Add support for extracting subtitles from MPD manifests

test/test_InfoExtractor.py

@@ -745,7 +745,7 @@ jwplayer("mediaplayer").setup({"abouttext":"Visit Indie DB","aboutlink":"http:\/
                 self.ie._sort_formats(formats)
                 expect_value(self, formats, expected_formats, None)
 
-    def test_parse_mpd_formats(self):
+    def test_parse_mpd_formats_subtitles(self):
         _TEST_CASES = [
             (
                 # https://github.com/ytdl-org/youtube-dl/issues/13919
@@ -829,7 +829,8 @@ jwplayer("mediaplayer").setup({"abouttext":"Visit Indie DB","aboutlink":"http:\/
                     'tbr': 5997.485,
                     'width': 1920,
                     'height': 1080,
-                }]
+                }],
+                {},
             ), (
                 # https://github.com/ytdl-org/youtube-dl/pull/14844
                 'urls_only',
@@ -912,7 +913,8 @@ jwplayer("mediaplayer").setup({"abouttext":"Visit Indie DB","aboutlink":"http:\/
                     'tbr': 4400,
                     'width': 1920,
                     'height': 1080,
-                }]
+                }],
+                {},
             ), (
                 # https://github.com/ytdl-org/youtube-dl/issues/20346
                 # Media considered unfragmented even though it contains
@@ -958,18 +960,116 @@ jwplayer("mediaplayer").setup({"abouttext":"Visit Indie DB","aboutlink":"http:\/
                     'width': 360,
                     'height': 360,
                     'fps': 30,
-                }]
+                }],
+                {},
+            ), (
+                'subtitles',
+                'https://example.com/streams/1/playlist/playlist.mpd',  # mpd_url
+                'https://example.com/streams/1/playlist',  # mpd_base_url
+                [{'acodec': 'mp4a.40.2',
+                  'asr': 48000,
+                  'container': 'm4a_dash',
+                  'ext': 'm4a',
+                  'filesize': None,
+                  'format_id': '131kbps',
+                  'format_note': 'DASH audio',
+                  'fps': None,
+                  'fragment_base_url': 'https://example.com/streams/1/playlist/',
+                  'fragments': [{'path': '../audio/1_stereo_131072/dash/init.mp4'},
+                                {'duration': 3989.0,
+                                 'path': '../audio/1_stereo_131072/dash/segment_0.m4s'},
+                                {'duration': 3989.0,
+                                 'path': '../audio/1_stereo_131072/dash/segment_1.m4s'}],
+                  'height': None,
+                  'language': 'de',
+                  'manifest_url': 'https://example.com/streams/1/playlist/playlist.mpd',
+                  'protocol': 'http_dash_segments',
+                  'tbr': 131.072,
+                  'url': 'https://example.com/streams/1/playlist/playlist.mpd',
+                  'vcodec': 'none',
+                  'width': None},
+                 {'acodec': 'mp4a.40.2',
+                  'asr': 48000,
+                  'container': 'm4a_dash',
+                  'ext': 'm4a',
+                  'filesize': None,
+                  'format_id': '196kbps',
+                  'format_note': 'DASH audio',
+                  'fps': None,
+                  'fragment_base_url': 'https://example.com/streams/1/playlist/',
+                  'fragments': [{'path': '../audio/1_stereo_196608/dash/init.mp4'},
+                                {'duration': 3989.0,
+                                 'path': '../audio/1_stereo_196608/dash/segment_0.m4s'},
+                                {'duration': 3989.0,
+                                 'path': '../audio/1_stereo_196608/dash/segment_1.m4s'}],
+                  'height': None,
+                  'language': 'de',
+                  'manifest_url': 'https://example.com/streams/1/playlist/playlist.mpd',
+                  'protocol': 'http_dash_segments',
+                  'tbr': 196.608,
+                  'url': 'https://example.com/streams/1/playlist/playlist.mpd',
+                  'vcodec': 'none',
+                  'width': None},
+                 {'acodec': 'none',
+                  'asr': None,
+                  'container': 'mp4_dash',
+                  'ext': 'mp4',
+                  'filesize': None,
+                  'format_id': '720p 1712kbps',
+                  'format_note': 'DASH video',
+                  'fps': 25,
+                  'fragment_base_url': 'https://example.com/streams/1/playlist/',
+                  'fragments': [{'path': '../video/720_1712128/dash/init.mp4'},
+                                {'duration': 4000.0,
+                                 'path': '../video/720_1712128/dash/segment_0.m4s'},
+                                {'duration': 4000.0,
+                                 'path': '../video/720_1712128/dash/segment_1.m4s'}],
+                  'height': 720,
+                  'language': None,
+                  'manifest_url': 'https://example.com/streams/1/playlist/playlist.mpd',
+                  'protocol': 'http_dash_segments',
+                  'tbr': 1712.128,
+                  'url': 'https://example.com/streams/1/playlist/playlist.mpd',
+                  'vcodec': 'avc1.42c00d',
+                  'width': 1280},
+                 {'acodec': 'none',
+                  'asr': None,
+                  'container': 'mp4_dash',
+                  'ext': 'mp4',
+                  'filesize': None,
+                  'format_id': '1080p 4669kbps',
+                  'format_note': 'DASH video',
+                  'fps': 25,
+                  'fragment_base_url': 'https://example.com/streams/1/playlist/',
+                  'fragments': [{'path': '../video/1080_4669440/dash/init.mp4'},
+                                {'duration': 4000.0,
+                                 'path': '../video/1080_4669440/dash/segment_0.m4s'},
+                                {'duration': 4000.0,
+                                 'path': '../video/1080_4669440/dash/segment_1.m4s'}],
+                  'height': 1080,
+                  'language': None,
+                  'manifest_url': 'https://example.com/streams/1/playlist/playlist.mpd',
+                  'protocol': 'http_dash_segments',
+                  'tbr': 4669.44,
+                  'url': 'https://example.com/streams/1/playlist/playlist.mpd',
+                  'vcodec': 'avc1.42c00d',
+                  'width': 1920}],
+                {'en': [{'ext': 'vtt',
+                         'url': 'https://example.com/streams/1/subtitles/sub_en.vtt'}],
+                 'fr': [{'ext': 'vtt',
+                         'url': 'https://example.com/streams/1/subtitles/sub_fr.vtt'}]},
             )
         ]
 
-        for mpd_file, mpd_url, mpd_base_url, expected_formats in _TEST_CASES:
+        for mpd_file, mpd_url, mpd_base_url, expected_formats, expected_subtitles in _TEST_CASES:
             with io.open('./test/testdata/mpd/%s.mpd' % mpd_file,
                          mode='r', encoding='utf-8') as f:
-                formats = self.ie._parse_mpd_formats(
+                formats, subtitles = self.ie._parse_mpd_formats_subtitles(
                     compat_etree_fromstring(f.read().encode('utf-8')),
                     mpd_base_url=mpd_base_url, mpd_url=mpd_url)
                 self.ie._sort_formats(formats)
                 expect_value(self, formats, expected_formats, None)
+                expect_value(self, subtitles, expected_subtitles, None)
 
     def test_parse_f4m_formats(self):
         _TEST_CASES = [

test/testdata/mpd/subtitles.mpd

@@ -0,0 +1,31 @@
+<?xml version='1.0' encoding='UTF-8'?>
+<MPD availabilityStartTime="2015-11-10T10:01:31Z" id="e66e8502-59d8-42d4-9fc7-1ca87dbed590" mediaPresentationDuration="PT1H43M14S" minBufferTime="PT1S" profiles="urn:mpeg:dash:profile:isoff-main:2011" timeShiftBufferDepth="PT30S" type="static">
+  <Period>
+    <AdaptationSet codecs="avc1.42c00d" mimeType="video/mp4">
+      <Representation bandwidth="1712128" frameRate="25" height="720" id="720p 1712kbps" width="1280">
+        <SegmentTemplate duration="100000" initialization="../video/720_1712128/dash/init.mp4" media="../video/720_1712128/dash/segment_$Number$.m4s" startNumber="0" timescale="25" />
+      </Representation>
+      <Representation bandwidth="4669440" frameRate="25" height="1080" id="1080p 4669kbps" width="1920">
+        <SegmentTemplate duration="100000" initialization="../video/1080_4669440/dash/init.mp4" media="../video/1080_4669440/dash/segment_$Number$.m4s" startNumber="0" timescale="25" />
+      </Representation>
+    </AdaptationSet>
+    <AdaptationSet codecs="mp4a.40.2" lang="de" mimeType="audio/mp4" streamName="Deutsch stereo">
+      <Representation audioSamplingRate="48000" bandwidth="131072" id="131kbps">
+        <SegmentTemplate duration="191472" initialization="../audio/1_stereo_131072/dash/init.mp4" media="../audio/1_stereo_131072/dash/segment_$Number$.m4s" startNumber="0" timescale="48" />
+      </Representation>
+      <Representation audioSamplingRate="48000" bandwidth="196608" id="196kbps">
+        <SegmentTemplate duration="191472" initialization="../audio/1_stereo_196608/dash/init.mp4" media="../audio/1_stereo_196608/dash/segment_$Number$.m4s" startNumber="0" timescale="48" />
+      </Representation>
+    </AdaptationSet>
+    <AdaptationSet lang="fr" mimeType="text/vtt" streamName="Französisch">
+      <Representation>
+        <BaseURL>../subtitles/sub_fr.vtt</BaseURL>
+      </Representation>
+    </AdaptationSet>
+    <AdaptationSet lang="en" mimeType="text/vtt" streamName="Englisch">
+      <Representation>
+        <BaseURL>../subtitles/sub_en.vtt</BaseURL>
+      </Representation>
+    </AdaptationSet>
+  </Period>
+</MPD>

youtube_dl/extractor/common.py

@@ -2023,7 +2023,10 @@ class InfoExtractor(object):
             })
         return entries
 
-    def _extract_mpd_formats(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}, data=None, headers={}, query={}):
+    def _extract_mpd_formats(self, *args, **kwargs):
+        return self._extract_mpd_formats_subtitles(*args, **kwargs)[0]
+
+    def _extract_mpd_formats_subtitles(self, mpd_url, video_id, mpd_id=None, note=None, errnote=None, fatal=True, formats_dict={}, data=None, headers={}, query={}):
         res = self._download_xml_handle(
             mpd_url, video_id,
             note=note or 'Downloading MPD manifest',
@@ -2036,13 +2039,16 @@ class InfoExtractor(object):
             return []
         mpd_base_url = base_url(urlh.geturl())
 
-        return self._parse_mpd_formats(
+        return self._parse_mpd_formats_subtitles(
             mpd_doc, mpd_id=mpd_id, mpd_base_url=mpd_base_url,
             formats_dict=formats_dict, mpd_url=mpd_url)
 
-    def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
+    def _parse_mpd_formats(self, *args, **kwargs):
+        return self._parse_mpd_formats_subtitles(*args, **kwargs)[0]
+
+    def _parse_mpd_formats_subtitles(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
         """
-        Parse formats from MPD manifest.
+        Parse formats and subtitles from MPD manifest.
         References:
          1. MPEG-DASH Standard, ISO/IEC 23009-1:2014(E),
             http://standards.iso.org/ittf/PubliclyAvailableStandards/c065274_ISO_IEC_23009-1_2014.zip
@@ -2119,6 +2125,7 @@ class InfoExtractor(object):
 
         mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
         formats = []
+        subtitles = {}
         for period in mpd_doc.findall(_add_ns('Period')):
             period_duration = parse_duration(period.get('duration')) or mpd_duration
             period_ms_info = extract_multisegment_info(period, {
@@ -2134,26 +2141,27 @@ class InfoExtractor(object):
                         continue
                     representation_attrib = adaptation_set.attrib.copy()
                     representation_attrib.update(representation.attrib)
-                    # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
+
-                    mime_type = representation_attrib['mimeType']
-                    content_type = mime_type.split('/')[0]
-                    if content_type == 'text':
-                        # TODO implement WebVTT downloading
-                        pass
-                    elif content_type in ('video', 'audio'):
                     base_url = ''
                     for element in (representation, adaptation_set, period, mpd_doc):
                         base_url_e = element.find(_add_ns('BaseURL'))
                         if base_url_e is not None:
-                                base_url = base_url_e.text + base_url
+                            base_url = compat_urlparse.urljoin(base_url_e.text, base_url)
                             if re.match(r'^https?://', base_url):
                                 break
                     if mpd_base_url and not re.match(r'^https?://', base_url):
                         if not mpd_base_url.endswith('/') and not base_url.startswith('/'):
                             mpd_base_url += '/'
-                            base_url = mpd_base_url + base_url
+                        base_url = compat_urlparse.urljoin(mpd_base_url, base_url)
                     representation_id = representation_attrib.get('id')
                     lang = representation_attrib.get('lang')
+
+                    # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
+                    mime_type = representation_attrib['mimeType']
+                    content_type = mime_type.split('/')[0]
+                    if content_type == 'text':
+                        subtitles[lang] = [{'ext': mimetype2ext(mime_type), 'url': base_url}]
+                    elif content_type in ('video', 'audio'):
                         url_el = representation.find(_add_ns('BaseURL'))
                         filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
                         bandwidth = int_or_none(representation_attrib.get('bandwidth'))
@@ -2329,7 +2337,7 @@ class InfoExtractor(object):
                         formats.append(full_info)
                     else:
                         self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
-        return formats
+        return formats, subtitles
 
     def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True, data=None, headers={}, query={}):
         res = self._download_xml_handle(

youtube_dl/extractor/generic.py

@@ -2390,11 +2390,14 @@ class GenericIE(InfoExtractor):
                         xspf_base_url=full_response.geturl()),
                     video_id)
             elif re.match(r'(?i)^(?:{[^}]+})?MPD$', doc.tag):
-                info_dict['formats'] = self._parse_mpd_formats(
+                formats, subtitles = self._parse_mpd_formats_subtitles(
                     doc,
                     mpd_base_url=full_response.geturl().rpartition('/')[0],
                     mpd_url=url)
-                self._sort_formats(info_dict['formats'])
+                self._sort_formats(formats)
+                info_dict['formats'] = formats
+                if subtitles:
+                    info_dict['subtitles'] = subtitles
                 return info_dict
             elif re.match(r'^{http://ns\.adobe\.com/f4m/[12]\.0}manifest$', doc.tag):
                 info_dict['formats'] = self._parse_f4m_formats(doc, url, video_id)