Move force_uniceode to utils package

ebook_converter/utils/config_base.py

@@ -7,10 +7,11 @@ import json
 import numbers
 import os
 import pickle
-import pkg_resources
 import re
 import traceback
 
+import pkg_resources
+
 from ebook_converter.constants_old import config_dir
 from ebook_converter.constants_old import filesystem_encoding
 from ebook_converter.constants_old import preferred_encoding
@@ -75,7 +76,8 @@ def from_json(obj):
     return obj
 
 
-def force_unicode(x):
+def _force_unicode(x):
+    # TODO(gryf): eliminate cases, when this kind of functions are needed.
     try:
         return x.decode(preferred_encoding)
     except UnicodeDecodeError:
@@ -87,7 +89,7 @@ def force_unicode(x):
 
 def force_unicode_recursive(obj):
     if isinstance(obj, bytes):
-        return force_unicode(obj)
+        return _force_unicode(obj)
     if isinstance(obj, (list, tuple)):
         return type(obj)(map(force_unicode_recursive, obj))
     if isinstance(obj, dict):

ebook_converter/utils/encoding.py

@@ -0,0 +1,20 @@
+from ebook_converter import constants_old
+
+
+def force_unicode(obj, enc=constants_old.preferred_encoding):
+    if isinstance(obj, bytes):
+        try:
+            obj = obj.decode(enc)
+        except Exception:
+            try:
+                obj = obj.decode(constants_old.filesystem_encoding
+                                 if enc == constants_old.preferred_encoding
+                                 else constants_old.preferred_encoding)
+            except Exception:
+                try:
+                    obj = obj.decode('utf-8')
+                except Exception:
+                    obj = repr(obj)
+                    if isinstance(obj, bytes):
+                        obj = obj.decode('utf-8')
+    return obj

ebook_converter/utils/filenames.py

@@ -9,9 +9,9 @@ import shutil
 from math import ceil
 
 from ebook_converter import constants_old
-from ebook_converter import force_unicode
 from ebook_converter.constants_old import (filesystem_encoding,
                                            preferred_encoding)
+from ebook_converter.utils import encoding as uenc
 from ebook_converter.utils.localization import get_udc
 
 
@@ -183,7 +183,7 @@ def case_preserving_open_file(path, mode='wb', mkdir_mode=0o777):
 
     path = os.path.abspath(path)
 
-    sep = force_unicode(os.sep, 'ascii')
+    sep = uenc.force_unicode(os.sep, 'ascii')
 
     if path.endswith(sep):
         path = path[:-1]

ebook_converter/utils/img.py

@@ -12,11 +12,11 @@ from threading import Thread
 #from PyQt5.QtCore import QBuffer, QByteArray, Qt
 #from PyQt5.QtGui import QColor, QImage, QImageReader, QImageWriter, QPixmap, QTransform
 
-from ebook_converter import force_unicode
 from ebook_converter.constants_old import plugins
 from ebook_converter.ptempfile import TemporaryDirectory
 from ebook_converter.utils.config_base import tweaks
 from ebook_converter.utils.filenames import atomic_rename
+from ebook_converter.utils import encoding as uenc
 from ebook_converter.utils.imghdr import what
 
 # Utilities {{{
@@ -586,7 +586,7 @@ def run_optimizer(file_path, cmd, as_filter=False, input_data=None):
             outw = Thread(name='CopyOutput', target=copy, args=(p.stdout, outf))
             outw.daemon = True
             outw.start()
-        raw = force_unicode(stderr.read())
+        raw = uenc.force_unicode(stderr.read())
         if p.wait() != 0:
             return raw
         else: