'frozen'):
-    base = sys.extensions_location if hasattr(sys, 'new_app_layout') else os.path.dirname(sys.executable)
+    base = os.path.dirname(sys.executable)
+    if hasattr(sys, 'new_app_layout'):
+        base = sys.extensions_location
     PDFTOHTML = os.path.join(base, 'pdftohtml.exe')
 if (islinux or isbsd) and getattr(sys, 'frozen', False):
     PDFTOHTML = os.path.join(sys.executables_location, 'bin', 'pdftohtml')
@@ -55,7 +58,7 @@ def pdftohtml(output_dir, pdf_path, no_images, as_xml=False):
 
         exe = PDFTOHTML
         cmd = [exe, '-enc', 'UTF-8', '-noframes', '-p', '-nomerge',
-                '-nodrm', a(pdfsrc), a(index)]
+               '-nodrm', a(pdfsrc), a(index)]
 
         if isbsd:
             cmd.remove('-nodrm')
@@ -67,7 +70,7 @@ def pdftohtml(output_dir, pdf_path, no_images, as_xml=False):
         logf = PersistentTemporaryFile('pdftohtml_log')
         try:
             p = popen(cmd, stderr=logf._fd, stdout=logf._fd,
-                    stdin=subprocess.PIPE)
+                      stdin=subprocess.PIPE)
         except OSError as err:
             if err.errno == errno.ENOENT:
                 raise ConversionError('Could not find pdftohtml, check it is '
@@ -79,7 +82,8 @@ def pdftohtml(output_dir, pdf_path, no_images, as_xml=False):
         logf.close()
         out = open(logf.name, 'rb').read().decode('utf-8', 'replace').strip()
         if ret != 0:
-            raise ConversionError('pdftohtml failed with return code: %d\n%s' % (ret, out))
+            raise ConversionError('pdftohtml failed with return code: '
+                                  '%d\n%s' % (ret, out))
         if out:
             prints("pdftohtml log:")
             prints(out)
@@ -90,22 +94,27 @@ def pdftohtml(output_dir, pdf_path, no_images, as_xml=False):
             with open(index, 'r+b') as i:
                 raw = i.read().decode('utf-8', 'replace')
                 raw = flip_images(raw)
-                raw = raw.replace('\n  \n  = 0.20 output self closing 
 tags, this
-                # breaks the pdf heuristics regexps, so replace them
+                # versions of pdftohtml >= 0.20 output self closing 
 tags,
+                # this breaks the pdf heuristics regexps, so replace them
                 raw = raw.replace('
', '
')
-                raw = re.sub(r' 2:
-            root = create_ncx(toc, (lambda x:x), 'pdftohtml', 'en', 'pdftohtml')
+            root = create_ncx(toc, (lambda x: x), 'pdftohtml', 'en',
+                              'pdftohtml')
             with open(os.path.join(output_dir, 'toc.ncx'), 'wb') as f:
-                f.write(etree.tostring(root, pretty_print=True, with_tail=False, encoding='utf-8', xml_declaration=True))
+                f.write(etree.tostring(root, pretty_print=True,
+                                       with_tail=False, encoding='utf-8',
+                                       xml_declaration=True))
 
 
 def flip_image(img, flip):
-    from ebook_converter.utils.img import flip_image, image_and_format_from_data, image_to_data
+    from ebook_converter.utils.img import image_to_data
+    from ebook_converter.utils.img import image_and_format_from_data
+    from ebook_converter.utils.img import flip_image
     with open(img, 'r+b') as f:
         img, fmt = image_and_format_from_data(f.read())
         img = flip_image(img, horizontal='x' in flip, vertical='y' in flip)
@@ -170,5 +183,5 @@ def flip_images(raw):
         if not os.path.exists(img):
             continue
         flip_image(img, flip)
-    raw = re.sub(r'\s*', '', raw, flags=re.I|re.DOTALL)
+    raw = re.sub(r'\s*', '', raw, flags=re.I | re.DOTALL)
     return raw
diff --git a/ebook_converter/ebooks/txt/markdownml.py b/ebook_converter/ebooks/txt/markdownml.py
index 1954179..0b862ee 100644
--- a/ebook_converter/ebooks/txt/markdownml.py
+++ b/ebook_converter/ebooks/txt/markdownml.py
@@ -5,8 +5,9 @@ import re
 
 from functools import partial
 
+from ebook_converter import constants as const
 from ebook_converter.ebooks.htmlz.oeb2html import OEB2HTML
-from ebook_converter.ebooks.oeb.base import XHTML, XHTML_NS, barename, namespace, rewrite_links
+from ebook_converter.ebooks.oeb.base import XHTML, barename, namespace, rewrite_links
 from ebook_converter.ebooks.oeb.stylizer import Stylizer
 
 
@@ -110,9 +111,9 @@ class MarkdownMLizer(OEB2HTML):
 
         # We can only processes tags. If there isn't a tag return any text.
         if not isinstance(elem.tag, (str, bytes)) \
-           or namespace(elem.tag) != XHTML_NS:
+           or namespace(elem.tag) != const.XHTML_NS:
             p = elem.getparent()
-            if p is not None and isinstance(p.tag, (str, bytes)) and namespace(p.tag) == XHTML_NS \
+            if p is not None and isinstance(p.tag, (str, bytes)) and namespace(p.tag) == const.XHTML_NS \
                     and elem.tail:
                 return [elem.tail]
             return ['']
diff --git a/ebook_converter/ebooks/txt/textileml.py b/ebook_converter/ebooks/txt/textileml.py
index 54e90d3..7b6f626 100644
--- a/ebook_converter/ebooks/txt/textileml.py
+++ b/ebook_converter/ebooks/txt/textileml.py
@@ -5,8 +5,10 @@ import re
 
 from functools import partial
 
+from ebook_converter import constants as const
 from ebook_converter.ebooks.htmlz.oeb2html import OEB2HTML
-from ebook_converter.ebooks.oeb.base import XHTML, XHTML_NS, barename, namespace, rewrite_links
+from ebook_converter.ebooks.oeb.base import XHTML, barename, namespace, \
+        rewrite_links
 from ebook_converter.ebooks.oeb.stylizer import Stylizer
 from ebook_converter.ebooks import unit_convert
 from ebook_converter.ebooks.textile.unsmarten import unsmarten
@@ -225,9 +227,9 @@ class TextileMLizer(OEB2HTML):
 
         # We can only processes tags. If there isn't a tag return any text.
         if not isinstance(elem.tag, (str, bytes)) \
-           or namespace(elem.tag) != XHTML_NS:
+           or namespace(elem.tag) != const.XHTML_NS:
             p = elem.getparent()
-            if p is not None and isinstance(p.tag, (str, bytes)) and namespace(p.tag) == XHTML_NS \
+            if p is not None and isinstance(p.tag, (str, bytes)) and namespace(p.tag) == const.XHTML_NS \
                     and elem.tail:
                 return [elem.tail]
             return ['']
diff --git a/ebook_converter/ebooks/txt/txtml.py b/ebook_converter/ebooks/txt/txtml.py
index 206b5a6..9b17a11 100644
--- a/ebook_converter/ebooks/txt/txtml.py
+++ b/ebook_converter/ebooks/txt/txtml.py
@@ -5,10 +5,11 @@ import re
 
 from lxml import etree
 
+from ebook_converter import constants as const
+from ebook_converter.ebooks.oeb import base
+from ebook_converter.ebooks.oeb import parse_utils
+from ebook_converter.ebooks.oeb.stylizer import Stylizer
 
-__license__ = 'GPL 3'
-__copyright__ = '2009, John Schember '
-__docformat__ = 'restructuredtext en'
 
 BLOCK_TAGS = [
     'div',
@@ -60,9 +61,6 @@ class TXTMLizer(object):
         return self.mlize_spine()
 
