Refactor DOCX metadata reading to use the container class

src/calibre/ebooks/docx/container.py

@@ -6,30 +6,90 @@ from __future__ import (unicode_literals, division, absolute_import,
 __license__ = 'GPL v3'
 __copyright__ = '2013, Kovid Goyal <kovid at kovidgoyal.net>'
 
-import os, sys
+import os, sys, shutil
 
 from lxml import etree
 
 from calibre import walk, guess_type
+from calibre.ebooks.metadata import string_to_authors
+from calibre.ebooks.metadata.book.base import Metadata
 from calibre.ebooks.docx import InvalidDOCX
-from calibre.ebooks.docx.names import DOCUMENT
+from calibre.ebooks.docx.names import DOCUMENT, DOCPROPS, XPath, APPPROPS
 from calibre.ptempfile import PersistentTemporaryDirectory
+from calibre.utils.localization import canonicalize_lang
 from calibre.utils.logging import default_log
 from calibre.utils.zipfile import ZipFile
+from calibre.ebooks.oeb.parse_utils import RECOVER_PARSER
+
+def fromstring(raw, parser=RECOVER_PARSER):
+    return etree.fromstring(raw, parser=parser)
+
+# Read metadata {{{
+def read_doc_props(raw, mi):
+    root = fromstring(raw)
+    titles = XPath('//dc:title')(root)
+    if titles:
+        title = titles[0].text
+        if title and title.strip():
+            mi.title = title.strip()
+    tags = []
+    for subject in XPath('//dc:subject')(root):
+        if subject.text and subject.text.strip():
+            tags.append(subject.text.strip().replace(',', '_'))
+    for keywords in XPath('//cp:keywords')(root):
+        if keywords.text and keywords.text.strip():
+            for x in keywords.text.split():
+                tags.extend(y.strip() for y in x.split(','))
+    if tags:
+        mi.tags = tags
+    authors = XPath('//dc:creator')(root)
+    aut = []
+    for author in authors:
+        if author.text and author.text.strip():
+            aut.extend(string_to_authors(author.text))
+    if aut:
+        mi.authors = aut
+
+    desc = XPath('//dc:description')(root)
+    if desc:
+        raw = etree.tostring(desc[0], method='text', encoding=unicode)
+        mi.comments = raw
+
+    langs = []
+    for lang in XPath('//dc:language')(root):
+        if lang.text and lang.text.strip():
+            l = canonicalize_lang(lang.text)
+            if l:
+                langs.append(l)
+    if langs:
+        mi.languages = langs
+
+def read_app_props(raw, mi):
+    root = fromstring(raw)
+    company = root.xpath('//*[local-name()="Company"]')
+    if company and company[0].text and company[0].text.strip():
+        mi.publisher = company[0].text.strip()
+# }}}
 
 class DOCX(object):
 
-    def __init__(self, path_or_stream, log=None):
+    def __init__(self, path_or_stream, log=None, extract=True):
         stream = path_or_stream if hasattr(path_or_stream, 'read') else open(path_or_stream, 'rb')
         self.name = getattr(stream, 'name', None) or '<stream>'
         self.log = log or default_log
-        self.tdir = PersistentTemporaryDirectory('docx_container')
+        if extract:
-
+            self.extract(stream)
-        self.extract(stream)
+        else:
+            self.init_zipfile(stream)
         self.read_content_types()
         self.read_package_relationships()
 
+    def init_zipfile(self, stream):
+        self.zipf = ZipFile(stream)
+        self.names = frozenset(self.zipf.namelist())
+
     def extract(self, stream):
+        self.tdir = PersistentTemporaryDirectory('docx_container')
         try:
             zf = ZipFile(stream)
             zf.extractall(self.tdir)
@@ -46,6 +106,8 @@ class DOCX(object):
             self.names[name] = f
 
     def read(self, name):
+        if hasattr(self, 'zipf'):
+            return self.zipf.open(name).read()
         path = self.names[name]
         with open(path, 'rb') as f:
             return f.read()
@@ -55,7 +117,7 @@ class DOCX(object):
             raw = self.read('[Content_Types].xml')
         except KeyError:
             raise InvalidDOCX('The file %s docx file has no [Content_Types].xml' % self.name)
-        root = etree.fromstring(raw)
+        root = fromstring(raw)
         self.content_types = {}
         self.default_content_types = {}
         for item in root.xpath('//*[local-name()="Types"]/*[local-name()="Default" and @Extension and @ContentType]'):
@@ -77,7 +139,7 @@ class DOCX(object):
             raw = self.read('_rels/.rels')
         except KeyError:
             raise InvalidDOCX('The file %s docx file has no _rels/.rels' % self.name)
-        root = etree.fromstring(raw)
+        root = fromstring(raw)
         self.relationships = {}
         self.relationships_rmap = {}
         for item in root.xpath('//*[local-name()="Relationships"]/*[local-name()="Relationship" and @Type and @Target]'):
@@ -94,7 +156,48 @@ class DOCX(object):
             if not names:
                 raise InvalidDOCX('The file %s docx file has no main document' % self.name)
             name = names[0]
-        return etree.fromstring(self.read(name))
+        return fromstring(self.read(name))
+
+    @property
+    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:
+            try:
+                raw = self.read(name)
+            except KeyError:
+                pass
+            else:
+                read_doc_props(raw, mi)
+
+        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]
+        if name:
+            try:
+                raw = self.read(name)
+            except KeyError:
+                pass
+            else:
+                read_app_props(raw, mi)
+
+        return mi
+
+    def close(self):
+        if hasattr(self, 'zipf'):
+            self.zipf.close()
+        else:
+            try:
+                shutil.rmtree(self.tdir)
+            except EnvironmentError:
+                pass
 
 if __name__ == '__main__':
-    d = DOCX(sys.argv[-1])
+    d = DOCX(sys.argv[-1], extract=False)
+    print (d.metadata)

src/calibre/ebooks/docx/names.py

@@ -6,7 +6,11 @@ from __future__ import (unicode_literals, division, absolute_import,
 __license__ = 'GPL v3'
 __copyright__ = '2013, Kovid Goyal <kovid at kovidgoyal.net>'
 
