debian/control: Remove trailing whitespace at EOF.

[youtubedl] / youtube_dl / extractor / discovery.py

diff --git a/youtube_dl/extractor/discovery.py b/youtube_dl/extractor/discovery.py
index d3e6675283cddcb8f6a6dfffbfbd1e1ea3da11bc..3589bd42831515d5d4b16bfd64dbb861830b5173 100644 (file)
--- a/youtube_dl/extractor/discovery.py
+++ b/youtube_dl/extractor/discovery.py
@@ -1,45 +1,102 @@
 from __future__ import unicode_literals
 
 from __future__ import unicode_literals
 

-from .common import InfoExtractor
+import random
+import re
+import string
+
+from .discoverygo import DiscoveryGoBaseIE
+from ..compat import (
+    compat_str,
+    compat_urllib_parse_unquote,
+)

 from ..utils import (
 from ..utils import (

-    parse_iso8601,
-    int_or_none,
+    ExtractorError,
+    try_get,

 )
 )

+from ..compat import compat_HTTPError

 
 
 
 

-class DiscoveryIE(InfoExtractor):
-    _VALID_URL = r'http://www\.discovery\.com\/[a-zA-Z0-9\-]*/[a-zA-Z0-9\-]*/videos/(?P<id>[a-zA-Z0-9_\-]*)(?:\.htm)?'
-    _TEST = {
-        'url': 'http://www.discovery.com/tv-shows/mythbusters/videos/mission-impossible-outtakes.htm',
-        'md5': '3c69d77d9b0d82bfd5e5932a60f26504',
+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': {
         'info_dict': {

-            'id': 'mission-impossible-outtakes',
-            'ext': 'flv',
-            'title': '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,
-            'timestamp': 1303099200,
-            'upload_date': '20110418',
+            '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):
 
     def _real_extract(self, url):

-        video_id = self._match_id(url)
-        webpage = self._download_webpage(url, video_id)
-
-        info = self._parse_json(self._search_regex(
-            r'(?s)<script type="application/ld\+json">(.*?)</script>',
-            webpage, 'video info'), video_id)
-
-        return {
-            'id': video_id,
-            'title': info['name'],
-            'url': info['contentURL'],
-            'description': info.get('description'),
-            'thumbnail': info.get('thumbnailUrl'),
-            'timestamp': parse_iso8601(info.get('uploadDate')),
-            'duration': int_or_none(info.get('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)