Implement updating metadata in DOCX files

src/calibre/customize/builtins.py

@@ -533,6 +533,17 @@ class TXTZMetadataWriter(MetadataWriterPlugin):
         from calibre.ebooks.metadata.extz import set_metadata
         set_metadata(stream, mi)
 
+class DocXMetadataWriter(MetadataWriterPlugin):
+
+    name        = 'Set DOCX metadata'
+    file_types  = set(['docx'])
+    description = _('Read metadata from %s files')%'DOCX'
+
+    def set_metadata(self, stream, mi, type):
+        from calibre.ebooks.metadata.docx import set_metadata
+        return set_metadata(stream, mi)
+
+
 plugins += [x for x in list(locals().values()) if isinstance(x, type) and
                                         x.__name__.endswith('MetadataWriter')]
 

src/calibre/ebooks/docx/container.py

@@ -200,17 +200,27 @@ class DOCX(object):
 
         return by_id, by_type
 
-    @property
+    def get_document_properties_names(self):
-    def metadata(self):
-        mi = Metadata(_('Unknown'))
         name = self.relationships.get(DOCPROPS, None)
         if name is None:
             names = tuple(n for n in self.names if n.lower() == 'docprops/core.xml')
             if names:
                 name = names[0]
-        if name:
+        yield name
+        name = self.relationships.get(APPPROPS, None)
+        if name is None:
+            names = tuple(n for n in self.names if n.lower() == 'docprops/app.xml')
+            if names:
+                name = names[0]
+        yield name
+
+    @property
+    def metadata(self):
+        mi = Metadata(_('Unknown'))
+        dp_name, ap_name = self.get_document_properties_names()
+        if dp_name:
             try:
-                raw = self.read(name)
+                raw = self.read(dp_name)
             except KeyError:
                 pass
             else:
@@ -223,14 +233,10 @@ class DOCX(object):
             else:
                 read_default_style_language(raw, mi)
 
-        name = self.relationships.get(APPPROPS, None)
+        ap_name = self.relationships.get(APPPROPS, None)
-        if name is None:
+        if ap_name:
-            names = tuple(n for n in self.names if n.lower() == 'docprops/app.xml')
-            if names:
-                name = names[0]
-        if name:
             try:
-                raw = self.read(name)
+                raw = self.read(ap_name)
             except KeyError:
                 pass
             else:

src/calibre/ebooks/docx/writer/container.py

@@ -30,7 +30,7 @@ def xml2str(root, pretty_print=False, with_tail=False):
 def update_doc_props(root, mi):
     def setm(name, text=None, ns='dc'):
         ans = root.makeelement('{%s}%s' % (namespaces[ns], name))
-        for child in root:
+        for child in tuple(root):
             if child.tag == ans.tag:
                 root.remove(child)
         ans.text = text

src/calibre/ebooks/metadata/docx.py

@@ -7,9 +7,13 @@ __license__   = 'GPL v3'
 __copyright__ = '2012, Kovid Goyal <kovid at kovidgoyal.net>'
 __docformat__ = 'restructuredtext en'
 
-from calibre.ebooks.docx.container import DOCX
+from io import BytesIO
-from calibre.ebooks.docx.names import XPath, get
 
+from lxml import etree
+
+from calibre.ebooks.docx.container import DOCX
+from calibre.ebooks.docx.writer.container import update_doc_props, xml2str, namespaces
+from calibre.ebooks.docx.names import XPath, get
 from calibre.utils.magick.draw import identify_data
 
 images = XPath('//*[name()="w:drawing" or name()="w:pict"]/descendant::*[(name()="a:blip" and @r:embed) or (name()="v:imagedata" and @r:id)][1]')
@@ -44,6 +48,30 @@ def get_metadata(stream):
 
     return mi
 
+def set_metadata(stream, mi):
+    from calibre.utils.zipfile import safe_replace
+    c = DOCX(stream, extract=False)
+    dp_name, ap_name = c.get_document_properties_names()
+    dp_raw = c.read(dp_name)
+    try:
+        ap_raw = c.read(ap_name)
+    except Exception:
+        ap_raw = None
+    cp = etree.fromstring(dp_raw)
+    update_doc_props(cp, mi)
+    replacements = {}
+    if ap_raw is not None:
+        ap = etree.fromstring(ap_raw)
+        comp = ap.makeelement('{%s}Company' % namespaces['ep'])
+        for child in tuple(ap):
+            if child.tag == comp.tag:
+                ap.remove(child)
+        comp.text = mi.publisher
+        ap.append(comp)
+        replacements[ap_name] = BytesIO(xml2str(ap))
+    stream.seek(0)
+    safe_replace(stream, dp_name, BytesIO(xml2str(cp)), extra_replacements=replacements)
+
 if __name__ == '__main__':
     import sys
     with open(sys.argv[-1], 'rb') as stream: