MOBI news download: Allow recipes to set a thumbnail for entries in the periodical table of contents. Fixes #900130 (Kindle MOBI TOC thumbnail)

src/calibre/ebooks/metadata/toc.py

@@ -31,7 +31,7 @@ class TOC(list):
 
     def __init__(self, href=None, fragment=None, text=None, parent=None, play_order=0,
                  base_path=os.getcwd(), type='unknown', author=None,
-                 description=None):
+                 description=None, toc_thumbnail=None):
         self.href = href
         self.fragment = fragment
         if not self.fragment:
@@ -43,6 +43,7 @@ class TOC(list):
         self.type = type
         self.author = author
         self.description = description
+        self.toc_thumbnail = toc_thumbnail
 
     def __str__(self):
         lines = ['TOC: %s#%s'%(self.href, self.fragment)]
@@ -72,12 +73,12 @@ class TOC(list):
         entry.parent = None
 
     def add_item(self, href, fragment, text, play_order=None, type='unknown',
-            author=None, description=None):
+            author=None, description=None, toc_thumbnail=None):
         if play_order is None:
             play_order = (self[-1].play_order if len(self) else self.play_order) + 1
         self.append(TOC(href=href, fragment=fragment, text=text, parent=self,
                         base_path=self.base_path, play_order=play_order,
-                        type=type, author=author, description=description))
+                        type=type, author=author, description=description, toc_thumbnail=toc_thumbnail))
         return self[-1]
 
     def top_level_items(self):
@@ -269,6 +270,9 @@ class TOC(list):
             if desc:
                 desc = re.sub(r'\s+', ' ', desc)
                 elem.append(C.meta(desc, name='description'))
+            idx = getattr(np, 'toc_thumbnail', None)
+            if idx:
+                elem.append(C.meta(idx, name='toc_thumbnail'))   
             parent.append(elem)
             for np2 in np:
                 navpoint(elem, np2)

src/calibre/ebooks/mobi/debug.py

@@ -656,11 +656,11 @@ class Tag(object): # {{{
                         ' image record associated with this article',
                         'image_index'),
                     70 : ('Description offset in cncx', 'desc_offset'),
-                    71 : ('Image attribution offset in cncx',
-                        'image_attr_offset'),
+                    71 : ('Author offset in cncx', 'author_offset'),
                     72 : ('Image caption offset in cncx',
                         'image_caption_offset'),
-                    73 : ('Author offset in cncx', 'author_offset'),
+                    73 : ('Image attribution offset in cncx',
+                        'image_attr_offset'),
             },
 
             'chapter_with_subchapters' : {

src/calibre/ebooks/mobi/writer2/indexer.py

@@ -136,7 +136,8 @@ class IndexEntry(object):
             'last_child_index': 23,
             'image_index': 69,
             'desc_offset': 70,
-            'author_offset': 73,
+            'author_offset': 71,
+
     }
     RTAG_MAP = {v:k for k, v in TAG_VALUES.iteritems()}
 
@@ -754,6 +755,13 @@ class Indexer(object): # {{{
                 normalized_articles.append(article)
                 article.author_offset = self.cncx[art.author]
                 article.desc_offset = self.cncx[art.description]
+                if getattr(art, 'toc_thumbnail', None) is not None:
+                    try:
+                        ii = self.serializer.images[art.toc_thumbnail] - 1
+                        if ii > -1:
+                            article.image_index = ii
+                    except KeyError:
+                        pass # Image not found in serializer
 
             if normalized_articles:
                 normalized_articles.sort(key=lambda x:x.offset)

src/calibre/ebooks/oeb/base.py

@@ -1629,9 +1629,10 @@ class TOC(object):
     :attr:`id`: Option unique identifier for this node.
     :attr:`author`: Optional author attribution for periodicals <mbp:>
     :attr:`description`: Optional description attribute for periodicals <mbp:>
+    :attr:`toc_thumbnail`: Optional toc thumbnail image
     """
     def __init__(self, title=None, href=None, klass=None, id=None,
-            play_order=None, author=None, description=None):
+            play_order=None, author=None, description=None, toc_thumbnail=None):
         self.title = title
         self.href = urlnormalize(href) if href else href
         self.klass = klass
@@ -1643,10 +1644,11 @@ class TOC(object):
         self.play_order = play_order
         self.author = author
         self.description = description
+        self.toc_thumbnail = toc_thumbnail
 
-    def add(self, title, href, klass=None, id=None, play_order=0, author=None, description=None):
+    def add(self, title, href, klass=None, id=None, play_order=0, author=None, description=None, toc_thumbnail=None):
         """Create and return a new sub-node of this node."""
-        node = TOC(title, href, klass, id, play_order, author, description)
+        node = TOC(title, href, klass, id, play_order, author, description, toc_thumbnail)
         self.nodes.append(node)
         return node
 

src/calibre/ebooks/oeb/reader.py

@@ -371,8 +371,15 @@ class OEBReader(object):
             else :
                 description = None
 
+            index_image = xpath(child,
+                    'descendant::calibre:meta[@name = "toc_thumbnail"]')
+            toc_thumbnail = (index_image[0].text if index_image else None)
+            if not toc_thumbnail or not toc_thumbnail.strip():
+                toc_thumbnail = None
+
             node = toc.add(title, href, id=id, klass=klass,
-                    play_order=po, description=description, author=author)
+                    play_order=po, description=description, author=author,
+                           toc_thumbnail=toc_thumbnail)
 
             self._toc_from_navpoint(item, node, child)
 

src/calibre/web/feeds/__init__.py

@@ -31,6 +31,7 @@ class Article(object):
         self._title = clean_ascii_chars(self._title)
         self.url = url
         self.author = author
+        self.toc_thumbnail = None
         if author and not isinstance(author, unicode):
             author = author.decode('utf-8', 'replace')
         self.summary = summary
@@ -88,8 +89,10 @@ URL         : %s
 Author      : %s
 Summary     : %s
 Date        : %s
+TOC thumb   : %s
 Has content : %s
-'''%(self.title, self.url, self.author, self.summary[:20]+'...', self.localtime.strftime('%a, %d %b, %Y %H:%M'),
+'''%(self.title, self.url, self.author, self.summary[:20]+'...',
+    self.toc_thumbnail, self.localtime.strftime('%a, %d %b, %Y %H:%M'),
      bool(self.content))).encode('utf-8')
 
     def __str__(self):

src/calibre/web/feeds/news.py

@@ -653,6 +653,25 @@ class BasicNewsRecipe(Recipe):
         '''
         raise NotImplementedError
 
+    def add_toc_thumbnail(self, article, src):
+        '''
+        Call this from populate_article_metadata with the src attribute of an
+        <img> tag from the article that is appropriate for use as the thumbnail
+        representing the article in the Table of Contents. Whether the
+        thumbnail is actually used is device dependent (currently only used by
+        the Kindles). Note that the referenced image must be one that was
+        successfully downloaded, otherwise it will be ignored.
+        '''
+        if not src or not hasattr(article, 'toc_thumbnail'):
+            return
+
+        src = src.replace('\\', '/')
+        if re.search(r'feed_\d+/article_\d+/images/img', src, flags=re.I) is None:
+            self.log.warn('Ignoring invalid TOC thumbnail image: %r'%src)
+            return
+        article.toc_thumbnail = re.sub(r'^.*?feed', 'feed',
+                src, flags=re.IGNORECASE)
+
     def populate_article_metadata(self, article, soup, first):
         '''
         Called when each HTML page belonging to article is downloaded.
@@ -1285,13 +1304,16 @@ class BasicNewsRecipe(Recipe):
                         desc = None
                     else:
                         desc = self.description_limiter(desc)
+                    tt = a.toc_thumbnail if a.toc_thumbnail else None
                     entries.append('%sindex.html'%adir)
                     po = self.play_order_map.get(entries[-1], None)
                     if po is None:
                         self.play_order_counter += 1
                         po = self.play_order_counter
-                    parent.add_item('%sindex.html'%adir, None, a.title if a.title else _('Untitled Article'),
-                                    play_order=po, author=auth, description=desc)
+                    parent.add_item('%sindex.html'%adir, None,
+                            a.title if a.title else _('Untitled Article'),
+                            play_order=po, author=auth,
+                            description=desc, toc_thumbnail=tt)
                     last = os.path.join(self.output_dir, ('%sindex.html'%adir).replace('/', os.sep))
                     for sp in a.sub_pages:
                         prefix = os.path.commonprefix([opf_path, sp])