EPUB Output: Fix converting a document with obfuscated fonts to EPUB 3 not working. Fixes #1990185 [Private bug](https://bugs.launchpad.net/calibre/+bug/1990185)

src/calibre/ebooks/conversion/plugins/epub_output.py

@@ -257,11 +257,11 @@ class EPUBOutput(OutputFormatPlugin):
             opf = [x for x in os.listdir(tdir) if x.endswith('.opf')][0]
             self.condense_ncx([os.path.join(tdir, x) for x in os.listdir(tdir)
                     if x.endswith('.ncx')][0])
-            if self.opts.epub_version == '3':
-                self.upgrade_to_epub3(tdir, opf)
             encryption = None
             if encrypted_fonts:
                 encryption = self.encrypt_fonts(encrypted_fonts, tdir, uuid)
+            if self.opts.epub_version == '3':
+                encryption = self.upgrade_to_epub3(tdir, opf, encryption)
 
             from calibre.ebooks.epub import initialize_container
             with initialize_container(output_path, os.path.basename(opf),
@@ -284,7 +284,7 @@ class EPUBOutput(OutputFormatPlugin):
                     zf.extractall(path=opts.extract_to)
                 self.log.info('EPUB extracted to', opts.extract_to)
 
-    def upgrade_to_epub3(self, tdir, opf):
+    def upgrade_to_epub3(self, tdir, opf, encryption=None):
         self.log.info('Upgrading to EPUB 3...')
         from calibre.ebooks.epub import simple_container_xml
         from calibre.ebooks.oeb.polish.cover import fix_conversion_titlepage_links_in_nav
@@ -294,6 +294,9 @@ class EPUBOutput(OutputFormatPlugin):
             pass
         with open(os.path.join(tdir, 'META-INF', 'container.xml'), 'wb') as f:
             f.write(simple_container_xml(os.path.basename(opf)).encode('utf-8'))
+        if encryption is not None:
+            with open(os.path.join(tdir, 'META-INF', 'encryption.xml'), 'wb') as ef:
+                ef.write(as_bytes(encryption))
         from calibre.ebooks.oeb.polish.container import EpubContainer
         container = EpubContainer(tdir, self.log)
         from calibre.ebooks.oeb.polish.upgrade import epub_2_to_3
@@ -304,10 +307,14 @@ class EPUBOutput(OutputFormatPlugin):
         fix_conversion_titlepage_links_in_nav(container)
         container.commit()
         os.remove(f.name)
+        if encryption is not None:
+            encryption = open(ef.name, 'rb').read()
+            os.remove(ef.name)
         try:
             os.rmdir(os.path.join(tdir, 'META-INF'))
         except OSError:
             pass
+        return encryption
 
     def encrypt_fonts(self, uris, tdir, uuid):  # {{{
         from polyglot.binary import from_hex_bytes

src/calibre/ebooks/oeb/polish/container.py

@@ -1283,24 +1283,7 @@ class EpubContainer(Container):
                     self.dirty('META-INF/encryption.xml')
         super().remove_item(name, remove_from_guide=remove_from_guide)
 
-    def process_encryption(self):
+    def read_raw_unique_identifier(self):
-        fonts = {}
-        enc = self.parsed('META-INF/encryption.xml')
-        for em in enc.xpath('//*[local-name()="EncryptionMethod" and @Algorithm]'):
-            alg = em.get('Algorithm')
-            if alg not in {ADOBE_OBFUSCATION, IDPF_OBFUSCATION}:
-                raise DRMError()
-            try:
-                cr = em.getparent().xpath('descendant::*[local-name()="CipherReference" and @URI]')[0]
-            except (IndexError, ValueError, KeyError):
-                continue
-            name = self.href_to_name(cr.get('URI'))
-            path = self.name_path_map.get(name, None)
-            if path is not None:
-                fonts[name] = alg
-        if not fonts:
-            return
-
         package_id = raw_unique_identifier = idpf_key = None
         for attrib, val in iteritems(self.opf.attrib):
             if attrib.endswith('unique-identifier'):
@@ -1315,6 +1298,32 @@ class EpubContainer(Container):
             idpf_key = raw_unique_identifier
             idpf_key = re.sub('[\u0020\u0009\u000d\u000a]', '', idpf_key)
             idpf_key = hashlib.sha1(idpf_key.encode('utf-8')).digest()
+        return package_id, raw_unique_identifier, idpf_key
+
+    def iter_encryption_entries(self):
+        if 'META-INF/encryption.xml' in self.name_path_map:
+            enc = self.parsed('META-INF/encryption.xml')
+            for em in enc.xpath('//*[local-name()="EncryptionMethod" and @Algorithm]'):
+                try:
+                    cr = em.getparent().xpath('descendant::*[local-name()="CipherReference" and @URI]')[0]
+                except Exception:
+                    cr = None
+                yield em, cr
+
+    def process_encryption(self):
+        fonts = {}
+        for em, cr in self.iter_encryption_entries():
+            alg = em.get('Algorithm')
+            if alg not in {ADOBE_OBFUSCATION, IDPF_OBFUSCATION}:
+                raise DRMError()
+            if cr is None:
+                continue
+            name = self.href_to_name(cr.get('URI'))
+            path = self.name_path_map.get(name, None)
+            if path is not None:
+                fonts[name] = alg
+
+        package_id, raw_unique_identifier, idpf_key = self.read_raw_unique_identifier()
         key = None
         for item in self.opf_xpath('//*[local-name()="metadata"]/*'
                                    '[local-name()="identifier"]'):

src/calibre/ebooks/oeb/polish/upgrade.py

@@ -4,14 +4,19 @@
 
 import sys
 
+from calibre.ebooks.conversion.plugins.epub_input import (
+    ADOBE_OBFUSCATION, IDPF_OBFUSCATION
+)
+from calibre.ebooks.metadata.opf3 import XPath
 from calibre.ebooks.metadata.opf_2_to_3 import upgrade_metadata
-from calibre.ebooks.oeb.base import EPUB_NS, OEB_DOCS, xpath
+from calibre.ebooks.oeb.base import DC, EPUB_NS, OEB_DOCS, xpath
 from calibre.ebooks.oeb.parse_utils import ensure_namespace_prefixes
-from calibre.ebooks.oeb.polish.utils import OEB_FONTS
 from calibre.ebooks.oeb.polish.opf import get_book_language
 from calibre.ebooks.oeb.polish.toc import (
     commit_nav_toc, find_existing_ncx_toc, get_landmarks, get_toc
 )
+from calibre.ebooks.oeb.polish.utils import OEB_FONTS
+from calibre.utils.short_uuid import uuid4
 
 
 def add_properties(item, *props):
@@ -31,6 +36,38 @@ def fix_font_mime_types(container):
     return changed
 
 
+def migrate_obfuscated_fonts(container):
+    if not container.obfuscated_fonts:
+        return
+    name_to_elem_map = {}
+    for em, cr in container.iter_encryption_entries():
+        alg = em.get('Algorithm')
+        if cr is None or alg not in {ADOBE_OBFUSCATION, IDPF_OBFUSCATION}:
+            continue
+        name = container.href_to_name(cr.get('URI'))
+        name_to_elem_map[name] = em, cr
+    package_id, raw_unique_identifier, idpf_key = container.read_raw_unique_identifier()
+    if not idpf_key:
+        if not package_id:
+            package_id = uuid4()
+            container.opf.set('unique-identifier', package_id)
+        metadata = XPath('./opf:metadata')(container.opf)[0]
+        ident = metadata.makeelement(DC('identifier'))
+        ident.text = uuid4()
+        metadata.append(ident)
+        package_id, raw_unique_identifier, idpf_key = container.read_raw_unique_identifier()
+    for name in tuple(container.obfuscated_fonts):
+        try:
+            em, cr = name_to_elem_map[name]
+        except KeyError:
+            container.obfuscated_fonts.pop(name)
+            continue
+        em.set('Algorithm', IDPF_OBFUSCATION)
+        cr.set('URI', container.name_to_href(name))
+        container.obfuscated_fonts[name] = (IDPF_OBFUSCATION, idpf_key)
+    container.commit_item('META-INF/encryption.xml')
+
+
 def collect_properties(container):
     for item in container.opf_xpath('//opf:manifest/opf:item[@href and @media-type]'):
         mt = item.get('media-type') or ''
@@ -128,6 +165,7 @@ def epub_2_to_3(container, report, previous_nav=None, remove_ncx=True):
     container.opf.set('version', '3.0')
     if fix_font_mime_types(container):
         container.refresh_mime_map()
+    migrate_obfuscated_fonts(container)
     container.dirty(container.opf_name)