[extractor/teamcoco] Fix extractor (#6437)

Closes #6339
Authored by: bashonly

yt_dlp/extractor/_extractors.py

@@ -1846,7 +1846,10 @@ from .teachertube import (
     TeacherTubeUserIE,
 )
 from .teachingchannel import TeachingChannelIE
-from .teamcoco import TeamcocoIE
+from .teamcoco import (
+    TeamcocoIE,
+    ConanClassicIE,
+)
 from .teamtreehouse import TeamTreeHouseIE
 from .techtalks import TechTalksIE
 from .ted import (

yt_dlp/extractor/teamcoco.py

@@ -1,57 +1,109 @@
 import json
+import re
 
 from .turner import TurnerBaseIE
 from ..utils import (
-    determine_ext,
     ExtractorError,
-    int_or_none,
+    clean_html,
+    determine_ext,
+    make_archive_id,
+    merge_dicts,
     mimetype2ext,
     parse_duration,
-    parse_iso8601,
+    parse_qs,
-    qualities,
+    traverse_obj,
+    unified_timestamp,
+    urljoin,
+    url_or_none,
 )
 
 
-class TeamcocoIE(TurnerBaseIE):
+class TeamcocoBaseIE(TurnerBaseIE):
-    _VALID_URL = r'https?://(?:\w+\.)?teamcoco\.com/(?P<id>([^/]+/)*[^/?#]+)'
+    _QUALITIES = {
+        'low': (480, 272),
+        'sd': (640, 360),
+        'hd': (1280, 720),
+        'uhd': (1920, 1080),
+    }
+
+    def _get_formats_and_subtitles(self, info, video_id):
+        formats, subtitles = [], {}
+
+        for src in traverse_obj(info, ('src', ..., {dict})):
+            format_id = src.get('label')
+            src_url = src.get('src')
+            if re.match(r'https?:/[^/]', src_url):
+                src_url = src_url.replace(':/', '://', 1)
+            ext = determine_ext(src_url, mimetype2ext(src.get('type')))
+
+            if not format_id or not src_url:
+                continue
+            elif format_id == 'hls' or ext == 'm3u8':
+                fmts, subs = self._extract_m3u8_formats_and_subtitles(
+                    src_url, video_id, 'mp4', m3u8_id=format_id, fatal=False)
+                formats.extend(fmts)
+                self._merge_subtitles(subs, target=subtitles)
+
+            elif format_id in self._QUALITIES:
+                if src_url.startswith('/mp4:protected/'):
+                    # TODO: Correct extraction for these files
+                    continue
+                formats.append({
+                    'url': src_url,
+                    'ext': ext,
+                    'format_id': format_id,
+                    'width': self._QUALITIES[format_id][0],
+                    'height': self._QUALITIES[format_id][1],
+                })
+
+        return formats, subtitles
+
+
+class TeamcocoIE(TeamcocoBaseIE):
+    _VALID_URL = r'https?://(?:www\.)?teamcoco\.com/(?P<id>([^/]+/)*[^/?#]+)'
     _TESTS = [
         {
             'url': 'http://teamcoco.com/video/mary-kay-remote',
-            'md5': '55d532f81992f5c92046ad02fec34d7d',
             'info_dict': {
                 'id': '80187',
+                'display_id': 'video_mary-kay-remote',
                 'ext': 'mp4',
                 'title': 'Conan Becomes A Mary Kay Beauty Consultant',
-                'description': 'Mary Kay is perhaps the most trusted name in female beauty, so of course Conan is a natural choice to sell their products.',
+                'description': 'md5:9fb64e45b5aef6b2af1b67612b36c162',
-                'duration': 495.0,
+                'thumbnail': 'https://teamcoco.com/image/thumb?id=80187',
                 'upload_date': '20140402',
-                'timestamp': 1396407600,
+                'timestamp': 1396440000,
-            }
+            },
+            'params': {
+                'skip_download': 'm3u8',
+            },
         }, {
             'url': 'http://teamcoco.com/video/louis-ck-interview-george-w-bush',
-            'md5': 'cde9ba0fa3506f5f017ce11ead928f9a',
             'info_dict': {
                 'id': '19705',
+                'display_id': 'video_louis-ck-interview-george-w-bush',
                 'ext': 'mp4',
-                'description': 'Louis C.K. got starstruck by George W. Bush, so what? Part one.',
                 'title': 'Louis C.K. Interview Pt. 1 11/3/11',
-                'duration': 288,
+                'description': 'Louis C.K. got starstruck by George W. Bush, so what? Part one.',
+                'thumbnail': 'https://teamcoco.com/image/thumb?id=19705',
                 'upload_date': '20111104',
-                'timestamp': 1320405840,
+                'timestamp': 1320408000,