+from lxml.etree import XPath as X
+
 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'
 
 namespaces = {
     'mo': 'http://schemas.microsoft.com/office/mac/office/2008/main',
@@ -38,3 +42,6 @@ namespaces = {
     'dcterms': 'http://purl.org/dc/terms/'
 }
 
+def XPath(expr):
+    return X(expr, namespaces=namespaces)
+

src/calibre/ebooks/docx/to_html.py

@@ -0,0 +1,41 @@
+#!/usr/bin/env python
+# vim:fileencoding=utf-8
+from __future__ import (unicode_literals, division, absolute_import,
+                        print_function)
+
+__license__ = 'GPL v3'
+__copyright__ = '2013, Kovid Goyal <kovid at kovidgoyal.net>'
+
+import sys, os
+
+from lxml import html
+from lxml.html.builder import (HTML, HEAD, TITLE, BODY, LINK, META)
+
+from calibre.ebooks.docx.container import Container
+
+class Convert(object):
+
+    def __init__(self, path_or_stream, dest_dir=None, log=None):
+        self.container = Container(path_or_stream, log=log)
+        self.log = self.container.log
+        self.dest_dir = dest_dir or os.getcwdu()
+        self.body = BODY()
+        self.html = HTML(
+            HEAD(
+                META(charset='utf-8'),
+                TITLE('TODO: read from metadata'),
+                LINK(rel='stylesheet', type='text/css', href='docx.css'),
+            ),
+            self.body
+        )
+
+    def __call__(self):
+        self.write()
+
+    def write(self):
+        raw = html.tostring(self.html, encoding='utf-8', doctype='<!DOCTYPE html>')
+        with open(os.path.join(self.dest_dir, 'index.html'), 'wb') as f:
+            f.write(raw)
+
+if __name__ == '__main__':
+    Convert(sys.argv[-1])()

src/calibre/ebooks/metadata/docx.py

@@ -7,80 +7,21 @@ __license__   = 'GPL v3'
 __copyright__ = '2012, Kovid Goyal <kovid at kovidgoyal.net>'
 __docformat__ = 'restructuredtext en'
 
-from lxml import etree
+from calibre.ebooks.docx.container import DOCX
 
-from calibre.ebooks.metadata.book.base import Metadata
-from calibre.utils.localization import canonicalize_lang
 from calibre.utils.zipfile import ZipFile
 from calibre.utils.magick.draw import identify_data
-from calibre.ebooks.oeb.base import DC11_NS
-from calibre.ebooks.oeb.parse_utils import RECOVER_PARSER
-
-NSMAP = {'dc':DC11_NS,
-'cp':'http://schemas.openxmlformats.org/package/2006/metadata/core-properties'}
-
-def XPath(expr):
-    return etree.XPath(expr, namespaces=NSMAP)
-
-def _read_doc_props(raw, mi):
-    from calibre.ebooks.metadata import string_to_authors
-    root = etree.fromstring(raw, parser=RECOVER_PARSER)
-    titles = XPath('//dc:title')(root)
-    if titles:
-        title = titles[0].text
-        if title and title.strip():
-            mi.title = title.strip()
-    tags = []
-    for subject in XPath('//dc:subject')(root):
-        if subject.text and subject.text.strip():
-            tags.append(subject.text.strip().replace(',', '_'))
-    for keywords in XPath('//cp:keywords')(root):
-        if keywords.text and keywords.text.strip():
-            for x in keywords.text.split():
-                tags.extend(y.strip() for y in x.split(','))
-    if tags:
-        mi.tags = tags
-    authors = XPath('//dc:creator')(root)
-    aut = []
-    for author in authors:
-        if author.text and author.text.strip():
-            aut.extend(string_to_authors(author.text))
-    if aut:
-        mi.authors = aut
-
-    desc = XPath('//dc:description')(root)
-    if desc:
-        raw = etree.tostring(desc[0], method='text', encoding=unicode)
-        mi.comments = raw
-
-    langs = []
-    for lang in XPath('//dc:language')(root):
-        if lang.text and lang.text.strip():
-            l = canonicalize_lang(lang.text)
-            if l:
-                langs.append(l)
-    if langs:
-        mi.languages = langs
-
-def _read_app_props(raw, mi):
-    root = etree.fromstring(raw, parser=RECOVER_PARSER)
-    company = root.xpath('//*[local-name()="Company"]')
-    if company and company[0].text and company[0].text.strip():
-        mi.publisher = company[0].text.strip()
 
 def get_metadata(stream):
+    c = DOCX(stream, extract=False)
+    mi = c.metadata
+    c.close()
+    stream.seek(0)
+    cdata = None
     with ZipFile(stream, 'r') as zf:
-
-        mi = Metadata(_('Unknown'))
-        cdata = None
-
         for zi in zf.infolist():
             ext = zi.filename.rpartition('.')[-1].lower()
-            if zi.filename.lower() == 'docprops/core.xml':
+            if cdata is None and ext in {'jpeg', 'jpg', 'png', 'gif'}:
-                _read_doc_props(zf.read(zi), mi)
-            elif zi.filename.lower() == 'docprops/app.xml':
-                _read_app_props(zf.read(zi), mi)
-            elif cdata is None and ext in {'jpeg', 'jpg', 'png', 'gif'}:
                 raw = zf.read(zi)
                 try:
                     width, height, fmt = identify_data(raw)