[youtube] Add a pseudo-extractor for truncated YouTube video IDs (#4610)

[youtube-dl] / youtube_dl / extractor / youtube.py

diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py
index 100ecbc0a11ab3d2765606d9cd90163e57797a07..8c7842ee84958d59116976bc49d25e4c77b86ad7 100644 (file)
--- a/youtube_dl/extractor/youtube.py
+++ b/youtube_dl/extractor/youtube.py
@@ -418,6 +418,19 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
                 'upload_date': '20140605',
             },
         },
+        # Age-gate video with encrypted signature
+        {
+            'url': 'http://www.youtube.com/watch?v=6kLq3WMV1nU',
+            'info_dict': {
+                'id': '6kLq3WMV1nU',
+                'ext': 'mp4',
+                'title': 'Dedication To My Ex (Miss That) (Lyric Video)',
+                'description': 'md5:33765bb339e1b47e7e72b5490139bb41',
+                'uploader': 'LloydVEVO',
+                'uploader_id': 'LloydVEVO',
+                'upload_date': '20110629',
+            },
+        },
         # video_info is None (https://github.com/rg3/youtube-dl/issues/4421)
         {
             'url': '__2ABJjxzNo',
@@ -527,8 +540,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
                 return 's[%s%s%s]' % (starts, ends, steps)
 
             step = None
-            start = '(Never used)'  # Quelch pyflakes warnings - start will be
-                                    # set as soon as step is set
+            # Quelch pyflakes warnings - start will be set when step is set
+            start = '(Never used)'
             for i, prev in zip(idxs[1:], idxs[:-1]):
                 if step is not None:
                     if i - prev == step:
@@ -599,24 +612,23 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
 
     def _get_available_subtitles(self, video_id, webpage):
         try:
-            sub_list = self._download_webpage(
+            subs_doc = self._download_xml(
                 'https://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id,
                 video_id, note=False)
         except ExtractorError as err:
             self._downloader.report_warning('unable to download video subtitles: %s' % compat_str(err))
             return {}
-        lang_list = re.findall(r'name="([^"]*)"[^>]+lang_code="([\w\-]+)"', sub_list)
 
         sub_lang_list = {}
-        for l in lang_list:
-            lang = l[1]
+        for track in subs_doc.findall('track'):
+            lang = track.attrib['lang_code']
             if lang in sub_lang_list:
                 continue
             params = compat_urllib_parse.urlencode({
                 'lang': lang,
                 'v': video_id,
                 'fmt': self._downloader.params.get('subtitlesformat', 'srt'),
-                'name': unescapeHTML(l[0]).encode('utf-8'),
+                'name': track.attrib['name'].encode('utf-8'),
             })
             url = 'https://www.youtube.com/api/timedtext?' + params
             sub_lang_list[lang] = url
@@ -649,10 +661,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
             list_url = caption_url + '&' + list_params
             caption_list = self._download_xml(list_url, video_id)
             original_lang_node = caption_list.find('track')
-            if original_lang_node is None or original_lang_node.attrib.get('kind') != 'asr':
+            if original_lang_node is None:
                 self._downloader.report_warning('Video doesn\'t have automatic captions')
                 return {}
             original_lang = original_lang_node.attrib['lang_code']
+            caption_kind = original_lang_node.attrib.get('kind', '')
 
             sub_lang_list = {}
             for lang_node in caption_list.findall('target'):
@@ -662,7 +675,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
                     'tlang': sub_lang,
                     'fmt': sub_format,
                     'ts': timestamp,
-                    'kind': 'asr',
+                    'kind': caption_kind,
                 })
                 sub_lang_list[sub_lang] = caption_url + '&' + params
             return sub_lang_list
@@ -766,11 +779,13 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
             age_gate = True
             # We simulate the access to the video from www.youtube.com/v/{video_id}
             # this can be viewed without login into Youtube
+            url = proto + '://www.youtube.com/embed/%s' % video_id
+            embed_webpage = self._download_webpage(url, video_id, 'Downloading embed webpage')
             data = compat_urllib_parse.urlencode({
                 'video_id': video_id,
                 'eurl': 'https://youtube.googleapis.com/v/' + video_id,
                 'sts': self._search_regex(
-                    r'"sts"\s*:\s*(\d+)', video_webpage, 'sts', default=''),
+                    r'"sts"\s*:\s*(\d+)', embed_webpage, 'sts', default=''),
             })
             video_info_url = proto + '://www.youtube.com/get_video_info?' + data
             video_info_webpage = self._download_webpage(
@@ -968,11 +983,10 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
                 elif 's' in url_data:
                     encrypted_sig = url_data['s'][0]
 
-                    if not age_gate:
-                        jsplayer_url_json = self._search_regex(
-                            r'"assets":.+?"js":\s*("[^"]+")',
-                            video_webpage, 'JS player URL')
-                        player_url = json.loads(jsplayer_url_json)
+                    jsplayer_url_json = self._search_regex(
+                        r'"assets":.+?"js":\s*("[^"]+")',
+                        embed_webpage if age_gate else video_webpage, 'JS player URL')
+                    player_url = json.loads(jsplayer_url_json)
                     if player_url is None:
                         player_url_json = self._search_regex(
                             r'ytplayer\.config.*?"url"\s*:\s*("[^"]+")',
@@ -1687,3 +1701,20 @@ class YoutubeTruncatedURLIE(InfoExtractor):
             '"http://www.youtube.com/watch?feature=foo&v=BaW_jenozKc" '
             ' or simply  youtube-dl BaW_jenozKc  .',
             expected=True)
+
+
+class YoutubeTruncatedIDIE(InfoExtractor):
+    IE_NAME = 'youtube:truncated_id'
+    IE_DESC = False  # Do not list
+    _VALID_URL = r'https?://(?:www\.)youtube\.com/watch\?v=(?P<id>[0-9A-Za-z_-]{1,10})$'
+
+    _TESTS = [{
+        'url': 'https://www.youtube.com/watch?v=N_708QY7Ob',
+        'only_matching': True,
+    }]
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
+        raise ExtractorError(
+            'Incomplete YouTube ID %s. URL %s looks truncated.' % (video_id, url),
+            expected=True)