gryf/ebook-converter
1
0
Fork 0
mirror of https://github.com/gryf/ebook-converter.git synced 2026-04-03 11:03:32 +02:00
Code Issues Projects Releases Wiki Activity

Next portion of logging adapt.

Browse Source
This commit is contained in:
gryf
2021-06-24 19:42:51 +02:00
parent 84abd62a61
commit 6dfcaea2dd
20 changed files with 105 additions and 122 deletions
Download Patch File Download Diff File Expand all files Collapse all files

20
ebook_converter/ebooks/conversion/plugins/txt_input.py
View File

@@ -167,13 +167,14 @@ class TXTInput(InputFormatPlugin):
if file_ext in {'md', 'textile', 'markdown'}:
options.formatting_type = {'md': 'markdown'}.get(file_ext, file_ext)
log.info('File extension indicates particular formatting. '
'Forcing formatting type to: %s'%options.formatting_type)
'Forcing formatting type to: %s',
options.formatting_type)
options.paragraph_type = 'off'
# Get the encoding of the document.
if options.input_encoding:
ienc = options.input_encoding
log.debug('Using user specified input encoding of %s' % ienc)
log.debug('Using user specified input encoding of %s', ienc)
else:
det_encoding = detect(txt[:4096])
det_encoding, confidence = det_encoding['encoding'], det_encoding['confidence']
@@ -184,10 +185,12 @@ class TXTInput(InputFormatPlugin):
# gb2312 instead of gbk. gbk is a superset of gb2312, anyway.
det_encoding = 'gbk'
ienc = det_encoding
log.debug('Detected input encoding as %s with a confidence of %s%%' % (ienc, confidence * 100))
log.debug('Detected input encoding as %s with a confidence of '
'%s%%', ienc, confidence * 100)
if not ienc:
ienc = 'utf-8'
log.debug('No input encoding specified and could not auto detect using %s' % ienc)
log.debug('No input encoding specified and could not auto detect '
'using %s', ienc)
# Remove BOM from start of txt as its presence can confuse markdown
import codecs
for bom in (codecs.BOM_UTF16_LE, codecs.BOM_UTF16_BE, codecs.BOM_UTF8, codecs.BOM_UTF32_LE, codecs.BOM_UTF32_BE):
@@ -206,15 +209,18 @@ class TXTInput(InputFormatPlugin):
if options.paragraph_type == 'auto':
options.paragraph_type = detect_paragraph_type(txt)
if options.paragraph_type == 'unknown':
log.debug('Could not reliably determine paragraph type using block')
log.debug('Could not reliably determine paragraph type using '
'block')
options.paragraph_type = 'block'
else:
log.debug('Auto detected paragraph type as %s' % options.paragraph_type)
log.debug('Auto detected paragraph type as %s',
options.paragraph_type)
# Detect formatting
if options.formatting_type == 'auto':
options.formatting_type = detect_formatting_type(txt)
log.debug('Auto detected formatting as %s' % options.formatting_type)
log.debug('Auto detected formatting as %s',
options.formatting_type)
if options.formatting_type == 'heuristic':
setattr(options, 'enable_heuristics', True)