[cloudflarestream] Add support for cloudflare streams (closes #16375)
[youtube-dl] / youtube_dl / extractor / cloudflarestream.py
diff --git a/youtube_dl/extractor/cloudflarestream.py b/youtube_dl/extractor/cloudflarestream.py
new file mode 100644 (file)
index 0000000..e6d92cc
--- /dev/null
@@ -0,0 +1,60 @@
+# coding: utf-8
+from __future__ import unicode_literals
+
+import re
+
+from .common import InfoExtractor
+
+
+class CloudflareStreamIE(InfoExtractor):
+    _VALID_URL = r'''(?x)
+                    https?://
+                        (?:
+                            (?:watch\.)?cloudflarestream\.com/|
+                            embed\.cloudflarestream\.com/embed/[^/]+\.js\?.*?\bvideo=
+                        )
+                        (?P<id>[\da-f]+)
+                    '''
+    _TESTS = [{
+        'url': 'https://embed.cloudflarestream.com/embed/we4g.fla9.latest.js?video=31c9291ab41fac05471db4e73aa11717',
+        'info_dict': {
+            'id': '31c9291ab41fac05471db4e73aa11717',
+            'ext': 'mp4',
+            'title': '31c9291ab41fac05471db4e73aa11717',
+        },
+        'params': {
+            'skip_download': True,
+        },
+    }, {
+        'url': 'https://watch.cloudflarestream.com/9df17203414fd1db3e3ed74abbe936c1',
+        'only_matching': True,
+    }, {
+        'url': 'https://cloudflarestream.com/31c9291ab41fac05471db4e73aa11717/manifest/video.mpd',
+        'only_matching': True,
+    }]
+
+    @staticmethod
+    def _extract_urls(webpage):
+        return [
+            mobj.group('url')
+            for mobj in re.finditer(
+                r'<script[^>]+\bsrc=(["\'])(?P<url>(?:https?:)?//embed\.cloudflarestream\.com/embed/[^/]+\.js\?.*?\bvideo=[\da-f]+?.*?)\1',
+                webpage)]
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
+
+        formats = self._extract_m3u8_formats(
+            'https://cloudflarestream.com/%s/manifest/video.m3u8' % video_id,
+            video_id, 'mp4', entry_protocol='m3u8_native', m3u8_id='hls',
+            fatal=False)
+        formats.extend(self._extract_mpd_formats(
+            'https://cloudflarestream.com/%s/manifest/video.mpd' % video_id,
+            video_id, mpd_id='dash', fatal=False))
+        self._sort_formats(formats)
+
+        return {
+            'id': video_id,
+            'title': video_id,
+            'formats': formats,
+        }