Fix handling of images and anchors for RB, FB2 and eReader output

src/calibre/ebooks/fb2/fb2ml.py

@@ -38,6 +38,14 @@ TAG_SPACE = [
     'br',
 ]
 
+TAG_IMAGES = [
+    'img',
+]
+
+TAG_LINKS = [
+    'a',
+]
+
 STYLES = [
     ('font-weight', {'bold'   : 'strong', 'bolder' : 'strong'}),
     ('font-style', {'italic' : 'emphasis'}),
@@ -48,6 +56,7 @@ class FB2MLizer(object):
     def __init__(self, log):
         self.log = log
         self.image_hrefs = {}
+        self.link_hrefs = {}
         
     def extract_content(self, oeb_book, opts):
         self.log.info('Converting XHTML to FB2 markup...')
@@ -57,6 +66,7 @@ class FB2MLizer(object):
         
     def fb2mlize_spine(self):
         self.image_hrefs = {}
+        self.link_hrefs = {}
         output = self.fb2_header()
         if 'titlepage' in self.oeb_book.guide:
             self.log.debug('Generating cover page...')
@@ -68,6 +78,7 @@ class FB2MLizer(object):
         for item in self.oeb_book.spine:
             self.log.debug('Converting %s to FictionBook2 XML' % item.href)
             stylizer = Stylizer(item.data, item.href, self.oeb_book, self.opts.output_profile)
+            output += self.add_page_anchor(item)
             output += self.dump_text(item.data.find(XHTML('body')), stylizer, item)
         output += self.fb2_body_footer()
         output += self.fb2mlize_images()
@@ -82,7 +93,7 @@ class FB2MLizer(object):
         
         if len(author_parts) == 1:
             author_last = author_parts[0]
-        elif len(author_parts == 2):
+        elif len(author_parts) == 2:
             author_first = author_parts[0]
             author_last = author_parts[1]
         else:
@@ -108,6 +119,17 @@ class FB2MLizer(object):
     def fb2_footer(self):
         return u'</FictionBook>'
 
+    def add_page_anchor(self, page):
+        return self.get_anchor(page, '')
+
+    def get_anchor(self, page, aid):
+        aid = prepare_string_for_xml(aid)
+        aid = '%s#%s' % (page.href, aid)
+        if aid not in self.link_hrefs.keys():
+            self.link_hrefs[aid] = 'calibre_link-%s' % len(self.link_hrefs.keys())
+        aid = self.link_hrefs[aid]
+        return '<v id="%s"></v>' % aid
+
     def fb2mlize_images(self):
         images = u''
         for item in self.oeb_book.manifest:
@@ -149,10 +171,32 @@ class FB2MLizer(object):
         tag = barename(elem.tag)
         tag_count = 0
 
-        if tag == 'img':
+        if tag in TAG_IMAGES:
-            if page.abshref(elem.attrib['src']) not in self.image_hrefs.keys():
+            if elem.attrib.get('src', None):
-                self.image_hrefs[page.abshref(elem.attrib['src'])] = '%s.jpg' % len(self.image_hrefs.keys())
+                if page.abshref(elem.attrib['src']) not in self.image_hrefs.keys():
-            fb2_text += '<image xlink:href="#%s" />' % self.image_hrefs[page.abshref(elem.attrib['src'])]
+                    self.image_hrefs[page.abshref(elem.attrib['src'])] = '%s.jpg' % len(self.image_hrefs.keys())
+                fb2_text += '<image xlink:href="#%s" />' % self.image_hrefs[page.abshref(elem.attrib['src'])]
+
+        if tag in TAG_LINKS:
+            href = elem.get('href')
+            if href:
+                href = prepare_string_for_xml(page.abshref(href))
+                if '://' in href:
+                    fb2_text += '<a xlink:href="%s">' % href
+                else:
+                    if '#' not in href:
+                        href += '#'
+                    if href not in self.link_hrefs.keys():
+                        self.link_hrefs[href] = 'calibre_link-%s' % len(self.link_hrefs.keys())
+                    href = self.link_hrefs[href]
+                    fb2_text += '<a xlink:href="#%s">' % href
+                tag_count += 1
+                tag_stack.append('a')
+
+        # Anchor ids
+        id_name = elem.get('id')
+        if id_name:
+            fb2_text += self.get_anchor(page, id_name)
 
         fb2_tag = TAG_MAP.get(tag, None)
         if fb2_tag and fb2_tag not in tag_stack:

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

@@ -9,7 +9,7 @@ import os
 class EreaderError(Exception):
     pass
     
-def image_name(name):
+def image_name(name, taken_names=[]):
     name = os.path.basename(name)
     
     if len(name) > 32:
@@ -18,6 +18,10 @@ def image_name(name):
         namee = name[10+cut:]
         name = names + namee
     
+    while name in taken_names:
+        for i in xrange(9999999999999999999999999999999):
+            name = '%s%s' % (name[:-len('%s' % i)], i)
+
     name = name.ljust(32, '\x00')[:32]
     
     return name

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

@@ -22,7 +22,6 @@ import cStringIO
 from calibre.ebooks.pdb.formatwriter import FormatWriter
 from calibre.ebooks.oeb.base import OEB_RASTER_IMAGES
 from calibre.ebooks.pdb.header import PdbHeaderBuilder
-from calibre.ebooks.pdb.ereader import image_name
 from calibre.ebooks.pml.pmlml import PMLMLizer
 
 IDENTITY = 'PNRdPPrs'
@@ -38,8 +37,8 @@ class Writer(FormatWriter):
         self.log = log
 
     def write_content(self, oeb_book, out_stream, metadata=None):
-        text = self._text(oeb_book)
+        text, image_hrefs = self._text(oeb_book)
-        images = self._images(oeb_book.manifest)
+        images = self._images(oeb_book.manifest, image_hrefs)
         metadata = [self._metadata(metadata)]
 
         hr = [self._header_record(len(text), len(images))]
@@ -66,13 +65,13 @@ class Writer(FormatWriter):
         for i in range(0, (len(pml) / MAX_RECORD_SIZE) + 1):
             pml_pages.append(zlib.compress(pml[i * MAX_RECORD_SIZE : (i * MAX_RECORD_SIZE) + MAX_RECORD_SIZE]))
 
-        return pml_pages
+        return pml_pages, pmlmlizer.image_hrefs
 
-    def _images(self, manifest):
+    def _images(self, manifest, image_hrefs):
         images = []
 
         for item in manifest:
-            if item.media_type in OEB_RASTER_IMAGES:
+            if item.media_type in OEB_RASTER_IMAGES and item.href in image_hrefs.keys():
                 try:
                     im = Image.open(cStringIO.StringIO(item.data)).convert('P')
                     im.thumbnail((300,300), Image.ANTIALIAS)
@@ -82,7 +81,7 @@ class Writer(FormatWriter):
                     data = data.getvalue()
 
                     header = 'PNG '
-                    header += image_name(item.href)
+                    header += image_hrefs[item.href].ljust(32, '\x00')[:32]
                     header = header.ljust(62, '\x00')
 
                     if len(data) + len(header) < 65505:

src/calibre/ebooks/pml/pmlml.py

@@ -53,6 +53,10 @@ LINK_TAGS = [
     'a',
 ]
 
+IMAGE_TAGS = [
+    'img',
+]
+
 SEPARATE_TAGS = [
     'h1',
     'h2',
@@ -69,6 +73,8 @@ SEPARATE_TAGS = [
 class PMLMLizer(object):
     def __init__(self, log):
         self.log = log
+        self.image_hrefs = {}
+        self.link_hrefs = {}
         
     def extract_content(self, oeb_book, opts):
         self.log.info('Converting XHTML to PML markup...')
@@ -77,6 +83,8 @@ class PMLMLizer(object):
         return self.pmlmlize_spine()
         
     def pmlmlize_spine(self):
+        self.image_hrefs = {}
+        self.link_hrefs = {}
         output = u''
         if 'titlepage' in self.oeb_book.guide:
             self.log.debug('Generating title page...')
@@ -84,19 +92,25 @@ class PMLMLizer(object):
             item = self.oeb_book.manifest.hrefs[href]
             if item.spine_position is None:
                 stylizer = Stylizer(item.data, item.href, self.oeb_book, self.opts.output_profile)
-                output += self.dump_text(item.data.find(XHTML('body')), stylizer)
+                output += self.dump_text(item.data.find(XHTML('body')), stylizer, item)
         for item in self.oeb_book.spine:
             self.log.debug('Converting %s to PML markup...' % item.href)
             stylizer = Stylizer(item.data, item.href, self.oeb_book, self.opts.output_profile)
-            output += self.add_page_anchor(item.href)
+            output += self.add_page_anchor(item)
-            output += self.dump_text(item.data.find(XHTML('body')), stylizer)
+            output += self.dump_text(item.data.find(XHTML('body')), stylizer, item)
         output = self.clean_text(output)
 
