merge youtube-dl master 22.09.2020

ChangeLog

@@ -1,3 +1,24 @@
+version 2020.09.20
+
+Core
+* [extractor/common] Relax interaction count extraction in _json_ld
++ [extractor/common] Extract author as uploader for VideoObject in _json_ld
+* [downloader/hls] Fix incorrect end byte in Range HTTP header for
+  media segments with EXT-X-BYTERANGE (#14748, #24512)
+* [extractor/common] Handle ssl.CertificateError in _request_webpage (#26601)
+* [downloader/http] Improve timeout detection when reading block of data
+  (#10935)
+* [downloader/http] Retry download when urlopen times out (#10935, #26603)
+
+Extractors
+* [redtube] Extend URL regular expression (#26506)
+* [twitch] Refactor
+* [twitch:stream] Switch to GraphQL and fix reruns (#26535)
++ [telequebec] Add support for brightcove videos (#25833)
+* [pornhub] Extract metadata from JSON-LD (#26614)
+* [pornhub] Fix view count extraction (#26621, #26614)
+
+
 version 2020.09.14
 
 Core

youtube_dlc/downloader/hls.py

@@ -141,7 +141,7 @@ class HlsFD(FragmentFD):
                     count = 0
                     headers = info_dict.get('http_headers', {})
                     if byte_range:
-                        headers['Range'] = 'bytes=%d-%d' % (byte_range['start'], byte_range['end'])
+                        headers['Range'] = 'bytes=%d-%d' % (byte_range['start'], byte_range['end'] - 1)
                     while count <= fragment_retries:
                         try:
                             success, frag_content = self._download_fragment(

youtube_dlc/downloader/http.py

@@ -106,7 +106,12 @@ class HttpFD(FileDownloader):
                 set_range(request, range_start, range_end)
             # Establish connection
             try:
-                ctx.data = self.ydl.urlopen(request)
+                try:
+                    ctx.data = self.ydl.urlopen(request)
+                except (compat_urllib_error.URLError, ) as err:
+                    if isinstance(err.reason, socket.timeout):
+                        raise RetryDownload(err)
+                    raise err
                 # When trying to resume, Content-Range HTTP header of response has to be checked
                 # to match the value of requested Range HTTP header. This is due to a webservers
                 # that don't support resuming and serve a whole file with no Content-Range
@@ -218,9 +223,10 @@ class HttpFD(FileDownloader):
 
             def retry(e):
                 to_stdout = ctx.tmpfilename == '-'
-                if not to_stdout:
-                    ctx.stream.close()
-                ctx.stream = None
+                if ctx.stream is not None:
+                    if not to_stdout:
+                        ctx.stream.close()
+                    ctx.stream = None
                 ctx.resume_len = byte_counter if to_stdout else os.path.getsize(encodeFilename(ctx.tmpfilename))
                 raise RetryDownload(e)
 
@@ -233,9 +239,11 @@ class HttpFD(FileDownloader):
                 except socket.timeout as e:
                     retry(e)
                 except socket.error as e:
-                    if e.errno not in (errno.ECONNRESET, errno.ETIMEDOUT):
-                        raise
-                    retry(e)
+                    # SSLError on python 2 (inherits socket.error) may have
+                    # no errno set but this error message
+                    if e.errno in (errno.ECONNRESET, errno.ETIMEDOUT) or getattr(e, 'message', None) == 'The read operation timed out':
+                        retry(e)
+                    raise
 
                 byte_counter += len(data_block)
 

youtube_dlc/extractor/common.py

@@ -10,6 +10,7 @@ import os
 import random
 import re
 import socket
+import ssl
 import sys
 import time
 import math
@@ -67,6 +68,7 @@ from ..utils import (
     sanitized_Request,
     sanitize_filename,
     str_or_none,
+    str_to_int,
     strip_or_none,
     unescapeHTML,
     unified_strdate,
@@ -623,9 +625,12 @@ class InfoExtractor(object):
                 url_or_request = update_url_query(url_or_request, query)
             if data is not None or headers:
                 url_or_request = sanitized_Request(url_or_request, data, headers)
+        exceptions = [compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error]
+        if hasattr(ssl, 'CertificateError'):
+            exceptions.append(ssl.CertificateError)
         try:
             return self._downloader.urlopen(url_or_request)
-        except (compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error) as err:
+        except tuple(exceptions) as err:
             if isinstance(err, compat_urllib_error.HTTPError):
                 if self.__can_accept_status_code(err, expected_status):
                     # Retain reference to error to prevent file object from
@@ -1244,7 +1249,10 @@ class InfoExtractor(object):
                 interaction_type = is_e.get('interactionType')
                 if not isinstance(interaction_type, compat_str):
                     continue
-                interaction_count = int_or_none(is_e.get('userInteractionCount'))
+                # For interaction count some sites provide string instead of
+                # an integer (as per spec) with non digit characters (e.g. ",")
+                # so extracting count with more relaxed str_to_int
+                interaction_count = str_to_int(is_e.get('userInteractionCount'))
                 if interaction_count is None:
                     continue
                 count_kind = INTERACTION_TYPE_MAP.get(interaction_type.split('/')[-1])
@@ -1264,6 +1272,7 @@ class InfoExtractor(object):
                 'thumbnail': url_or_none(e.get('thumbnailUrl') or e.get('thumbnailURL')),
                 'duration': parse_duration(e.get('duration')),
                 'timestamp': unified_timestamp(e.get('uploadDate')),
+                'uploader': str_or_none(e.get('author')),
                 'filesize': float_or_none(e.get('contentSize')),
                 'tbr': int_or_none(e.get('bitrate')),
                 'width': int_or_none(e.get('width')),

youtube_dlc/extractor/pornhub.py

@@ -17,6 +17,7 @@ from ..utils import (
     determine_ext,
     ExtractorError,
     int_or_none,
+    merge_dicts,
     NO_DEFAULT,
     orderedSet,
     remove_quotes,
@@ -59,13 +60,14 @@ class PornHubIE(PornHubBaseIE):
                     '''
     _TESTS = [{
         'url': 'http://www.pornhub.com/view_video.php?viewkey=648719015',
-        'md5': '1e19b41231a02eba417839222ac9d58e',
+        'md5': 'a6391306d050e4547f62b3f485dd9ba9',
         'info_dict': {
             'id': '648719015',
             'ext': 'mp4',
             'title': 'Seductive Indian beauty strips down and fingers her pink pussy',
             'uploader': 'Babes',
             'upload_date': '20130628',
+            'timestamp': 1372447216,
             'duration': 361,
             'view_count': int,
             'like_count': int,
@@ -82,8 +84,8 @@ class PornHubIE(PornHubBaseIE):
             'id': '1331683002',
             'ext': 'mp4',
             'title': '重庆婷婷女王足交',
-            'uploader': 'Unknown',
             'upload_date': '20150213',
+            'timestamp': 1423804862,
             'duration': 1753,
             'view_count': int,
             'like_count': int,
@@ -121,6 +123,7 @@ class PornHubIE(PornHubBaseIE):
         'params': {
             'skip_download': True,
         },
+        'skip': 'This video has been disabled',
     }, {
         'url': 'http://www.pornhub.com/view_video.php?viewkey=ph557bbb6676d2d',
         'only_matching': True,
@@ -338,10 +341,10 @@ class PornHubIE(PornHubBaseIE):
 
         video_uploader = self._html_search_regex(
             r'(?s)From:&nbsp;.+?<(?:a\b[^>]+\bhref=["\']/(?:(?:user|channel)s|model|pornstar)/|span\b[^>]+\bclass=["\']username)[^>]+>(.+?)<',
-            webpage, 'uploader', fatal=False)
+            webpage, 'uploader', default=None)
 
         view_count = self._extract_count(
-            r'<span class="count">([\d,\.]+)</span> views', webpage, 'view')
+            r'<span class="count">([\d,\.]+)</span> [Vv]iews', webpage, 'view')
         like_count = self._extract_count(
             r'<span class="votesUp">([\d,\.]+)</span>', webpage, 'like')
         dislike_count = self._extract_count(
@@ -356,7 +359,11 @@ class PornHubIE(PornHubBaseIE):
             if div:
                 return re.findall(r'<a[^>]+\bhref=[^>]+>([^<]+)', div)
 