-            }
+            },
+            'params': {
+                'skip_download': 'm3u8',
+            },
         }, {
             'url': 'http://teamcoco.com/video/timothy-olyphant-drinking-whiskey',
             'info_dict': {
                 'id': '88748',
+                'display_id': 'video_timothy-olyphant-drinking-whiskey',
                 'ext': 'mp4',
                 'title': 'Timothy Olyphant Raises A Toast To “Justified”',
                 'description': 'md5:15501f23f020e793aeca761205e42c24',
                 'upload_date': '20150415',
-                'timestamp': 1429088400,
+                'timestamp': 1429099200,
+                'thumbnail': 'https://teamcoco.com/image/thumb?id=88748',
             },
-            'params': {
-                'skip_download': True,  # m3u8 downloads
-            }
         }, {
             'url': 'http://teamcoco.com/video/full-episode-mon-6-1-joel-mchale-jake-tapper-and-musical-guest-courtney-barnett?playlist=x;eyJ0eXBlIjoidGFnIiwiaWQiOjl9',
             'info_dict': {
@@ -60,9 +112,6 @@ class TeamcocoIE(TurnerBaseIE):
                 'title': 'Full Episode - Mon. 6/1 - Joel McHale, Jake Tapper, And Musical Guest Courtney Barnett',
                 'description': 'Guests: Joel McHale, Jake Tapper, And Musical Guest Courtney Barnett',
             },
-            'params': {
-                'skip_download': True,  # m3u8 downloads
-            },
             'skip': 'This video is no longer available.',
         }, {
             'url': 'http://teamcoco.com/video/the-conan-audiencey-awards-for-04/25/18',
@@ -76,126 +125,156 @@ class TeamcocoIE(TurnerBaseIE):
         }, {
             'url': 'http://teamcoco.com/israel/conan-hits-the-streets-beaches-of-tel-aviv',
             'only_matching': True,
+        },
+    ]
+
+    def _real_extract(self, url):
+        display_id = self._match_id(url).replace('/', '_')
+        webpage = self._download_webpage(url, display_id)
+        data = self._search_nextjs_data(webpage, display_id)['props']['pageProps']['pageData']
+        info = merge_dicts(*traverse_obj(data, (
+            'blocks', lambda _, v: v['name'] in ('meta-tags', 'video-player', 'video-info'), 'props', {dict})))
+
+        thumbnail = traverse_obj(
+            info, (('image', 'poster'), {lambda x: urljoin('https://teamcoco.com/', x)}), get_all=False)
+        video_id = traverse_obj(parse_qs(thumbnail), ('id', 0)) or display_id
+
+        formats, subtitles = self._get_formats_and_subtitles(info, video_id)
+
+        return {
+            'id': video_id,
+            'display_id': display_id,
+            'formats': formats,
+            'subtitles': subtitles,
+            'thumbnail': thumbnail,
+            **traverse_obj(info, {
+                'title': 'title',
+                'description': (('descriptionHtml', 'description'), {clean_html}),
+                'timestamp': ('publishedOn', {lambda x: f'{x} 12:00AM'}, {unified_timestamp}),
+            }, get_all=False),
+        }
+
+
+class ConanClassicIE(TeamcocoBaseIE):
+    _VALID_URL = r'https?://(?:(?:www\.)?conanclassic|conan25\.teamcoco)\.com/(?P<id>([^/]+/)*[^/?#]+)'
+    _TESTS = [{
+        'url': 'https://conanclassic.com/video/ice-cube-kevin-hart-conan-share-lyft',
+        'info_dict': {
+            'id': '74709',
+            'ext': 'mp4',
+            'title': 'Ice Cube, Kevin Hart, & Conan Share A Lyft Car',
+            'display_id': 'video/ice-cube-kevin-hart-conan-share-lyft',
+            'description': 'The stars of "Ride Along" teach Conan how to roll around Hollywood.',
+            'thumbnail': 'http://cdn.teamcococdn.com/image/640x360/lyft-5bd75f82b616c.png',
+            'duration': 570.0,
+            'upload_date': '20131211',
+            'timestamp': 1386721620,
+            '_old_archive_ids': ['teamcoco 74709'],
+        },
+        'params': {'skip_download': 'm3u8'},
     }, {
         'url': 'https://conan25.teamcoco.com/video/ice-cube-kevin-hart-conan-share-lyft',
         'only_matching': True,
-        }
+    }]
-    ]
+
-    _RECORD_TEMPL = '''id
+    _GRAPHQL_QUERY = '''query find($id: ID!) {
+  findRecord(id: $id) {
+
+... on MetaInterface {
+  id
   title
   teaser
   publishOn
+  slug
   thumb {
-          preview
-        }
-        tags {
-          name
-        }
-        duration
-        turnerMediaId
-        turnerMediaAuthToken'''
 
-    def _graphql_call(self, query_template, object_type, object_id):
+... on FileInterface {
-        find_object = 'find' + object_type
+  id
-        return self._download_json(
+  path
-            'https://teamcoco.com/graphql', object_id, data=json.dumps({
+  preview
-                'query': query_template % (find_object, object_id)
+  mime
-            }).encode(), headers={
+}
-                'Content-Type': 'application/json',
+
-            })['data'][find_object]
+  }
+}
+
+... on Video {
+  videoType
+  duration
+  isLive
+  youtubeId
+  turnerMediaId
+  turnerMediaAuthToken
+  airDate
+}
+
+... on Episode {
+  airDate
+  seasonNumber
+  episodeNumber
+  guestNames
+}
+
+  }
+  findRecordVideoMetadata(id: $id) {
+    turnerMediaId
+    turnerMediaAuthToken
+    duration
+    src
+  }
+}'''
 
     def _real_extract(self, url):
         display_id = self._match_id(url)
+        webpage = self._download_webpage(url, display_id)
+        data = self._search_nextjs_data(webpage, display_id)['props']['pageProps']['pageData']
+        video_id = traverse_obj(
+            data, ('blocks', ..., 'props', 'fieldDefs', lambda _, v: v['name'] == 'incomingVideoId', 'value'),
+            ('blocks', ..., 'props', 'fields', 'incomingVideoRecord', 'id'), get_all=False)
+        if not video_id:
+            self.raise_no_formats('Unable to extract video ID from webpage', expected=True)
 
-        response = self._graphql_call('''{
+        response = self._download_json(
-  %%s(slug: "%%s") {
+            'https://conanclassic.com/api/legacy/graphql', video_id, data=json.dumps({
-    ... on RecordSlug {
+                'query': self._GRAPHQL_QUERY,
-      record {
+                'variables': {'id': video_id},
-        %s
+            }, separators=(',', ':')).encode(), headers={
-      }
+                'Content-Type': 'application/json',
-    }
+            })
-    ... on PageSlug {
-      child {
-        id
-      }
-    }
-    ... on NotFoundSlug {
-      status
-    }
-  }
-}''' % self._RECORD_TEMPL, 'Slug', display_id)
-        if response.get('status'):
-            raise ExtractorError('This video is no longer available.', expected=True)
 
-        child = response.get('child')
+        info = traverse_obj(response, ('data', 'findRecord', {
-        if child:
+            'title': 'title',
-            record = self._graphql_call('''{
+            'description': 'teaser',
-  %%s(id: "%%s") {
+            'thumbnail': ('thumb', 'preview', {url_or_none}),
-    ... on Video {
+            'duration': ('duration', {parse_duration}),
-      %s
+            'timestamp': ('publishOn', {unified_timestamp}),
-    }
+        }))
-  }
-}''' % self._RECORD_TEMPL, 'Record', child['id'])
-        else:
-            record = response['record']
-        video_id = record['id']
 
-        info = {
+        media_id = traverse_obj(
-            'id': video_id,
+            response, ('data', ('findRecord', 'findRecordVideoMetadata'), 'turnerMediaId'), get_all=False)
-            'display_id': display_id,
-            'title': record['title'],
-            'thumbnail': record.get('thumb', {}).get('preview'),
-            'description': record.get('teaser'),
-            'duration': parse_duration(record.get('duration')),
-            'timestamp': parse_iso8601(record.get('publishOn')),
-        }
-
-        media_id = record.get('turnerMediaId')
         if media_id:
+            token = traverse_obj(
+                response, ('data', ('findRecord', 'findRecordVideoMetadata'), 'turnerMediaAuthToken'), get_all=False)
+            if not token:
+                raise ExtractorError('No Turner Media auth token found in API response')
             self._initialize_geo_bypass({
                 'countries': ['US'],
             })
             info.update(self._extract_ngtv_info(media_id, {
-                'accessToken': record['turnerMediaAuthToken'],
+                'accessToken': token,
                 'accessTokenType': 'jws',
             }))
         else:
-            video_sources = self._download_json(
+            formats, subtitles = self._get_formats_and_subtitles(
-                'https://teamcoco.com/_truman/d/' + video_id,
+                traverse_obj(response, ('data', 'findRecordVideoMetadata')), video_id)
-                video_id)['meta']['src']
+            info.update({
-            if isinstance(video_sources, dict):
+                'formats': formats,
-                video_sources = video_sources.values()
+                'subtitles': subtitles,
-
-            formats = []
-            get_quality = qualities(['low', 'sd', 'hd', 'uhd'])
-            for src in video_sources:
-                if not isinstance(src, dict):
-                    continue
-                src_url = src.get('src')
-                if not src_url:
-                    continue
-                format_id = src.get('label')
-                ext = determine_ext(src_url, mimetype2ext(src.get('type')))
-                if format_id == 'hls' or ext == 'm3u8':
-                    # compat_urllib_parse.urljoin does not work here
-                    if src_url.startswith('/'):
-                        src_url = 'http://ht.cdn.turner.com/tbs/big/teamcoco' + src_url
-                    formats.extend(self._extract_m3u8_formats(
-                        src_url, video_id, 'mp4', m3u8_id=format_id, fatal=False))
-                else:
-                    if src_url.startswith('/mp4:protected/'):
-                        # TODO Correct extraction for these files
-                        continue
-                    tbr = int_or_none(self._search_regex(
-                        r'(\d+)k\.mp4', src_url, 'tbr', default=None))
-
-                    formats.append({
-                        'url': src_url,
-                        'ext': ext,
-                        'tbr': tbr,
-                        'format_id': format_id,
-                        'quality': get_quality(format_id),
             })
-            info['formats'] = formats
 
-        return info
+        return {
+            'id': video_id,
+            'display_id': display_id,
+            '_old_archive_ids': [make_archive_id('Teamcoco', video_id)],
+            **info,
+        }