     def mlize_spine(self):
-        from ebook_converter.ebooks.oeb.base import XHTML
-        from ebook_converter.ebooks.oeb.stylizer import Stylizer
-        from ebook_converter.utils.xml_parse import safe_xml_fromstring
         output = [u'']
         output.append(self.get_toc())
         for item in self.oeb_book.spine:
@@ -72,9 +70,11 @@ class TXTMLizer(object):
                     x.text = x.text.replace('--', '__')
             content = etree.tostring(item.data, encoding='unicode')
             content = self.remove_newlines(content)
-            content = safe_xml_fromstring(content)
-            stylizer = Stylizer(content, item.href, self.oeb_book, self.opts, self.opts.output_profile)
-            output += self.dump_text(content.find(XHTML('body')), stylizer, item)
+            content = etree.fromstring(content)
+            stylizer = Stylizer(content, item.href, self.oeb_book, self.opts,
+                                self.opts.output_profile)
+            output += self.dump_text(content.find(base.tag('xhtml', 'body')),
+                                     stylizer, item)
             output += '\n\n\n\n\n\n'
         output = ''.join(output)
         output = '\n'.join(l.rstrip() for l in output.splitlines())
@@ -130,8 +130,12 @@ class TXTMLizer(object):
         text = re.sub('\n[ ]+\n', '\n\n', text)
         if self.opts.remove_paragraph_spacing:
             text = re.sub('\n{2,}', '\n', text)
-            text = re.sub(r'(?msu)^(?P[^\t\n]+?)$', lambda mo: u'%s\n\n' % mo.group('t'), text)
-            text = re.sub(r'(?msu)(?P[^\n])\n+(?P[^\t\n]+?)(?=\n)', lambda mo: '%s\n\n\n\n\n\n%s' % (mo.group('b'), mo.group('t')), text)
+            text = re.sub(r'(?msu)^(?P[^\t\n]+?)$', lambda mo: u'%s\n\n' %
+                          mo.group('t'), text)
+            text = re.sub(r'(?msu)(?P[^\n])\n+(?P[^\t\n]+?)(?=\n)',
+                          lambda mo: '%s\n\n\n\n\n\n%s' % (mo.group('b'),
+                                                           mo.group('t')),
+                          text)
         else:
             text = re.sub('\n{7,}', '\n\n\n\n\n\n', text)
 
@@ -146,7 +150,8 @@ class TXTMLizer(object):
 
         if self.opts.max_line_length:
             max_length = self.opts.max_line_length
-            if self.opts.max_line_length < 25 and not self.opts.force_max_line_length:
+            if (self.opts.max_line_length < 25 and not
+                    self.opts.force_max_line_length):
                 max_length = 25
             short_lines = []
             lines = text.splitlines()
@@ -186,13 +191,13 @@ class TXTMLizer(object):
         @stylizer: The style information attached to the element.
         @page: OEB page used to determine absolute urls.
         '''
-        from ebook_converter.ebooks.oeb.base import XHTML_NS, barename, namespace
 
         if not isinstance(elem.tag, (str, bytes)) \
-           or namespace(elem.tag) != XHTML_NS:
+           or parse_utils.namespace(elem.tag) != const.XHTML_NS:
             p = elem.getparent()
-            if p is not None and isinstance(p.tag, (str, bytes)) and namespace(p.tag) == XHTML_NS \
-                    and elem.tail:
+            if (p is not None and isinstance(p.tag, (str, bytes)) and
+                    parse_utils.namespace(p.tag) == const.XHTML_NS and
+                    elem.tail):
                 return [elem.tail]
             return ['']
 
@@ -205,14 +210,15 @@ class TXTMLizer(object):
                 return [elem.tail]
             return ['']
 
-        tag = barename(elem.tag)
+        tag = parse_utils.barename(elem.tag)
         tag_id = elem.attrib.get('id', None)
         in_block = False
         in_heading = False
 
         # Are we in a heading?
         # This can either be a heading tag or a TOC item.
-        if tag in HEADING_TAGS or '%s#%s' % (page.href, tag_id) in self.toc_ids:
+        if tag in HEADING_TAGS or '%s#%s' % (page.href,
+                                             tag_id) in self.toc_ids:
             in_heading = True
             if not self.last_was_heading:
                 text.append('\n\n\n\n\n\n')
@@ -234,7 +240,7 @@ class TXTMLizer(object):
             ems = int(round((float(style.marginTop) / style.fontSize) - 1))
             if ems >= 1:
                 text.append('\n' * ems)
-        except:
+        except Exception:
             pass
 
         # Process tags that contain text.
diff --git a/ebook_converter/startup.py b/ebook_converter/startup.py
index bfa6ce8..fbfb960 100644
--- a/ebook_converter/startup.py
+++ b/ebook_converter/startup.py
@@ -10,7 +10,7 @@ import builtins
 import locale
 import sys
 
-from ebook_converter import constants
+from ebook_converter import constants_old
 
 # For backwards compat with some third party plugins
 builtins.__dict__['dynamic_property'] = lambda func: func(None)
@@ -41,8 +41,8 @@ if not _run_once:
 
     #
     # Platform specific modules
-    if constants.iswindows:
-        winutil, winutilerror = constants.plugins['winutil']
+    if constants_old.iswindows:
+        winutil, winutilerror = constants_old.plugins['winutil']
         if not winutil:
             raise RuntimeError('Failed to load the winutil plugin: %s'%winutilerror)
         if len(sys.argv) > 1 and not isinstance(sys.argv[1], str):
@@ -57,8 +57,8 @@ if not _run_once:
 
     #
     # Convert command line arguments to unicode
-    enc = constants.preferred_encoding
-    if constants.isosx:
+    enc = constants_old.preferred_encoding
+    if constants_old.isosx:
         enc = 'utf-8'
     for i in range(1, len(sys.argv)):
         if not isinstance(sys.argv[i], str):
@@ -66,7 +66,7 @@ if not _run_once:
 
     #
     # Ensure that the max number of open files is at least 1024
-    if constants.iswindows:
+    if constants_old.iswindows:
         # See https://msdn.microsoft.com/en-us/library/6e3b887c.aspx
         if hasattr(winutil, 'setmaxstdio'):
             winutil.setmaxstdio(max(1024, winutil.getmaxstdio()))
@@ -77,7 +77,7 @@ if not _run_once:
             try:
                 resource.setrlimit(resource.RLIMIT_NOFILE, (min(1024, hard), hard))
             except Exception:
-                if constants.DEBUG:
+                if constants_old.DEBUG:
                     import traceback
                     traceback.print_exc()
 
@@ -122,7 +122,7 @@ if not _run_once:
         bound_signal.connect(slot, **kw)
     builtins.__dict__['connect_lambda'] = connect_lambda
 
-    if constants.islinux or constants.isosx or constants.isfreebsd:
+    if constants_old.islinux or constants_old.isosx or constants_old.isfreebsd:
         # Name all threads at the OS level created using the threading module, see
         # http://bugs.python.org/issue15500
         import threading
@@ -140,7 +140,7 @@ if not _run_once:
                 if name:
                     if isinstance(name, str):
                         name = name.encode('ascii', 'replace').decode('ascii')
-                    constants.plugins['speedup'][0].set_thread_name(name[:15])
+                    constants_old.plugins['speedup'][0].set_thread_name(name[:15])
             except Exception:
                 pass  # Don't care about failure to set name
         threading.Thread.start = new_start
@@ -152,7 +152,7 @@ def test_lopen():
     n = 'f\xe4llen'
     print('testing open()')
 
-    if constants.iswindows:
+    if constants_old.iswindows:
         import msvcrt, win32api
 
         def assert_not_inheritable(f):
diff --git a/ebook_converter/utils/xml_parse.py b/ebook_converter/utils/xml_parse.py
index c83bd5a..2b9a8ef 100644
--- a/ebook_converter/utils/xml_parse.py
+++ b/ebook_converter/utils/xml_parse.py
@@ -1,3 +1,7 @@
+import os
+import tempfile
+import unittest
+
 from lxml import etree
 
 
@@ -24,7 +28,6 @@ def safe_xml_fromstring(string_or_bytes, recover=True):
 
 
 def find_tests():
-    import unittest, tempfile, os
 
     class TestXMLParse(unittest.TestCase):
 
@@ -37,9 +40,11 @@ def find_tests():
             os.remove(self.temp_file)
 
         def test_safe_xml_fromstring(self):
-            templ = ''' ]>&e;'''
+            templ = ' ]>&e;'
             external = 'file:///' + self.temp_file.replace(os.sep, '/')
-            self.assertEqual(etree.fromstring(templ.format(id='SYSTEM', val=external)).text, 'external')
+            self.assertEqual(etree.fromstring(templ.format(id='SYSTEM',
+                                                           val=external)).text,
+                             'external')
             for eid, val, expected in (
                 ('', 'normal entity', 'normal entity'),
                 ('', external, external),
@@ -50,7 +55,8 @@ def find_tests():
                 ('PUBLIC', external, None),
                 ('PUBLIC', 'http://example.com', None),
             ):
-                got = getattr(safe_xml_fromstring(templ.format(id=eid, val=val)), 'text', None)
+                got = getattr(etree.fromstring(templ.format(id=eid, val=val)),
+                              'text', None)
                 self.assertEqual(got, expected)
 
     return unittest.defaultTestLoader.loadTestsFromTestCase(TestXMLParse)
diff --git a/ebook_converter/utils/zipfile.py b/ebook_converter/utils/zipfile.py
index d4dc785..b08d006 100644
--- a/ebook_converter/utils/zipfile.py
+++ b/ebook_converter/utils/zipfile.py
@@ -8,7 +8,7 @@ from contextlib import closing
 from tempfile import SpooledTemporaryFile
 
 from ebook_converter import sanitize_file_name
-from ebook_converter.constants import filesystem_encoding
+from ebook_converter.constants_old import filesystem_encoding
 from ebook_converter.ebooks.chardet import detect
 from ebook_converter.polyglot.builtins import as_bytes