about summary refs log tree commit diff
path: root/youtube_dl/extractor/twitter.py
diff options
context:
space:
mode:
authorSergey M․ <dstftw@gmail.com>2015-06-28 01:22:25 +0600
committerSergey M․ <dstftw@gmail.com>2015-06-28 01:22:25 +0600
commit23e7cba87fbcec9aa2b1fbccf60c2d560df3d7ad (patch)
tree38dadf6a598bdec2ce99039c3f926a5990bb08dc /youtube_dl/extractor/twitter.py
parent6c4d20cd6fae8e47c1f0c6771e2e91660bf10eb5 (diff)
downloadyoutube-dl-23e7cba87fbcec9aa2b1fbccf60c2d560df3d7ad.tar.gz
youtube-dl-23e7cba87fbcec9aa2b1fbccf60c2d560df3d7ad.tar.xz
youtube-dl-23e7cba87fbcec9aa2b1fbccf60c2d560df3d7ad.zip
[twitter:card] Add extractor (#5239)
Diffstat (limited to 'youtube_dl/extractor/twitter.py')
-rw-r--r--youtube_dl/extractor/twitter.py72
1 files changed, 72 insertions, 0 deletions
diff --git a/youtube_dl/extractor/twitter.py b/youtube_dl/extractor/twitter.py
new file mode 100644
index 000000000..1aaa06305
--- /dev/null
+++ b/youtube_dl/extractor/twitter.py
@@ -0,0 +1,72 @@
+from __future__ import unicode_literals
+
+import re
+
+from .common import InfoExtractor
+from ..compat import compat_urllib_request
+from ..utils import (
+    float_or_none,
+    unescapeHTML,
+)
+
+
+class TwitterCardIE(InfoExtractor):
+    _VALID_URL = r'https?://(?:www\.)?twitter\.com/i/cards/tfw/v1/(?P<id>\d+)'
+    _TEST = {
+        'url': 'https://twitter.com/i/cards/tfw/v1/560070183650213889',
+        'md5': 'a74f50b310c83170319ba16de6955192',
+        'info_dict': {
+            'id': '560070183650213889',
+            'ext': 'mp4',
+            'title': 'TwitterCard',
+            'thumbnail': 're:^https?://.*\.jpg$',
+            'duration': 30.033,
+        },
+    }
+
+    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 = compat_urllib_request.Request(url)
+            request.add_header('User-Agent', user_agent)
+            webpage = self._download_webpage(request, video_id)
+
+            config = self._parse_json(
+                unescapeHTML(self._search_regex(
+                    r'data-player-config="([^"]+)"', webpage, 'data player config')),
+                video_id)
+
+            video_url = config['playlist'][0]['source']
+
+            f = {
+                'url': video_url,
+            }
+
+            m = re.search(r'/(?P<width>\d+)x(?P<height>\d+)/', video_url)
+            if m:
+                f.update({
+                    'width': int(m.group('width')),
+                    'height': int(m.group('height')),
+                })
+            formats.append(f)
+        self._sort_formats(formats)
+
+        thumbnail = config.get('posterImageUrl')
+        duration = float_or_none(config.get('duration'))
+
+        return {
+            'id': video_id,
+            'title': 'TwitterCard',
+            'thumbnail': thumbnail,
+            'duration': duration,
+            'formats': formats,
+        }