         return output
 
-    def add_page_anchor(self, href):
+    def add_page_anchor(self, page):
-        href = os.path.splitext(os.path.basename(href))[0]
+        return self.get_anchor(page, '')
-        return u'\\Q="%s"' % href
+
+    def get_anchor(self, page, aid):
+        aid = '%s#%s' % (page.href, aid)
+        if aid not in self.link_hrefs.keys():
+            self.link_hrefs[aid] = 'calibre_link-%s' % len(self.link_hrefs.keys())
+        aid = self.link_hrefs[aid]
+        return u'\\Q="%s"' % aid
 
     def clean_text(self, text):
         # Remove excess spaces at beginning and end of lines
@@ -123,7 +137,7 @@ class PMLMLizer(object):
         
         return text
 
-    def dump_text(self, elem, stylizer, tag_stack=[]):
+    def dump_text(self, elem, stylizer, page, tag_stack=[]):
         if not isinstance(elem.tag, basestring) \
            or namespace(elem.tag) != XHTML_NS:
             return u''
@@ -146,8 +160,11 @@ class PMLMLizer(object):
         
         # Process tags that need special processing and that do not have inner
         # text. Usually these require an argument
-        if tag == 'img':
+        if tag in IMAGE_TAGS:
-            text += '\\m="%s"' % image_name(os.path.basename(elem.get('src'))).strip('\x00')
+            if elem.attrib.get('src', None):
+                if page.abshref(elem.attrib['src']) not in self.image_hrefs.keys():
+                    self.image_hrefs[page.abshref(elem.attrib['src'])] = image_name('%s' % len(self.image_hrefs.keys()), self.image_hrefs.keys()).strip('\x00')
+                text += '\\m="%s"' % self.image_hrefs[page.abshref(elem.attrib['src'])]
         if tag == 'hr':
             text += '\\w'
             width = elem.get('width')
@@ -171,17 +188,22 @@ class PMLMLizer(object):
         # Anchors links
         if tag in LINK_TAGS and 'q' not in tag_stack:
             href = elem.get('href')
-            if href and '://' not in href:
+            if href:
-                if '#' in href:
+                href = page.abshref(href)
-                    href = href.partition('#')[2]
+                if '://' not in href:
-                href = os.path.splitext(os.path.basename(href))[0]
+                    if '#' not in href:
+                        href += '#'
+                    if href not in self.link_hrefs.keys():
+                        self.link_hrefs[href] = 'calibre_link-%s' % len(self.link_hrefs.keys())
+                    href = self.link_hrefs[href]
+                    text += '\\q="#%s"' % href
                 tag_count += 1
-                text += '\\q="#%s"' % href
                 tag_stack.append('q')
+
         # Anchor ids
         id_name = elem.get('id')
         if id_name:
-            text += '\\Q="%s"' % os.path.splitext(id_name)[0]
+            text += self.get_anchor(page, id_name)
 
         # Processes style information
         for s in STYLES:
@@ -197,7 +219,7 @@ class PMLMLizer(object):
             text += self.elem_text(elem, tag_stack)
             
         for item in elem:
-            text += self.dump_text(item, stylizer, tag_stack)
+            text += self.dump_text(item, stylizer, page, tag_stack)
         
         close_tag_list = []
         for i in range(0, tag_count):

src/calibre/ebooks/rb/rbml.py

@@ -45,6 +45,10 @@ LINK_TAGS = [
     'a',
 ]
 
+IMAGE_TAGS = [
+    'img',
+]
+
 STYLES = [
     ('font-weight', {'bold'   : 'b', 'bolder' : 'b'}),
     ('font-style', {'italic' : 'i'}),
@@ -56,6 +60,7 @@ class RBMLizer(object):
     def __init__(self, log, name_map={}):
         self.log = log
         self.name_map = name_map
+        self.link_hrefs = {}
 
     def extract_content(self, oeb_book, opts):
         self.log.info('Converting XHTML to RB markup...')
@@ -65,6 +70,7 @@ class RBMLizer(object):
 
 
     def mlize_spine(self):
+        self.link_hrefs = {}
         output = u'<HTML><HEAD><TITLE></TITLE></HEAD><BODY>'
         if 'titlepage' in self.oeb_book.guide:
             self.log.debug('Generating cover page...')
@@ -72,19 +78,25 @@ class RBMLizer(object):
             item = self.oeb_book.manifest.hrefs[href]
             if item.spine_position is None:
                 stylizer = Stylizer(item.data, item.href, self.oeb_book, self.opts.output_profile)
-                output += self.dump_text(item.data.find(XHTML('body')), stylizer)
+                output += self.dump_text(item.data.find(XHTML('body')), stylizer, item)
         for item in self.oeb_book.spine:
             self.log.debug('Converting %s to RocketBook HTML...' % item.href)
             stylizer = Stylizer(item.data, item.href, self.oeb_book, self.opts.output_profile)
-            output += self.add_page_anchor(item.href)
+            output += self.add_page_anchor(item)
-            output += self.dump_text(item.data.find(XHTML('body')), stylizer)
+            output += self.dump_text(item.data.find(XHTML('body')), stylizer, item)
         output += u'</BODY></HTML>'
         output = self.clean_text(output)
         return output
 
-    def add_page_anchor(self, href):
+    def add_page_anchor(self, page):
-        href = os.path.splitext(os.path.basename(href))[0]
+        return self.get_anchor(page, '')
-        return u'<A NAME="%s"></A>' % href
+
+    def get_anchor(self, page, aid):
+        aid = '%s#%s' % (page.href, aid)
+        if aid not in self.link_hrefs.keys():
+            self.link_hrefs[aid] = 'calibre_link-%s' % len(self.link_hrefs.keys())
+        aid = self.link_hrefs[aid]
+        return u'<A NAME="%s"></A>' % aid
 
     def clean_text(self, text):        
         # Remove anchors that do not have links
@@ -95,7 +107,7 @@ class RBMLizer(object):
 
         return text
 
-    def dump_text(self, elem, stylizer, tag_stack=[]):
+    def dump_text(self, elem, stylizer, page, tag_stack=[]):
         if not isinstance(elem.tag, basestring) \
            or namespace(elem.tag) != XHTML_NS:
             return u''
@@ -112,10 +124,11 @@ class RBMLizer(object):
         
         # Process tags that need special processing and that do not have inner
         # text. Usually these require an argument
-        if tag == 'img':
+        if tag in IMAGE_TAGS:
-            src = os.path.basename(elem.get('src'))
+            if elem.attrib.get('src', None):
-            name = self.name_map.get(src, src)
+                if page.abshref(elem.attrib['src']) not in self.name_map.keys():
-            text += '<IMG SRC="%s">' % name
+                    self.name_map[page.abshref(elem.attrib['src'])] = unique_name('%s' % len(self.image_hrefs.keys()), self.image_hrefs.keys(), self.name_map.keys())
+                text += '<IMG SRC="%s">' % self.name_map[page.abshref(elem.attrib['src'])]
 
         rb_tag = tag.upper() if tag in TAGS else None
         if rb_tag:
@@ -123,21 +136,25 @@ class RBMLizer(object):
             text += '<%s>' % rb_tag
             tag_stack.append(rb_tag)
 
+        # Anchors links
         if tag in LINK_TAGS:
             href = elem.get('href')
             if href:
+                href = page.abshref(href)
                 if '://' not in href:
-                    if '#' in href:
+                    if '#' not in href:
-                        href = href.partition('#')[2]
+                        href += '#'
-                    href = os.path.splitext(os.path.basename(href))[0]
+                    if href not in self.link_hrefs.keys():
+                        self.link_hrefs[href] = 'calibre_link-%s' % len(self.link_hrefs.keys())
+                    href = self.link_hrefs[href]
+                    text += '<A HREF="#%s">' % href
                 tag_count += 1
-                text += '<A HREF="#%s">' % href
                 tag_stack.append('A')
 
         # Anchor ids
         id_name = elem.get('id')
         if id_name:
-            text += '<A NAME="%s"></A>' % os.path.splitext(id_name)[0]
+            text += self.get_anchor(page, id_name)
 
         # Processes style information
         for s in STYLES:
@@ -153,7 +170,7 @@ class RBMLizer(object):
             text += prepare_string_for_xml(elem.text)
 
         for item in elem:
-            text += self.dump_text(item, stylizer, tag_stack)
+            text += self.dump_text(item, stylizer, page, tag_stack)
 
         close_tag_list = []
         for i in range(0, tag_count):

src/calibre/ebooks/rb/writer.py

@@ -125,10 +125,10 @@ class RBWriter(object):
                     im.save(data, 'PNG')
                     data = data.getvalue()
 
-                    name = '%s.png' % os.path.splitext(os.path.basename(item.href))[0]
+                    name = '%s.png' % len(used_names)
                     name = unique_name(name, used_names)
                     used_names.append(name)
-                    self.name_map[os.path.basename(item.href)] = name
+                    self.name_map[item.href] = name
 
                     images.append((name, data))
                 except Exception as e: