AZW3: Preserve the primary-writing-mopde EXTH header field when round…

…-tripping AZW3 files. Should fix editing/conversion of RTL AZW3 files causing page turning to become left-to-right on the Kindle

src/calibre/ebooks/metadata/opf2.py

@@ -1163,6 +1163,11 @@ def page_progression_direction(self):
                 if k == 'page-progression-direction' or k.endswith('}page-progression-direction'):
                     return v
 
+    @property
+    def primary_writing_mode(self):
+        for m in self.XPath('//*[local-name()="meta" and @name="primary-writing-mode" and @content]')(self.root):
+            return m.get('content')
+
     def guess_cover(self):
         '''
         Try to guess a cover. Needed for some old/badly formed OPF files.
@@ -1384,6 +1389,7 @@ def __init__(self, base_path, other):
         Metadata.__init__(self, title='', other=other)
         self.base_path = os.path.abspath(base_path)
         self.page_progression_direction = None
+        self.primary_writing_mode = None
         if self.application_id is None:
             self.application_id = str(uuid.uuid4())
         if not isinstance(self.toc, TOC):
@@ -1546,6 +1552,8 @@ def CAL_ELEM(name, content):
             from calibre.ebooks.metadata.book.json_codec import object_to_unicode
             a(CAL_ELEM('calibre:user_categories',
                        json.dumps(object_to_unicode(self.user_categories))))
+        if self.primary_writing_mode:
+            a(M.meta(name='primary-writing-mode', content=self.primary_writing_mode))
         manifest = E.manifest()
         if self.manifest is not None:
             for ref in self.manifest:

src/calibre/ebooks/mobi/reader/headers.py

@@ -34,6 +34,7 @@ def __init__(self, raw, codec, title):
         self.kf8_header = None
         self.uuid = self.cdetype = None
         self.page_progression_direction = None
+        self.primary_writing_mode = None
 
         self.decode = lambda x : clean_ascii_chars(x.decode(codec, 'replace'))
 
@@ -83,6 +84,13 @@ def __init__(self, raw, codec, title):
                         self.mi.language = lang
                 except:
                     pass
+            elif idx == 525:
+                try:
+                    pwm = content.decode(codec)
+                    if pwm:
+                        self.primary_writing_mode = pwm
+                except Exception:
+                    pass
             elif idx == 527:
                 try:
                     ppd = content.decode(codec)
@@ -336,4 +344,3 @@ def section_data(self, number):
         except OverflowError:
             self.stream.seek(start)
             return self.stream.read()
-

src/calibre/ebooks/mobi/reader/mobi8.py

@@ -512,6 +512,9 @@ def exclude(path):
         ppd = getattr(self.header.exth, 'page_progression_direction', None)
         if ppd in {'ltr', 'rtl', 'default'}:
             opf.page_progression_direction = ppd
+        pwm = getattr(self.header.exth, 'primary_writing_mode', None)
+        if pwm is not None:
+            opf.primary_writing_mode = pwm
 
         with open('metadata.opf', 'wb') as of, open('toc.ncx', 'wb') as ncx:
             opf.render(of, ncx, 'toc.ncx')

src/calibre/ebooks/mobi/writer8/exth.py

@@ -40,6 +40,7 @@
     'lastupdatetime': 502,
     'title': 503,
     'language': 524,
+    'primary_writing_mode': 525,
     'page_progression_direction': 527,
 }
 
@@ -50,7 +51,7 @@ def build_exth(metadata, prefer_author_sort=False, is_periodical=False,
         share_not_sync=True, cover_offset=None, thumbnail_offset=None,
         start_offset=None, mobi_doctype=2, num_of_resources=None,
         kf8_unknown_count=0, be_kindlegen2=False, kf8_header_index=None,
-        page_progression_direction=None):
+        page_progression_direction=None, primary_writing_mode=None):
     exth = BytesIO()
     nrecs = 0
 
@@ -208,6 +209,12 @@ def build_exth(metadata, prefer_author_sort=False, is_periodical=False,
             kf8_unknown_count))
         nrecs += 1
 
+    if primary_writing_mode:
+        pwm = primary_writing_mode.encode('utf-8')
+        exth.write(pack(b'>II', EXTH_CODES['primary_writing_mode'], len(pwm) + 8))
+        exth.write(pwm)
+        nrecs += 1
+
     if page_progression_direction in {'rtl', 'ltr', 'default'}:
         ppd = bytes(page_progression_direction)
         exth.write(pack(b'>II', EXTH_CODES['page_progression_direction'], len(ppd) + 8))
@@ -219,5 +226,3 @@ def build_exth(metadata, prefer_author_sort=False, is_periodical=False,
     pad = b'\0' * (4 - trail)  # Always pad w/ at least 1 byte
     exth = [b'EXTH', pack(b'>II', len(exth) + 12, nrecs), exth, pad]
     return b''.join(exth)
-
-

src/calibre/ebooks/mobi/writer8/mobi.py

@@ -209,6 +209,7 @@ def format_value(self, name, val):
 
 # }}}
 
+
 HEADER_FIELDS = {'compression', 'text_length', 'last_text_record', 'book_type',
                     'first_non_text_record', 'title_length', 'language_code',
                     'first_resource_record', 'exth_flags', 'fdst_record',
@@ -223,6 +224,7 @@ def __init__(self, writer, for_joint=False):
         self.build_records(writer, for_joint)
         self.used_images = writer.used_images
         self.page_progression_direction = writer.oeb.spine.page_progression_direction
+        self.primary_writing_mode = writer.oeb.metadata.primary_writing_mode
 
     def build_records(self, writer, for_joint):
         metadata = writer.oeb.metadata
@@ -311,7 +313,8 @@ def record0(self):
             num_of_resources=self.num_of_resources,
             kf8_unknown_count=self.kuc, be_kindlegen2=True,
             start_offset=self.start_offset, mobi_doctype=self.book_type,
-            page_progression_direction=self.page_progression_direction
+            page_progression_direction=self.page_progression_direction,
+            primary_writing_mode=self.primary_writing_mode
         )
 
         kwargs = {field:getattr(self, field) for field in HEADER_FIELDS}
@@ -342,4 +345,3 @@ def write(self, outpath):
 
             for rec in records:
                 f.write(rec)
-

src/calibre/ebooks/oeb/base.py

@@ -782,6 +782,7 @@ def to_opf2(self, parent=None, nsrmap={}):
     def __init__(self, oeb):
         self.oeb = oeb
         self.items = defaultdict(list)
+        self.primary_writing_mode = None
 
     def add(self, term, value, attrib={}, nsmap={}, **kwargs):
         """Add a new metadata item."""
@@ -864,6 +865,8 @@ def to_opf2(self, parent=None):
         for term in self.items:
             for item in self.items[term]:
                 item.to_opf2(elem, nsrmap=nsrmap)
+        if self.primary_writing_mode:
+            elem.append(elem.makeelement(OPF('meta'), attrib={'name':'primary-writing-mode', 'content':self.primary_writing_mode}))
         return elem
 
 

src/calibre/ebooks/oeb/reader.py

@@ -137,7 +137,11 @@ def _metadata_from_opf(self, opf):
         from calibre.ebooks.metadata.opf2 import OPF
         from calibre.ebooks.oeb.transforms.metadata import meta_info_to_oeb_metadata
         stream = cStringIO.StringIO(etree.tostring(opf, xml_declaration=True, encoding='utf-8'))
-        mi = OPF(stream).to_book_metadata()
+        o = OPF(stream)
+        pwm = o.primary_writing_mode
+        if pwm:
+            self.oeb.metadata.primary_writing_mode = pwm
+        mi = o.to_book_metadata()
         if not mi.language:
             mi.language = get_lang().replace('_', '-')
         self.oeb.metadata.add('language', mi.language)