-        return {
+        info = self._search_json_ld(webpage, video_id, default={})
+        # description provided in JSON-LD is irrelevant
+        info['description'] = None
+
+        return merge_dicts({
             'id': video_id,
             'uploader': video_uploader,
             'upload_date': upload_date,
@@ -372,7 +379,7 @@ class PornHubIE(PornHubBaseIE):
             'tags': extract_list('tags'),
             'categories': extract_list('categories'),
             'subtitles': subtitles,
-        }
+        }, info)
 
 
 class PornHubPlaylistBaseIE(PornHubBaseIE):

youtube_dlc/extractor/redtube.py

@@ -15,7 +15,7 @@ from ..utils import (
 
 
 class RedTubeIE(InfoExtractor):
-    _VALID_URL = r'https?://(?:(?:www\.)?redtube\.com/|embed\.redtube\.com/\?.*?\bid=)(?P<id>[0-9]+)'
+    _VALID_URL = r'https?://(?:(?:\w+\.)?redtube\.com/|embed\.redtube\.com/\?.*?\bid=)(?P<id>[0-9]+)'
     _TESTS = [{
         'url': 'http://www.redtube.com/66418',
         'md5': 'fc08071233725f26b8f014dba9590005',
@@ -31,6 +31,9 @@ class RedTubeIE(InfoExtractor):
     }, {
         'url': 'http://embed.redtube.com/?bgcolor=000000&id=1443286',
         'only_matching': True,
+    }, {
+        'url': 'http://it.redtube.com/66418',
+        'only_matching': True,
     }]
 
     @staticmethod

youtube_dlc/extractor/telequebec.py

@@ -13,14 +13,24 @@ from ..utils import (
 
 class TeleQuebecBaseIE(InfoExtractor):
     @staticmethod
-    def _limelight_result(media_id):
+    def _result(url, ie_key):
         return {
             '_type': 'url_transparent',
-            'url': smuggle_url(
-                'limelight:media:' + media_id, {'geo_countries': ['CA']}),
-            'ie_key': 'LimelightMedia',
+            'url': smuggle_url(url, {'geo_countries': ['CA']}),
+            'ie_key': ie_key,
         }
 
+    @staticmethod
+    def _limelight_result(media_id):
+        return TeleQuebecBaseIE._result(
+            'limelight:media:' + media_id, 'LimelightMedia')
+
+    @staticmethod
+    def _brightcove_result(brightcove_id):
+        return TeleQuebecBaseIE._result(
+            'http://players.brightcove.net/6150020952001/default_default/index.html?videoId=%s'
+            % brightcove_id, 'BrightcoveNew')
+
 
 class TeleQuebecIE(TeleQuebecBaseIE):
     _VALID_URL = r'''(?x)
@@ -37,11 +47,27 @@ class TeleQuebecIE(TeleQuebecBaseIE):
             'id': '577116881b4b439084e6b1cf4ef8b1b3',
             'ext': 'mp4',
             'title': 'Un petit choc et puis repart!',
-            'description': 'md5:b04a7e6b3f74e32d7b294cffe8658374',
+            'description': 'md5:067bc84bd6afecad85e69d1000730907',
         },
         'params': {
             'skip_download': True,
         },
+    }, {
+        'url': 'https://zonevideo.telequebec.tv/media/55267/le-soleil/passe-partout',
+        'info_dict': {
+            'id': '6167180337001',
+            'ext': 'mp4',
+            'title': 'Le soleil',
+            'description': 'md5:64289c922a8de2abbe99c354daffde02',
+            'uploader_id': '6150020952001',
+            'upload_date': '20200625',
+            'timestamp': 1593090307,
+        },
+        'params': {
+            'format': 'bestvideo',
+            'skip_download': True,
+        },
+        'add_ie': ['BrightcoveNew'],
     }, {
         # no description
         'url': 'http://zonevideo.telequebec.tv/media/30261',
@@ -58,7 +84,14 @@ class TeleQuebecIE(TeleQuebecBaseIE):
             'https://mnmedias.api.telequebec.tv/api/v2/media/' + media_id,
             media_id)['media']
 
