diff --git a/src/calibre/ebooks/epub/split.py b/src/calibre/ebooks/epub/split.py
index 33c6360495..38f51ae3be 100644
--- a/src/calibre/ebooks/epub/split.py
+++ b/src/calibre/ebooks/epub/split.py
@@ -7,7 +7,7 @@ __docformat__ = 'restructuredtext en'
Split the flows in an epub file to conform to size limitations.
'''
-import os, math, copy, logging, functools
+import os, math, copy, logging, functools, collections
from lxml.etree import XPath as _XPath
from lxml import etree, html
@@ -234,7 +234,7 @@ class Splitter(LoggingInterface):
all anchors in the original tree. Internal links are re-directed. The
original file is deleted and the split files are saved.
'''
- self.anchor_map = {None:self.base%0}
+ self.anchor_map = collections.defaultdict(lambda :self.base%0)
self.files = []
for i, tree in enumerate(self.trees):
diff --git a/src/calibre/ebooks/html.py b/src/calibre/ebooks/html.py
index 147e4a8a79..1433976113 100644
--- a/src/calibre/ebooks/html.py
+++ b/src/calibre/ebooks/html.py
@@ -252,15 +252,7 @@ def opf_traverse(opf_reader, verbose=0, encoding=None):
class PreProcessor(object):
PREPROCESS = []
- # Fix Baen markup
- BAEN = [
- (re.compile(r'page-break-before:\s*\w+([\s;\}])', re.IGNORECASE),
- lambda match: match.group(1)),
- (re.compile(r'\s*(\s*)\s*
', re.IGNORECASE),
- lambda match: match.group(1)),
- (re.compile(r'<\s*a\s+id="p[0-9]+"\s+name="p[0-9]+"\s*>\s*', re.IGNORECASE),
- lambda match: ''),
- ]
+
# Fix pdftohtml markup
PDFTOHTML = [
# Remove
tags
@@ -275,6 +267,9 @@ class PreProcessor(object):
# Remove hyphenation
(re.compile(r'-\n\r?'), lambda match: ''),
+ # Remove gray background
+ (re.compile(r']+>'), lambda match : '')
+
]
# Fix Book Designer markup
@@ -305,7 +300,7 @@ class PreProcessor(object):
def preprocess(self, html):
if self.is_baen(html):
- rules = self.BAEN
+ rules = []
elif self.is_book_designer(html):
rules = self.BOOK_DESIGNER
elif self.is_pdftohtml(html):