PML Output: Generate \CX Tags as chapter anchors. PDB eReader Output: Use \CX tags to generate chapter index.

src/calibre/ebooks/pdb/ereader/writer.py

@@ -42,8 +42,8 @@ class Writer(FormatWriter):
         pml = unicode(pmlmlizer.extract_content(oeb_book, self.opts)).encode('cp1252', 'replace')
 
         text, text_sizes = self._text(pml)
-        chapter_index = self._chapter_index(pml)
+        chapter_index = self._index_item(r'(?s)\\C(?P<val>\d)="(?P<text>.+?)"', pml)
-        link_index = self._link_index(pml)
+        link_index = self._index_item(r'(?s)\\Q="(?P<text>.+?)"', pml)
         images = self._images(oeb_book.manifest, pmlmlizer.image_hrefs)
         metadata = [self._metadata(metadata)]
         hr = [self._header_record(len(text), len(chapter_index), len(link_index), len(images))]
@@ -101,38 +101,24 @@ class Writer(FormatWriter):
 
         return pml_pages, text_sizes
 
-    def _index_item(self, mo):
+    def _index_item(self, regex, pml):
-        index = ''
-        if 'text' in mo.groupdict().keys():
-            index += struct.pack('>L', mo.start())
-            text = mo.group('text')
-            # Strip all PML tags from text
-            text = re.sub(r'\\U[0-9a-z]{4}', '', text)
-            text = re.sub(r'\\a\d{3}', '', text)
-            text = re.sub(r'\\.', '', text)
-            # Add appropriate spacing to denote the various levels of headings
-            if 'val' in mo.groupdict().keys():
-                text = '%s%s' % (' ' * 4 * int(mo.group('val')), text)
-            index += text
-            index += '\x00'
-        return index
-
-    def _chapter_index(self, pml):
-        chapter_marks = [
-            r'(?s)\\x(?P<text>.+?)\\x',
-            r'(?s)\\X(?P<val>[0-4])(?P<text>.*?)\\X[0-4]',
-            r'(?s)\\C(?P<val>\d)="(?P<text>.+?)"',
-        ]
         index = []
-        for chapter_mark in chapter_marks:
+        for mo in re.finditer(regex, pml):
-            for mo in re.finditer(chapter_mark, pml):
+            item = ''
-                index.append(self._index_item(mo))
+            if 'text' in mo.groupdict().keys():
-        return index
+                item += struct.pack('>L', mo.start())
-
+                text = mo.group('text')
-    def _link_index(self, pml):
+                # Strip all PML tags from text
-        index = []
+                text = re.sub(r'\\U[0-9a-z]{4}', '', text)
-        for mo in re.finditer(r'(?s)\\Q="(?P<text>.+?)"', pml):
+                text = re.sub(r'\\a\d{3}', '', text)
-            index.append(self._index_item(mo))
+                text = re.sub(r'\\.', '', text)
+                # Add appropriate spacing to denote the various levels of headings
+                if 'val' in mo.groupdict().keys():
+                    text = '%s%s' % (' ' * 4 * int(mo.group('val')), text)
+                item += text
+                item += '\x00'
+            if item:
+                index.append(item)
         return index
 
     def _images(self, manifest, image_hrefs):

src/calibre/ebooks/pml/pmlml.py

@@ -79,6 +79,16 @@ class PMLMLizer(object):
         self.log.info('Converting XHTML to PML markup...')
         self.oeb_book = oeb_book
         self.opts = opts
+
+        # This is used for adding \CX tags chapter markers. This is separate
+        # from the optional inline toc.
+        self.toc = {}
+        for item in oeb_book.toc:
+            page, mid, id = item.href.partition('#')
+            if not self.toc.get(page, None):
+                self.toc[page] = {}
+            self.toc[page][id] = item.title
+
         return self.pmlmlize_spine()
 
     def pmlmlize_spine(self):
@@ -107,7 +117,11 @@ class PMLMLizer(object):
         return output
 
     def get_toc(self):
-        toc = [u'']
+        '''
+        Generation of inline TOC
+        '''
+
+        toc = []
         if self.opts.inline_toc:
             self.log.debug('Generating table of contents...')
             toc.append(u'\\X0%s\\X0\n\n' % _('Table of Contents:'))
@@ -177,14 +191,14 @@ class PMLMLizer(object):
     def dump_text(self, elem, stylizer, page, tag_stack=[]):
         if not isinstance(elem.tag, basestring) \
            or namespace(elem.tag) != XHTML_NS:
-            return [u'']
+            return []
 
-        text = [u'']
+        text = []
         style = stylizer.style(elem)
 
         if style['display'] in ('none', 'oeb-page-head', 'oeb-page-foot') \
            or style['visibility'] == 'hidden':
-            return [u'']
+            return []
 
         tag = barename(elem.tag)
         tag_count = 0
@@ -213,6 +227,12 @@ class PMLMLizer(object):
             else:
                 w += '="50%"'
             text.append(w)
+        toc_id = elem.attrib.get('id', None)
+        if toc_id:
+            if self.toc.get(page.href, None):
+                toc_title = self.toc[page.href].get(toc_id, None)
+                if toc_title:
+                    text.append('\\C1="%s"' % toc_title)
 
         # Process style information that needs holds a single tag
         # Commented out because every page in an OEB book starts with this style
@@ -287,4 +307,3 @@ class PMLMLizer(object):
             if tag != 'block':
                 text.append('\\%s' % tag)
         return text
-