Merge pull request #8092 from bpfoley/twitter-thumbnail

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

diff --git a/youtube_dl/extractor/twitter.py b/youtube_dl/extractor/twitter.py
index 5cbb536f97e3c3f2d1b771996b40d71dce32dacd..e70b2ab3c8d564cd907e8763fd8136e1b3827ac5 100644 (file)
--- a/youtube_dl/extractor/twitter.py
+++ b/youtube_dl/extractor/twitter.py
@@ -10,7 +10,6 @@ from ..utils import (
     remove_end,
     int_or_none,
     ExtractorError,
-    sanitized_Request,
 )
 
 
@@ -87,55 +86,83 @@ class TwitterCardIE(TwitterBaseIE):
     def _real_extract(self, url):
         video_id = self._match_id(url)
 
-        # Different formats served for different User-Agents
-        USER_AGENTS = [
-            'Mozilla/5.0 (X11; Linux x86_64; rv:10.0) Gecko/20150101 Firefox/20.0 (Chrome)',  # mp4
-            'Mozilla/5.0 (Windows NT 5.2; WOW64; rv:38.0) Gecko/20100101 Firefox/38.0',  # webm
-        ]
-
         config = None
         formats = []
-        for user_agent in USER_AGENTS:
-            request = sanitized_Request(url)
-            request.add_header('User-Agent', user_agent)
-            webpage = self._download_webpage(request, video_id)
-
-            iframe_url = self._html_search_regex(
-                r'<iframe[^>]+src="((?:https?:)?//(?:www.youtube.com/embed/[^"]+|(?:www\.)?vine\.co/v/\w+/card))"',
-                webpage, 'video iframe', default=None)
-            if iframe_url:
-                return self.url_result(iframe_url)
-
-            config = self._parse_json(self._html_search_regex(
-                r'data-(?:player-)?config="([^"]+)"', webpage, 'data player config'),
-                video_id)
-            if 'playlist' not in config:
-                vmap_url = config.get('vmapUrl') or config.get('vmap_url')
-                if vmap_url:
-                    formats.append({
-                        'url': self._get_vmap_video_url(vmap_url, video_id),
-                    })
-                    break   # same video regardless of UA
-                continue
-
-            video_url = config['playlist'][0]['source']
+        duration = None
 
-            f = {
-                'url': video_url,
-            }
+        webpage = self._download_webpage(url, video_id)
+
+        iframe_url = self._html_search_regex(
+            r'<iframe[^>]+src="((?:https?:)?//(?:www.youtube.com/embed/[^"]+|(?:www\.)?vine\.co/v/\w+/card))"',
+            webpage, 'video iframe', default=None)
+        if iframe_url:
+            return self.url_result(iframe_url)
 
+        config = self._parse_json(self._html_search_regex(
+            r'data-(?:player-)?config="([^"]+)"', webpage, 'data player config'),
+            video_id)
+
+        def _search_dimensions_in_video_url(a_format, video_url):
             m = re.search(r'/(?P<width>\d+)x(?P<height>\d+)/', video_url)
             if m:
-                f.update({
+                a_format.update({
                     'width': int(m.group('width')),
                     'height': int(m.group('height')),
                 })
+
+        playlist = config.get('playlist')
+        if playlist:
+            video_url = playlist[0]['source']
+
+            f = {
+                'url': video_url,
+            }
+
+            _search_dimensions_in_video_url(f, video_url)
+
             formats.append(f)
+
+        vmap_url = config.get('vmapUrl') or config.get('vmap_url')
+        if vmap_url:
+            formats.append({
+                'url': self._get_vmap_video_url(vmap_url, video_id),
+            })
+
+        media_info = None
+
+        for entity in config.get('status', {}).get('entities', []):
+            if 'mediaInfo' in entity:
+                media_info = entity['mediaInfo']
+
+        if media_info:
+            for media_variant in media_info['variants']:
+                media_url = media_variant['url']
+                if media_url.endswith('.m3u8'):
+                    formats.extend(self._extract_m3u8_formats(media_url, video_id, ext='mp4', m3u8_id='hls'))
+                elif media_url.endswith('.mpd'):
+                    formats.extend(self._extract_mpd_formats(media_url, video_id, mpd_id='dash'))
+                else:
+                    vbr = int_or_none(media_variant.get('bitRate'), scale=1000)
+                    a_format = {
+                        'url': media_url,
+                        'format_id': 'http-%d' % vbr if vbr else 'http',
+                        'vbr': vbr,
+                    }
+                    # Reported bitRate may be zero
+                    if not a_format['vbr']:
+                        del a_format['vbr']
+
+                    _search_dimensions_in_video_url(a_format, media_url)
+
+                    formats.append(a_format)
+
+            duration = float_or_none(media_info.get('duration', {}).get('nanos'), scale=1e9)
+
         self._sort_formats(formats)
 
         title = self._search_regex(r'<title>([^<]+)</title>', webpage, 'title')
         thumbnail = config.get('posterImageUrl') or config.get('image_src')
-        duration = float_or_none(config.get('duration'))
+        duration = float_or_none(config.get('duration')) or duration
 
         return {
             'id': video_id,
@@ -153,7 +180,6 @@ class TwitterIE(InfoExtractor):
 
     _TESTS = [{
         'url': 'https://twitter.com/freethenipple/status/643211948184596480',
-        # MD5 checksums are different in different places
         'info_dict': {
             'id': '643211948184596480',
             'ext': 'mp4',
@@ -164,6 +190,9 @@ class TwitterIE(InfoExtractor):
             'uploader': 'FREE THE NIPPLE',
             'uploader_id': 'freethenipple',
         },
+        'params': {
+            'skip_download': True,  # requires ffmpeg
+        },
     }, {
         'url': 'https://twitter.com/giphz/status/657991469417025536/photo/1',
         'md5': 'f36dcd5fb92bf7057f155e7d927eeb42',
@@ -203,6 +232,21 @@ class TwitterIE(InfoExtractor):
             # Test case of TwitterCardIE
             'skip_download': True,
         },
+    }, {
+        'url': 'https://twitter.com/jaydingeer/status/700207533655363584',
+        'md5': '',
+        'info_dict': {
+            'id': '700207533655363584',
+            'ext': 'mp4',
+            'title': 'jay - BEAT PROD: @suhmeduh #Damndaniel',
+            'description': 'jay on Twitter: "BEAT PROD: @suhmeduh  https://t.co/HBrQ4AfpvZ #Damndaniel https://t.co/byBooq2ejZ"',
+            'thumbnail': 're:^https?://.*\.jpg',
+            'uploader': 'jay',
+            'uploader_id': 'jaydingeer',
+        },
+        'params': {
+            'skip_download': True,  # requires ffmpeg
+        },
     }]
 
     def _real_extract(self, url):