DOCX: Bookmarks

src/calibre/ebooks/docx/names.py

@@ -6,8 +6,12 @@ from __future__ import (unicode_literals, division, absolute_import,
 __license__ = 'GPL v3'
 __copyright__ = '2013, Kovid Goyal <kovid at kovidgoyal.net>'
 
+import re
+
 from lxml.etree import XPath as X
 
+from calibre.utils.filenames import ascii_text
+
 DOCUMENT  = 'http://schemas.openxmlformats.org/officeDocument/2006/relationships/officeDocument'
 DOCPROPS  = 'http://schemas.openxmlformats.org/package/2006/relationships/metadata/core-properties'
 APPPROPS  = 'http://schemas.openxmlformats.org/officeDocument/2006/relationships/extended-properties'
@@ -66,7 +70,27 @@ def barename(x):
 def XML(x):
     return '{%s}%s' % (namespaces['xml'], x)
 
-def get(x, attr, default=None):
+def expand(name):
-    ns, name = attr.partition(':')[0::2]
+    ns, tag = name.partition(':')[0::2]
-    return x.attrib.get('{%s}%s' % (namespaces[ns], name), default)
+    if ns:
+        tag = '{%s}%s' % (namespaces[ns], tag)
+    return tag
+
+def get(x, attr, default=None):
+    return x.attrib.get(expand(attr), default)
+
+def ancestor(elem, name):
+    tag = expand(name)
+    while elem is not None:
+        elem = elem.getparent()
+        if getattr(elem, 'tag', None) == tag:
+            return elem
+
+def generate_anchor(name, existing):
+    x = y = 'id_' + re.sub(r'[^0-9a-zA-Z_]', '', ascii_text(name)).lstrip('_')
+    c = 1
+    while y in existing:
+        y = '%s_%d' % (x, c)
+        c += 1
+    return y
 

src/calibre/ebooks/docx/to_html.py

@@ -14,7 +14,7 @@ from lxml.html.builder import (
     HTML, HEAD, TITLE, BODY, LINK, META, P, SPAN, BR, DIV)
 
 from calibre.ebooks.docx.container import DOCX, fromstring
-from calibre.ebooks.docx.names import XPath, is_tag, XML, STYLES, NUMBERING, FONTS
+from calibre.ebooks.docx.names import XPath, is_tag, XML, STYLES, NUMBERING, FONTS, expand, get, generate_anchor
 from calibre.ebooks.docx.styles import Styles, inherit, PageProperties
 from calibre.ebooks.docx.numbering import Numbering
 from calibre.ebooks.docx.fonts import Fonts
@@ -70,6 +70,7 @@ class Convert(object):
         self.layers = OrderedDict()
         self.framed = [[]]
         self.framed_map = {}
+        self.anchor_map = {}
 
         self.read_page_properties(doc)
         for wp, page_properties in self.page_map.iteritems():
@@ -141,7 +142,7 @@ class Convert(object):
             if name is None:
                 cname = self.docx.document_name.split('/')
                 cname[-1] = defname
-                if self.docx.exists(cname):
+                if self.docx.exists('/'.join(cname)):
                     name = name
             return name
 
@@ -193,10 +194,21 @@ class Convert(object):
         style = self.styles.resolve_paragraph(p)
         self.layers[p] = []
         self.add_frame(dest, style.frame)
-        for run in XPath('descendant::w:r')(p):
+
-            span = self.convert_run(run)
+        current_anchor = None
-            dest.append(span)
+
-            self.layers[p].append(run)
+        for x in p.iterdescendants(expand('w:r'), expand('w:bookmarkStart')):
+            if x.tag.endswith('}r'):
+                span = self.convert_run(x)
+                if current_anchor is not None:
+                    (dest if len(dest) == 0 else span).set('id', current_anchor)
+                    current_anchor = None
+                dest.append(span)
+                self.layers[p].append(x)
+            elif x.tag.endswith('}bookmarkStart'):
+                anchor = get(x, 'w:name')
+                if anchor and anchor not in self.anchor_map:
+                    self.anchor_map[anchor] = current_anchor = generate_anchor(anchor, frozenset(self.anchor_map.itervalues()))
 
         m = re.match(r'heading\s+(\d+)$', style.style_name or '', re.IGNORECASE)
         if m is not None: