Update old transcripts metadata present in video xml with the new transcripts for backward compatibility.

common/lib/xmodule/xmodule/tests/test_video.py

@@ -716,7 +716,12 @@ class VideoExportTestCase(VideoDescriptorTestBase):
         Test that we write the correct XML on export.
         """
         edx_video_id = u'test_edx_video_id'
-        mock_val_api.export_to_xml = Mock(return_value=etree.Element('video_asset'))
+        mock_val_api.export_to_xml = Mock(
+            return_value=dict(
+                xml=etree.Element('video_asset'),
+                transcripts={}
+            )
+        )
         self.descriptor.youtube_id_0_75 = 'izygArpw-Qo'
         self.descriptor.youtube_id_1_0 = 'p2Q6BrNhdh8'
         self.descriptor.youtube_id_1_25 = '1EeWXzPdhSA'
@@ -736,14 +741,23 @@ class VideoExportTestCase(VideoDescriptorTestBase):
         xml = self.descriptor.definition_to_xml(self.file_system)
         parser = etree.XMLParser(remove_blank_text=True)
         xml_string = '''\
-         <video url_name="SampleProblem" start_time="0:00:01" youtube="0.75:izygArpw-Qo,1.00:p2Q6BrNhdh8,1.25:1EeWXzPdhSA,1.50:rABDYkeK0x8" show_captions="false" end_time="0:01:00" download_video="true" download_track="true">
+         <video
+            url_name="SampleProblem"
+            start_time="0:00:01"
+            show_captions="false"
+            end_time="0:01:00"
+            download_video="true"
+            download_track="true"
+            youtube="0.75:izygArpw-Qo,1.00:p2Q6BrNhdh8,1.25:1EeWXzPdhSA,1.50:rABDYkeK0x8"
+            transcripts='{"ge": "german_translation.srt", "ua": "ukrainian_translation.srt"}'
+         >
            <source src="http://www.example.com/source.mp4"/>
            <source src="http://www.example.com/source1.ogg"/>
            <track src="http://www.example.com/track"/>
            <handout src="http://www.example.com/handout"/>
+           <video_asset />
            <transcript language="ge" src="german_translation.srt" />
            <transcript language="ua" src="ukrainian_translation.srt" />
-           <video_asset />
          </video>
         '''
         expected = etree.XML(xml_string, parser=parser)

common/lib/xmodule/xmodule/video_module/video_module.py

@@ -739,13 +739,9 @@ class VideoDescriptor(VideoFields, VideoTranscriptsMixin, VideoStudioViewHandler
             ele.set('src', self.handout)
             xml.append(ele)
 
+        transcripts = {}
         if self.transcripts is not None:
-            # sorting for easy testing of resulting xml
-            for transcript_language in sorted(self.transcripts.keys()):
-                ele = etree.Element('transcript')
-                ele.set('language', transcript_language)
-                ele.set('src', self.transcripts[transcript_language])
-                xml.append(ele)
+            transcripts.update(self.transcripts)
 
         edx_video_id = clean_video_id(self.edx_video_id)
         if edxval_api and edx_video_id:
@@ -753,17 +749,42 @@ class VideoDescriptor(VideoFields, VideoTranscriptsMixin, VideoStudioViewHandler
                 # Create static dir if not created earlier.
                 resource_fs.makedirs(EXPORT_IMPORT_STATIC_DIR, recreate=True)
 
-                xml.append(
-                    edxval_api.export_to_xml(
-                        video_id=edx_video_id,
-                        resource_fs=resource_fs,
-                        static_dir=EXPORT_IMPORT_STATIC_DIR,
-                        course_id=unicode(self.runtime.course_id.for_branch(None))
-                    )
+                # Backward compatible exports
+                # edxval exports new transcripts into the course OLX and returns a transcript
+                # files map so that it can also be rewritten in old transcript metadata fields
+                # (i.e. `self.transcripts`) on import and older open-releases (<= ginkgo),
+                # who do not have deprecated contentstore yet, can also import and use new-style
+                # transcripts into their openedX instances.
+
+                exported_metadata = edxval_api.export_to_xml(
+                    video_id=edx_video_id,
+                    resource_fs=resource_fs,
+                    static_dir=EXPORT_IMPORT_STATIC_DIR,
+                    course_id=unicode(self.runtime.course_id.for_branch(None))
                 )
+                # Update xml with edxval metadata
+                xml.append(exported_metadata['xml'])
+
+                # we don't need sub if english transcript
+                # is also in new transcripts.
+                new_transcripts = exported_metadata['transcripts']
+                transcripts.update(new_transcripts)
+                if new_transcripts.get('en'):
+                    xml.set('sub', '')
+
+                # Update `transcripts` attribute in the xml
+                xml.set('transcripts', json.dumps(transcripts))
+
             except edxval_api.ValVideoNotFoundError:
                 pass
 
+            # Sorting transcripts for easy testing of resulting xml
+            for transcript_language in sorted(transcripts.keys()):
+                ele = etree.Element('transcript')
+                ele.set('language', transcript_language)
+                ele.set('src', transcripts[transcript_language])
+                xml.append(ele)
+
         # handle license specifically
         self.add_license_to_xml(xml)
 

common/lib/xmodule/xmodule/xml_module.py

@@ -139,6 +139,14 @@ class XmlParserMixin(object):
                          # Used for storing xml attributes between import and export, for roundtrips
                          'xml_attributes')
 
+    # This is a categories to fields map that contains the block category specific fields which should not be
+    # cleaned and/or override while adding xml to node.
+    metadata_to_not_to_clean = {
+        # A category `video` having `sub` and `transcripts` fields
+        # which should not be cleaned/override in an xml object.
+        'video': ('sub', 'transcripts')
+    }
+
     metadata_to_export_to_policy = ('discussion_topics',)
 
     @staticmethod
@@ -165,13 +173,15 @@ class XmlParserMixin(object):
         raise NotImplementedError("%s does not implement definition_from_xml" % cls.__name__)
 
     @classmethod
-    def clean_metadata_from_xml(cls, xml_object):
+    def clean_metadata_from_xml(cls, xml_object, excluded_fields=()):
         """
         Remove any attribute named for a field with scope Scope.settings from the supplied
         xml_object
         """
         for field_name, field in cls.fields.items():
-            if field.scope == Scope.settings and xml_object.get(field_name) is not None:
+            if (field.scope == Scope.settings
+                    and field_name not in excluded_fields
+                    and xml_object.get(field_name) is not None):
                 del xml_object.attrib[field_name]
 
     @classmethod
@@ -448,7 +458,8 @@ class XmlParserMixin(object):
                 aside.add_xml_to_node(aside_node)
                 xml_object.append(aside_node)
 
-        self.clean_metadata_from_xml(xml_object)
+        not_to_clean_fields = self.metadata_to_not_to_clean.get(self.category, ())
+        self.clean_metadata_from_xml(xml_object, excluded_fields=not_to_clean_fields)
 
         # Set the tag on both nodes so we get the file path right.
         xml_object.tag = self.category
@@ -457,7 +468,9 @@ class XmlParserMixin(object):
         # Add the non-inherited metadata
         for attr in sorted(own_metadata(self)):
             # don't want e.g. data_dir
-            if attr not in self.metadata_to_strip and attr not in self.metadata_to_export_to_policy:
+            if (attr not in self.metadata_to_strip
+                    and attr not in self.metadata_to_export_to_policy
+                    and attr not in not_to_clean_fields):
                 val = serialize_field(self._field_data.get(self, attr))
                 try:
                     xml_object.set(attr, val)