Merge pull request #7320 from remitamine/adobetv

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

diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py
index 5482aac3b5ccf332867c451637092df6f994db10..9b39505ba71cf09880e6d8fcec1910b8806204c0 100644 (file)
--- a/youtube_dl/extractor/youtube.py
+++ b/youtube_dl/extractor/youtube.py
@@ -20,7 +20,6 @@ from ..compat import (
     compat_urllib_parse_unquote,
     compat_urllib_parse_unquote_plus,
     compat_urllib_parse_urlparse,
-    compat_urllib_request,
     compat_urlparse,
     compat_str,
 )
@@ -35,6 +34,7 @@ from ..utils import (
     orderedSet,
     parse_duration,
     remove_start,
+    sanitized_Request,
     smuggle_url,
     str_to_int,
     unescapeHTML,
@@ -114,7 +114,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
 
         login_data = compat_urllib_parse.urlencode(encode_dict(login_form_strs)).encode('ascii')
 
-        req = compat_urllib_request.Request(self._LOGIN_URL, login_data)
+        req = sanitized_Request(self._LOGIN_URL, login_data)
         login_results = self._download_webpage(
             req, None,
             note='Logging in', errnote='unable to log in', fatal=False)
@@ -147,7 +147,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor):
 
             tfa_data = compat_urllib_parse.urlencode(encode_dict(tfa_form_strs)).encode('ascii')
 
-            tfa_req = compat_urllib_request.Request(self._TWOFACTOR_URL, tfa_data)
+            tfa_req = sanitized_Request(self._TWOFACTOR_URL, tfa_data)
             tfa_results = self._download_webpage(
                 tfa_req, None,
                 note='Submitting TFA code', errnote='unable to submit tfa', fatal=False)
@@ -258,7 +258,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                              |(?:                                             # or the v= param in all its forms
                                  (?:(?:watch|movie)(?:_popup)?(?:\.php)?/?)?  # preceding watch(_popup|.php) or nothing (like /?v=xxxx)
                                  (?:\?|\#!?)                                  # the params delimiter ? or # or #!
-                                 (?:.*?&)??                                   # any other preceding param (like /?s=tuff&v=xxxx)
+                                 (?:.*?[&;])??                                # any other preceding param (like /?s=tuff&v=xxxx or ?s=tuff&v=V36LpHqtcDY)
                                  v=
                              )
                          ))
@@ -346,6 +346,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
         '247': {'ext': 'webm', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
         '248': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
         '271': {'ext': 'webm', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
+        # itag 272 videos are either 3840x2160 (e.g. RtoitU2A-3E) or 7680x4320 (sLprVF6d7Ug)
         '272': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40},
         '302': {'ext': 'webm', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'vp9'},
         '303': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'vp9'},
@@ -426,7 +427,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
                 'title': 'Principal Sexually Assaults A Teacher - Episode 117 - 8th June 2012',
                 'description': 'md5:09b78bd971f1e3e289601dfba15ca4f7',
                 'uploader': 'SET India',
-                'uploader_id': 'setindia'
+                'uploader_id': 'setindia',
+                'age_limit': 18,
             }
         },
         {
@@ -563,7 +565,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             'info_dict': {
                 'id': 'lqQg6PlCWgI',
                 'ext': 'mp4',
-                'upload_date': '20120724',
+                'upload_date': '20150827',
                 'uploader_id': 'olympic',
                 'description': 'HO09  - Women -  GER-AUS - Hockey - 31 July 2012 - London 2012 Olympic Games',
                 'uploader': 'Olympics',
@@ -713,6 +715,26 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             'url': 'https://www.youtube.com/watch?v=Ms7iBXnlUO8',
             'only_matching': True,
         },
+        {
+            # Video with yt:stretch=17:0
+            'url': 'https://www.youtube.com/watch?v=Q39EVAstoRM',
+            'info_dict': {
+                'id': 'Q39EVAstoRM',
+                'ext': 'mp4',
+                'title': 'Clash Of Clans#14 Dicas De Ataque Para CV 4',
+                'description': 'md5:ee18a25c350637c8faff806845bddee9',
+                'upload_date': '20151107',
+                'uploader_id': 'UCCr7TALkRbo3EtFzETQF1LA',
+                'uploader': 'CH GAMER DROID',
+            },
+            'params': {
+                'skip_download': True,
+            },
+        },
+        {
+            'url': 'https://www.youtube.com/watch?feature=player_embedded&v=V36LpHqtcDY',
+            'only_matching': True,
+        }
     ]
 
     def __init__(self, *args, **kwargs):
@@ -1458,6 +1480,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             manifest_url = video_info['hlsvp'][0]
             url_map = self._extract_from_m3u8(manifest_url, video_id)
             formats = _map_to_format_list(url_map)
+            # Accept-Encoding header causes failures in live streams on Youtube and Youtube Gaming
+            for a_format in formats:
+                a_format.setdefault('http_headers', {})['Youtubedl-no-compression'] = 'True'
         else:
             raise ExtractorError('no conn, hlsvp or url_encoded_fmt_stream_map information found in video info')
 
@@ -1495,10 +1520,15 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
             r'<meta\s+property="og:video:tag".*?content="yt:stretch=(?P<w>[0-9]+):(?P<h>[0-9]+)">',
             video_webpage)
         if stretched_m:
-            ratio = float(stretched_m.group('w')) / float(stretched_m.group('h'))
-            for f in formats:
-                if f.get('vcodec') != 'none':
-                    f['stretched_ratio'] = ratio
+            w = float(stretched_m.group('w'))
+            h = float(stretched_m.group('h'))
+            # yt:stretch may hold invalid ratio data (e.g. for Q39EVAstoRM ratio is 17:0).
+            # We will only process correct ratios.
+            if w > 0 and h > 0:
+                ratio = w / h
+                for f in formats:
+                    if f.get('vcodec') != 'none':
+                        f['stretched_ratio'] = ratio
 
         self._sort_formats(formats)
 
@@ -1537,7 +1567,7 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor, YoutubePlaylistBaseInfoExtract
                         youtube\.com/
                         (?:
                            (?:course|view_play_list|my_playlists|artist|playlist|watch|embed/videoseries)
-                           \? (?:.*?&)*? (?:p|a|list)=
+                           \? (?:.*?[&;])*? (?:p|a|list)=
                         |  p/
                         )
                         (