IGN:Improved TOC detection for html2epub and miscellaneous bug fixes

src/calibre/__init__.py

@@ -289,7 +289,7 @@ class LoggingInterface:
         formatter = logging.Formatter()
         level     = logging.INFO
         if verbosity > 0:
-            formatter = ColoredFormatter('[%(levelname)s] %(filename)s:%(lineno)s: %(message)s') if verbosity > 1 else \
+            formatter = ColoredFormatter('[%(levelname)s] %(message)s') if verbosity > 1 else \
                         ColoredFormatter('%(levelname)s: %(message)s')
             level     = logging.DEBUG
             if verbosity > 1:

src/calibre/ebooks/epub/from_any.py

@@ -74,7 +74,7 @@ MAP = {
        'txt'  : txt2opf,
        'pdf'  : pdf2opf,
        }
-SOURCE_FORMATS = ['lit', 'mobi', 'prc', 'fb2', 'rtf', 'txt', 'pdf', 'rar', 'zip']    
+SOURCE_FORMATS = ['lit', 'mobi', 'prc', 'fb2', 'rtf', 'txt', 'pdf', 'rar', 'zip', 'htm', 'html']    
 
 def unarchive(path, tdir):
     extract(path, tdir)

src/calibre/ebooks/epub/from_html.py

@@ -149,7 +149,7 @@ def convert(htmlfile, opts, notification=None):
         buf = cStringIO.StringIO()
         if mi.toc:
             rebase_toc(mi.toc, htmlfile_map, tdir)
-        if opts.use_auto_toc or mi.toc is None or len(mi.toc) < 2:
+        if opts.use_auto_toc or mi.toc is None or len(list(mi.toc.flat())) < 2:
             mi.toc = generated_toc
         for item in mi.manifest:
             if getattr(item, 'mime_type', None) == 'text/html':
@@ -184,4 +184,4 @@ def main(args=sys.argv):
     return 0
     
 if __name__ == '__main__':
-    sys.exit(main())
+    sys.exit(main())

src/calibre/ebooks/epub/split.py

@@ -48,8 +48,14 @@ class Splitter(LoggingInterface):
         css = XPath('//link[@type = "text/css" and @rel = "stylesheet"]')(root)
         if css:
             cssp = os.path.join('content', *(css[0].get('href').split('/')))
-            self.log_debug('\t\tParsing stylesheet...') 
+            self.log_debug('\t\tParsing stylesheet...')
-            stylesheet = CSSParser().parseString(open(cssp, 'rb').read())
+            try: 
+                stylesheet = CSSParser().parseString(open(cssp, 'rb').read())
+            except:
+                self.log_warn('Failed to parse CSS. Splitting on page-breaks is disabled')
+                if self.opts.verbose > 1:
+                    self.log_exception('')
+                stylesheet = None
         else:
             stylesheet = None
         self.page_breaks = []

src/calibre/ebooks/html.py

@@ -123,6 +123,8 @@ class HTMLFile(object):
             if encoding is None:
                 encoding = xml_to_unicode(src[:4096], verbose=verbose)[-1]
                 self.encoding = encoding
+            else:
+                self.encoding = encoding
 
             src = src.decode(encoding, 'replace')
             match = self.TITLE_PAT.search(src)
@@ -200,6 +202,8 @@ def traverse(path_to_html_file, max_levels=sys.maxint, verbose=0, encoding=None)
                     continue
                 try:
                     nf = HTMLFile(link.path, level, encoding, verbose, referrer=hf)
+                    if nf.is_binary:
+                        raise IgnoreFile('%s is a binary file'%nf.path, -1)
                     nl.append(nf)
                     flat.append(nf)
                 except IgnoreFile, err:
@@ -235,7 +239,7 @@ def opf_traverse(opf_reader, verbose=0, encoding=None):
             if path not in flat:
                 flat.append(path)
     flat = [HTMLFile(path, 0, encoding, verbose) for path in flat]
-    return flat
+    return [f for f in flat if not f.is_binary]
             
 
 
@@ -521,8 +525,6 @@ class Processor(Parser):
         Remove all CSS information from the document and store in self.raw_css. 
         This includes <font> tags.
         '''
-        counter = 0
-        
         def get_id(chapter, counter, prefix='calibre_css_'):
             new_id = '%s_%d'%(prefix, counter)
             if chapter.tag.lower() == 'a' and  'name' in chapter.keys():
@@ -667,7 +669,10 @@ def get_filelist(htmlfile, opts):
     opf = search_for_opf(dir)
     filelist = None
     if opf is not None:
-        filelist = opf_traverse(opf, verbose=opts.verbose, encoding=opts.encoding)
+        try:
+            filelist = opf_traverse(opf, verbose=opts.verbose, encoding=opts.encoding)
+        except:
+            pass
     if not filelist:
         filelist = traverse(htmlfile, max_levels=int(opts.max_levels), 
                             verbose=opts.verbose, encoding=opts.encoding)\

src/calibre/ebooks/metadata/opf2.py

@@ -7,7 +7,7 @@ __docformat__ = 'restructuredtext en'
 lxml based OPF parser.
 '''
 
-import sys, unittest, functools, os, mimetypes, uuid
+import sys, unittest, functools, os, mimetypes, uuid, glob
 from urllib import unquote
 from urlparse import urlparse
 
@@ -446,10 +446,43 @@ class OPF(object):
             self.spine = Spine.from_opf_spine_element(s, self.manifest)
         self.guide = None
         guide = self.guide_path(self.root)
-        if guide:
+        self.guide = Guide.from_opf_guide(guide, basedir) if guide else None
-            self.guide = Guide.from_opf_guide(guide, basedir)
         self.cover_data = (None, None)
+        self.find_toc()
         
+    def find_toc(self):
+        self.toc = None
+        try:
+            spine = self.XPath('descendant::*[re:match(name(), "spine", "i")]')(self.root)
+            toc = None
+            if spine:
+                spine = spine[0]
+                toc = spine.get('toc', None)
+            if toc is None and self.guide:
+                for item in self.guide:
+                    if item.type and item.type.lower() == 'toc':
+                        toc = item.path
+            if toc is None:
+                for item in self.manifest:
+                    if 'toc' in item.href().lower():
+                        toc = item.path
+            
+            if toc is None: return
+            self.toc = TOC(base_path=self.base_dir)
+            if toc.lower() in ('ncx', 'ncxtoc'):
+                path = self.manifest.path_for_id(toc)
+                if path:
+                    self.toc.read_ncx_toc(path)
+                else:
+                    f = glob.glob(os.path.join(self.base_dir, '*.ncx'))
+                    if f:
+                        self.toc.read_ncx_toc(f[0])
+            else:
+                self.toc.read_html_toc(toc)
+        except:
+            pass        
+            
+    
         
     def get_text(self, elem):
         return u''.join(self.TEXT(elem))