Merge e84b4d1c9ea50e11b687df93f3407bfabba9b729 into 3eb8d22ddb8982ca4fb56bb7a8d6517538bf14c6

* thx seproDev, yt-dlp/yt-dlp/pull/12694/commits/ae69e3c
* also add documenting comment in `InfoExtractor`

youtube_dl/extractor/bokecc.py

@@ -32,7 +32,7 @@ class BokeCCBaseIE(InfoExtractor):
 
 
 class BokeCCIE(BokeCCBaseIE):
-    _IE_DESC = 'CC视频'
+    IE_DESC = 'CC视频'
     _VALID_URL = r'https?://union\.bokecc\.com/playvideo\.bo\?(?P<query>.*)'
 
     _TESTS = [{

youtube_dl/extractor/cloudy.py

@@ -9,7 +9,7 @@ from ..utils import (
 
 
 class CloudyIE(InfoExtractor):
-    _IE_DESC = 'cloudy.ec'
+    IE_DESC = 'cloudy.ec'
     _VALID_URL = r'https?://(?:www\.)?cloudy\.ec/(?:v/|embed\.php\?.*?\bid=)(?P<id>[A-Za-z0-9]+)'
     _TESTS = [{
         'url': 'https://www.cloudy.ec/v/af511e2527aac',

youtube_dl/extractor/common.py

@@ -422,6 +422,8 @@ class InfoExtractor(object):
     _GEO_COUNTRIES = None
     _GEO_IP_BLOCKS = None
     _WORKING = True
+    # supply this in public subclasses: used in supported sites list, etc
+    # IE_DESC = 'short description of IE'
 
     def __init__(self, downloader=None):
         """Constructor. Receives an optional downloader."""

youtube_dl/extractor/itv.py

@@ -35,15 +35,6 @@ from ..utils import (
 
 class ITVBaseIE(InfoExtractor):
 
-    def _search_nextjs_data(self, webpage, video_id, **kw):
-        transform_source = kw.pop('transform_source', None)
-        fatal = kw.pop('fatal', True)
-        return self._parse_json(
-            self._search_regex(
-                r'''<script\b[^>]+\bid=('|")__NEXT_DATA__\1[^>]*>(?P<js>[^<]+)</script>''',
-                webpage, 'next.js data', group='js', fatal=fatal, **kw),
-            video_id, transform_source=transform_source, fatal=fatal)
-
     def __handle_request_webpage_error(self, err, video_id=None, errnote=None, fatal=True):
         if errnote is False:
             return False
@@ -109,7 +100,9 @@ class ITVBaseIE(InfoExtractor):
 
 class ITVIE(ITVBaseIE):
     _VALID_URL = r'https?://(?:www\.)?itv\.com/(?:(?P<w>watch)|hub)/[^/]+/(?(w)[\w-]+/)(?P<id>\w+)'
-    _IE_DESC = 'ITVX'
+    IE_DESC = 'ITVX'
+    _WORKING = False
+
     _TESTS = [{
         'note': 'Hub URLs redirect to ITVX',
         'url': 'https://www.itv.com/hub/liar/2a4547a0012',
@@ -270,7 +263,7 @@ class ITVIE(ITVBaseIE):
                 'ext': determine_ext(href, 'vtt'),
             })
 
-        next_data = self._search_nextjs_data(webpage, video_id, fatal=False, default='{}')
+        next_data = self._search_nextjs_data(webpage, video_id, fatal=False, default={})
         video_data.update(traverse_obj(next_data, ('props', 'pageProps', ('title', 'episode')), expected_type=dict)[0] or {})
         title = traverse_obj(video_data, 'headerTitle', 'episodeTitle')
         info = self._og_extract(webpage, require_title=not title)
@@ -323,7 +316,7 @@ class ITVIE(ITVBaseIE):
 
 class ITVBTCCIE(ITVBaseIE):
     _VALID_URL = r'https?://(?:www\.)?itv\.com/(?!(?:watch|hub)/)(?:[^/]+/)+(?P<id>[^/?#&]+)'
-    _IE_DESC = 'ITV articles: News, British Touring Car Championship'
+    IE_DESC = 'ITV articles: News, British Touring Car Championship'
     _TESTS = [{
         'note': 'British Touring Car Championship',
         'url': 'https://www.itv.com/btcc/articles/btcc-2018-all-the-action-from-brands-hatch',

youtube_dl/extractor/porn91.py

@@ -1,11 +1,21 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
+import re
+
 from .common import InfoExtractor
+from ..compat import (
+    compat_urllib_parse_unquote,
+)
 from ..utils import (
+    clean_html,
+    extract_attributes,
+    get_elements_by_class,
     parse_duration,
-    int_or_none,
     ExtractorError,
+    str_to_int,
+    strip_or_none,
+    unified_strdate,
 )
 
 
@@ -15,18 +25,21 @@ class Porn91IE(InfoExtractor):
 