-        info = self._limelight_result(media_data['streamInfo']['sourceId'])
+        source_id = media_data['streamInfo']['sourceId']
+        source = (try_get(
+            media_data, lambda x: x['streamInfo']['source'],
+            compat_str) or 'limelight').lower()
+        if source == 'brightcove':
+            info = self._brightcove_result(source_id)
+        else:
+            info = self._limelight_result(source_id)
         info.update({
             'title': media_data.get('title'),
             'description': try_get(

youtube_dlc/extractor/twitch.py

@@ -24,7 +24,6 @@ from ..utils import (
     parse_duration,
     parse_iso8601,
     qualities,
-    str_or_none,
     try_get,
     unified_timestamp,
     update_url_query,
@@ -337,19 +336,27 @@ def _make_video_result(node):
 class TwitchGraphQLBaseIE(TwitchBaseIE):
     _PAGE_LIMIT = 100
 
-    def _download_gql(self, video_id, op, variables, sha256_hash, note, fatal=True):
+    _OPERATION_HASHES = {
+        'CollectionSideBar': '27111f1b382effad0b6def325caef1909c733fe6a4fbabf54f8d491ef2cf2f14',
+        'FilterableVideoTower_Videos': 'a937f1d22e269e39a03b509f65a7490f9fc247d7f83d6ac1421523e3b68042cb',
+        'ClipsCards__User': 'b73ad2bfaecfd30a9e6c28fada15bd97032c83ec77a0440766a56fe0bd632777',
+        'ChannelCollectionsContent': '07e3691a1bad77a36aba590c351180439a40baefc1c275356f40fc7082419a84',
+        'StreamMetadata': '1c719a40e481453e5c48d9bb585d971b8b372f8ebb105b17076722264dfa5b3e',
+        'ComscoreStreamingQuery': 'e1edae8122517d013405f237ffcc124515dc6ded82480a88daef69c83b53ac01',
+        'VideoPreviewOverlay': '3006e77e51b128d838fa4e835723ca4dc9a05c5efd4466c1085215c6e437e65c',
+    }
+
+    def _download_gql(self, video_id, ops, note, fatal=True):
+        for op in ops:
+            op['extensions'] = {
+                'persistedQuery': {
+                    'version': 1,
+                    'sha256Hash': self._OPERATION_HASHES[op['operationName']],
+                }
+            }
         return self._download_json(
             'https://gql.twitch.tv/gql', video_id, note,
-            data=json.dumps({
-                'operationName': op,
-                'variables': variables,
-                'extensions': {
-                    'persistedQuery': {
-                        'version': 1,
-                        'sha256Hash': sha256_hash,
-                    }
-                }
-            }).encode(),
+            data=json.dumps(ops).encode(),
             headers={
                 'Content-Type': 'text/plain;charset=UTF-8',
                 'Client-ID': self._CLIENT_ID,
@@ -369,14 +376,15 @@ class TwitchCollectionIE(TwitchGraphQLBaseIE):
     }]
 
     _OPERATION_NAME = 'CollectionSideBar'
-    _SHA256_HASH = '27111f1b382effad0b6def325caef1909c733fe6a4fbabf54f8d491ef2cf2f14'
 
     def _real_extract(self, url):
         collection_id = self._match_id(url)
         collection = self._download_gql(
-            collection_id, self._OPERATION_NAME,
-            {'collectionID': collection_id}, self._SHA256_HASH,
-            'Downloading collection GraphQL')['data']['collection']
+            collection_id, [{
+                'operationName': self._OPERATION_NAME,
+                'variables': {'collectionID': collection_id},
+            }],
+            'Downloading collection GraphQL')[0]['data']['collection']
         title = collection.get('title')
         entries = []
         for edge in collection['items']['edges']:
@@ -403,14 +411,16 @@ class TwitchPlaylistBaseIE(TwitchGraphQLBaseIE):
             if cursor:
                 variables['cursor'] = cursor
             page = self._download_gql(
-                channel_name, self._OPERATION_NAME, variables,
-                self._SHA256_HASH,
+                channel_name, [{
+                    'operationName': self._OPERATION_NAME,
+                    'variables': variables,
+                }],
                 'Downloading %ss GraphQL page %s' % (self._NODE_KIND, page_num),
                 fatal=False)
             if not page:
                 break
             edges = try_get(
-                page, lambda x: x['data']['user'][entries_key]['edges'], list)
+                page, lambda x: x[0]['data']['user'][entries_key]['edges'], list)
             if not edges:
                 break
             for edge in edges:
@@ -553,7 +563,6 @@ class TwitchVideosIE(TwitchPlaylistBaseIE):
         'views': 'Popular',
     }
 
-    _SHA256_HASH = 'a937f1d22e269e39a03b509f65a7490f9fc247d7f83d6ac1421523e3b68042cb'
     _OPERATION_NAME = 'FilterableVideoTower_Videos'
     _ENTRY_KIND = 'video'
     _EDGE_KIND = 'VideoEdge'
@@ -622,7 +631,6 @@ class TwitchVideosClipsIE(TwitchPlaylistBaseIE):
     # NB: values other than 20 result in skipped videos
     _PAGE_LIMIT = 20
 
-    _SHA256_HASH = 'b73ad2bfaecfd30a9e6c28fada15bd97032c83ec77a0440766a56fe0bd632777'
     _OPERATION_NAME = 'ClipsCards__User'
     _ENTRY_KIND = 'clip'
     _EDGE_KIND = 'ClipEdge'
@@ -680,7 +688,6 @@ class TwitchVideosCollectionsIE(TwitchPlaylistBaseIE):
         'playlist_mincount': 3,
     }]
 
-    _SHA256_HASH = '07e3691a1bad77a36aba590c351180439a40baefc1c275356f40fc7082419a84'
     _OPERATION_NAME = 'ChannelCollectionsContent'
     _ENTRY_KIND = 'collection'
     _EDGE_KIND = 'CollectionsItemEdge'
@@ -717,7 +724,7 @@ class TwitchVideosCollectionsIE(TwitchPlaylistBaseIE):
             playlist_title='%s - Collections' % channel_name)
 
 
-class TwitchStreamIE(TwitchBaseIE):
+class TwitchStreamIE(TwitchGraphQLBaseIE):
     IE_NAME = 'twitch:stream'
     _VALID_URL = r'''(?x)
                     https?://
@@ -774,28 +781,43 @@ class TwitchStreamIE(TwitchBaseIE):
                 else super(TwitchStreamIE, cls).suitable(url))
 
     def _real_extract(self, url):
-        channel_name = self._match_id(url)
+        channel_name = self._match_id(url).lower()
 
-        access_token = self._download_access_token(channel_name)
+        gql = self._download_gql(
+            channel_name, [{
+                'operationName': 'StreamMetadata',
+                'variables': {'channelLogin': channel_name},
+            }, {
+                'operationName': 'ComscoreStreamingQuery',
+                'variables': {
+                    'channel': channel_name,
+                    'clipSlug': '',
+                    'isClip': False,
+                    'isLive': True,
+                    'isVodOrCollection': False,
+                    'vodID': '',
+                },
+            }, {
+                'operationName': 'VideoPreviewOverlay',
+                'variables': {'login': channel_name},
+            }],
+            'Downloading stream GraphQL')
 
-        token = access_token['token']
-        channel_id = self._extract_channel_id(token, channel_name)
+        user = gql[0]['data']['user']
 
-        stream = self._call_api(
-            'kraken/streams/%s?stream_type=all' % channel_id,
-            channel_id, 'Downloading stream JSON').get('stream')
+        if not user:
+            raise ExtractorError(
+                '%s does not exist' % channel_name, expected=True)
+
+        stream = user['stream']
 
         if not stream:
-            raise ExtractorError('%s is offline' % channel_id, expected=True)
+            raise ExtractorError('%s is offline' % channel_name, expected=True)
 
-        # Channel name may be typed if different case than the original channel name
-        # (e.g. http://www.twitch.tv/TWITCHPLAYSPOKEMON) that will lead to constructing
-        # an invalid m3u8 URL. Working around by use of original channel name from stream
-        # JSON and fallback to lowercase if it's not available.
-        channel_name = try_get(
-            stream, lambda x: x['channel']['name'],
-            compat_str) or channel_name.lower()
+        access_token = self._download_access_token(channel_name)
+        token = access_token['token']
 
+        stream_id = stream.get('id') or channel_name
         query = {
             'allow_source': 'true',
             'allow_audio_only': 'true',
@@ -808,41 +830,39 @@ class TwitchStreamIE(TwitchBaseIE):
             'token': token.encode('utf-8'),
         }
         formats = self._extract_m3u8_formats(
-            '%s/api/channel/hls/%s.m3u8?%s'
-            % (self._USHER_BASE, channel_name, compat_urllib_parse_urlencode(query)),
-            channel_id, 'mp4')
+            '%s/api/channel/hls/%s.m3u8' % (self._USHER_BASE, channel_name),
+            stream_id, 'mp4', query=query)
         self._prefer_source(formats)
 
         view_count = stream.get('viewers')
-        timestamp = parse_iso8601(stream.get('created_at'))
+        timestamp = unified_timestamp(stream.get('createdAt'))
 
-        channel = stream['channel']
-        title = self._live_title(channel.get('display_name') or channel.get('name'))
-        description = channel.get('status')
+        sq_user = try_get(gql, lambda x: x[1]['data']['user'], dict) or {}
+        uploader = sq_user.get('displayName')
+        description = try_get(
+            sq_user, lambda x: x['broadcastSettings']['title'], compat_str)
 
-        thumbnails = []
-        for thumbnail_key, thumbnail_url in stream['preview'].items():
-            m = re.search(r'(?P<width>\d+)x(?P<height>\d+)\.jpg$', thumbnail_key)
-            if not m:
-                continue
-            thumbnails.append({
-                'url': thumbnail_url,
-                'width': int(m.group('width')),
-                'height': int(m.group('height')),
-            })
+        thumbnail = url_or_none(try_get(
+            gql, lambda x: x[2]['data']['user']['stream']['previewImageURL'],
+            compat_str))
+
+        title = uploader or channel_name
+        stream_type = stream.get('type')
+        if stream_type in ['rerun', 'live']:
+            title += ' (%s)' % stream_type
 
         return {
-            'id': str_or_none(stream.get('_id')) or channel_id,
+            'id': stream_id,
             'display_id': channel_name,
-            'title': title,
+            'title': self._live_title(title),
             'description': description,
-            'thumbnails': thumbnails,
-            'uploader': channel.get('display_name'),
-            'uploader_id': channel.get('name'),
+            'thumbnail': thumbnail,
+            'uploader': uploader,
+            'uploader_id': channel_name,
             'timestamp': timestamp,
             'view_count': view_count,
             'formats': formats,
-            'is_live': True,
+            'is_live': stream_type == 'live',
         }