]> Raphaƫl G. Git Repositories - youtubedl/blobdiff - youtube_dl/extractor/discovery.py
debian/control: Mark compliance with Debian policy 4.2.1.
[youtubedl] / youtube_dl / extractor / discovery.py
index 885944c5e8bf95e7af78c5c3ed4eb5f69cc7a35f..3589bd42831515d5d4b16bfd64dbb861830b5173 100644 (file)
 from __future__ import unicode_literals
 
+import random
 import re
-import json
+import string
 
-from .common import InfoExtractor
+from .discoverygo import DiscoveryGoBaseIE
+from ..compat import (
+    compat_str,
+    compat_urllib_parse_unquote,
+)
+from ..utils import (
+    ExtractorError,
+    try_get,
+)
+from ..compat import compat_HTTPError
 
 
-class DiscoveryIE(InfoExtractor):
-    _VALID_URL = r'http://dsc\.discovery\.com\/[a-zA-Z0-9\-]*/[a-zA-Z0-9\-]*/videos/(?P<id>[a-zA-Z0-9\-]*)(.htm)?'
-    _TEST = {
-        'url': 'http://dsc.discovery.com/tv-shows/mythbusters/videos/mission-impossible-outtakes.htm',
-        'file': '614784.mp4',
-        'md5': 'e12614f9ee303a6ccef415cb0793eba2',
+class DiscoveryIE(DiscoveryGoBaseIE):
+    _VALID_URL = r'''(?x)https?://(?:www\.)?(?P<site>
+            discovery|
+            investigationdiscovery|
+            discoverylife|
+            animalplanet|
+            ahctv|
+            destinationamerica|
+            sciencechannel|
+            tlc|
+            velocity
+        )\.com(?P<path>/tv-shows/[^/]+/(?:video|full-episode)s/(?P<id>[^./?#]+))'''
+    _TESTS = [{
+        'url': 'https://www.discovery.com/tv-shows/cash-cab/videos/dave-foley',
         'info_dict': {
-            'title': 'MythBusters: Mission Impossible Outtakes',
-            'description': ('Watch Jamie Hyneman and Adam Savage practice being'
-                ' each other -- to the point of confusing Jamie\'s dog -- and '
-                'don\'t miss Adam moon-walking as Jamie ... behind Jamie\'s'
-                ' back.'),
-            'duration': 156,
+            'id': '5a2d9b4d6b66d17a5026e1fd',
+            'ext': 'mp4',
+            'title': 'Dave Foley',
+            'description': 'md5:4b39bcafccf9167ca42810eb5f28b01f',
+            'duration': 608,
         },
-    }
+        'params': {
+            'skip_download': True,  # requires ffmpeg
+        }
+    }, {
+        'url': 'https://www.investigationdiscovery.com/tv-shows/final-vision/full-episodes/final-vision',
+        'only_matching': True,
+    }]
+    _GEO_COUNTRIES = ['US']
+    _GEO_BYPASS = False
 
     def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
-        video_id = mobj.group('id')
-        webpage = self._download_webpage(url, video_id)
-
-        video_list_json = self._search_regex(r'var videoListJSON = ({.*?});',
-            webpage, 'video list', flags=re.DOTALL)
-        video_list = json.loads(video_list_json)
-        info = video_list['clips'][0]
-        formats = []
-        for f in info['mp4']:
-            formats.append(
-                {'url': f['src'], r'ext': r'mp4', 'tbr': int(f['bitrate'][:-1])})
-
-        return {
-            'id': info['contentId'],
-            'title': video_list['name'],
-            'formats': formats,
-            'description': info['videoCaption'],
-            'thumbnail': info.get('videoStillURL') or info.get('thumbnailURL'),
-            'duration': info['duration'],
-        }
+        site, path, display_id = re.match(self._VALID_URL, url).groups()
+        webpage = self._download_webpage(url, display_id)
+
+        react_data = self._parse_json(self._search_regex(
+            r'window\.__reactTransmitPacket\s*=\s*({.+?});',
+            webpage, 'react data'), display_id)
+        content_blocks = react_data['layout'][path]['contentBlocks']
+        video = next(cb for cb in content_blocks if cb.get('type') == 'video')['content']['items'][0]
+        video_id = video['id']
+
+        access_token = None
+        cookies = self._get_cookies(url)
+
+        # prefer Affiliate Auth Token over Anonymous Auth Token
+        auth_storage_cookie = cookies.get('eosAf') or cookies.get('eosAn')
+        if auth_storage_cookie and auth_storage_cookie.value:
+            auth_storage = self._parse_json(compat_urllib_parse_unquote(
+                compat_urllib_parse_unquote(auth_storage_cookie.value)),
+                video_id, fatal=False) or {}
+            access_token = auth_storage.get('a') or auth_storage.get('access_token')
+
+        if not access_token:
+            access_token = self._download_json(
+                'https://www.%s.com/anonymous' % site, display_id, query={
+                    'authRel': 'authorization',
+                    'client_id': try_get(
+                        react_data, lambda x: x['application']['apiClientId'],
+                        compat_str) or '3020a40c2356a645b4b4',
+                    'nonce': ''.join([random.choice(string.ascii_letters) for _ in range(32)]),
+                    'redirectUri': 'https://fusion.ddmcdn.com/app/mercury-sdk/180/redirectHandler.html?https://www.%s.com' % site,
+                })['access_token']
+
+        try:
+            stream = self._download_json(
+                'https://api.discovery.com/v1/streaming/video/' + video_id,
+                display_id, headers={
+                    'Authorization': 'Bearer ' + access_token,
+                })
+        except ExtractorError as e:
+            if isinstance(e.cause, compat_HTTPError) and e.cause.code in (401, 403):
+                e_description = self._parse_json(
+                    e.cause.read().decode(), display_id)['description']
+                if 'resource not available for country' in e_description:
+                    self.raise_geo_restricted(countries=self._GEO_COUNTRIES)
+                if 'Authorized Networks' in e_description:
+                    raise ExtractorError(
+                        'This video is only available via cable service provider subscription that'
+                        ' is not currently supported. You may want to use --cookies.', expected=True)
+                raise ExtractorError(e_description)
+            raise
+
+        return self._extract_video_info(video, stream, display_id)