     _TEST = {
         'url': 'http://91porn.com/view_video.php?viewkey=7e42283b4f5ab36da134',
-        'md5': '7fcdb5349354f40d41689bd0fa8db05a',
+        'md5': 'd869db281402e0ef4ddef3c38b866f86',
         'info_dict': {
             'id': '7e42283b4f5ab36da134',
             'title': '18岁大一漂亮学妹，水嫩性感，再爽一次！',
             'ext': 'mp4',
             'duration': 431,
             'age_limit': 18,
+            'upload_date': '20150520',
+            'uploader': '千岁九王爷',
         }
     }
 
     def _real_extract(self, url):
         video_id = self._match_id(url)
+        # set language for page to be extracted
         self._set_cookie('91porn.com', 'language', 'cn_CN')
 
         webpage = self._download_webpage(
@@ -35,29 +48,53 @@ class Porn91IE(InfoExtractor):
         if '作为游客，你每天只可观看10个视频' in webpage:
             raise ExtractorError('91 Porn says: Daily limit 10 videos exceeded', expected=True)
 
-        title = self._search_regex(
-            r'<div id="viewvideo-title">([^<]+)</div>', webpage, 'title')
-        title = title.replace('\n', '')
+        title = self._html_search_regex(
+            r'(?s)<title\b[^>]*>\s*(.+?)\s*(?:Chinese\s+homemade\s+video\s*)?</title', webpage, 'title')
 
         video_link_url = self._search_regex(
-            r'<textarea[^>]+id=["\']fm-video_link[^>]+>([^<]+)</textarea>',
-            webpage, 'video link')
-        videopage = self._download_webpage(video_link_url, video_id)
+            r'''document\s*\.\s*write\s*\(\s*strencode2\s*\((?P<q>"|')(?P<enc_str>[%\da-fA-F]+)(?P=q)\s*\)''',
+            webpage, 'video link', group='enc_str')
+        video_link_url = compat_urllib_parse_unquote(video_link_url)
 
-        info_dict = self._parse_html5_media_entries(url, videopage, video_id)[0]
+        info_dict = self._parse_html5_media_entries(url, '<video>%s</video>' % (video_link_url, ), video_id)[0]
 
-        duration = parse_duration(self._search_regex(
-            r'时长:\s*</span>\s*(\d+:\d+)', webpage, 'duration', fatal=False))
+        # extract various fields in <tag class=info>Name: value</tag>
+        FIELD_MAP = {
+            # cn_CN name: (yt-dl key, value parser, en name)
+            '时长': ('duration', parse_duration, 'Runtime', ),
+            '查看': ('view_count', str_to_int, 'Views', ),
+            '留言': ('comment_count', str_to_int, 'Comments', ),
+            '收藏': ('like_count', str_to_int, 'Favorites', ),
+            '添加时间': ('upload_date', unified_strdate, 'Added', ),
+            # same as title for en, not description for cn_CN
+            '__ignore__': ('description', strip_or_none, 'Description', ),
+            '作者': ('uploader', strip_or_none, 'From', ),
+        }
+        # yt-dl's original implementation of get_elements_by_class() uses regex
+        # yt-dlp uses an actual HTML parser, and can be confused by bad HTML fragments
+        for elt in get_elements_by_class(
+                'info',
+                # concatenate <span>s ...
+                re.sub(r'(?i)</span>\s*<span\b[^>]*?>', '',
+                       # ... and strip out possibly unbalanced <font> for yt-dlp
+                       re.sub(r'(?i)(?:<font\b[^>]*?>|</font\s*>)', '', webpage))) or []:
+            elt = re.split(r':\s*', clean_html(elt), 1)
+            if len(elt) != 2 or elt[1] == '':
+                continue
+            parm = FIELD_MAP.get(elt[0].strip())
+            if parm and elt[1] is not None:
+                info_dict[parm[0]] = parm[1](elt[1]) if parm[1] else elt[1]
 
-        comment_count = int_or_none(self._search_regex(
-            r'留言:\s*</span>\s*(\d+)', webpage, 'comment count', fatal=False))
+        thumbnail = extract_attributes(
+            self._search_regex(
+                r'''(<video\b[^>]*\bid\s*=\s*(?P<q>"|')player_one(?P=q)[^>]*>)''',
+                webpage, 'poster', default='')).get('poster')
 
         info_dict.update({
             'id': video_id,
             'title': title,
-            'duration': duration,
-            'comment_count': comment_count,
             'age_limit': self._rta_search(webpage),
+            'thumbnail': thumbnail,
         })
 
         return info_dict

youtube_dl/extractor/senateisvp.py

@@ -47,7 +47,7 @@ class SenateISVPIE(InfoExtractor):
         ['vetaff', '76462', 'http://vetaff-f.akamaihd.net'],
         ['arch', '', 'http://ussenate-f.akamaihd.net/']
     ]
-    _IE_NAME = 'senate.gov'
+    IE_NAME = 'senate.gov'
     _VALID_URL = r'https?://(?:www\.)?senate\.gov/isvp/?\?(?P<qs>.+)'
     _TESTS = [{
         'url': 'http://www.senate.gov/isvp/?comm=judiciary&type=live&stt=&filename=judiciary031715&auto_play=false&wmode=transparent&poster=http%3A%2F%2Fwww.judiciary.senate.gov%2Fthemes%2Fjudiciary%2Fimages%2Fvideo-poster-flash-fit.png',

youtube_dl/jsinterp.py

@@ -686,6 +686,8 @@ class JSInterpreter(object):
             raise self.Exception('Cannot get index {idx!r:.100}'.format(**locals()), expr=repr(obj), cause=e)
 
     def _dump(self, obj, namespace):
+        if obj is JS_Undefined:
+            return 'undefined'
         try:
             return json.dumps(obj)
         except TypeError: