]> Raphaël G. Git Repositories - youtubedl/blobdiff - youtube_dl/extractor/condenast.py
Update upstream source from tag 'upstream/2018.06.18'
[youtubedl] / youtube_dl / extractor / condenast.py
index f336a3c620a04e8bb643309b4812725e8f50e1d1..ed278fefc67089f44760ed8e0635cf1c5cbb1fde 100644 (file)
@@ -1,15 +1,22 @@
 # coding: utf-8
+from __future__ import unicode_literals
 
 import re
-import json
 
 from .common import InfoExtractor
-from ..utils import (
-    compat_urllib_parse,
-    orderedSet,
+from ..compat import (
     compat_urllib_parse_urlparse,
     compat_urlparse,
 )
+from ..utils import (
+    determine_ext,
+    extract_attributes,
+    int_or_none,
+    js_to_json,
+    mimetype2ext,
+    orderedSet,
+    parse_iso8601,
+)
 
 
 class CondeNastIE(InfoExtractor):
@@ -20,87 +27,206 @@ class CondeNastIE(InfoExtractor):
 
     # The keys are the supported sites and the values are the name to be shown
     # to the user and in the extractor description.
-    _SITES = {'wired': u'WIRED',
-              'gq': u'GQ',
-              'vogue': u'Vogue',
-              'glamour': u'Glamour',
-              'wmagazine': u'W Magazine',
-              'vanityfair': u'Vanity Fair',
-              }
-
-    _VALID_URL = r'http://(video|www).(?P<site>%s).com/(?P<type>watch|series|video)/(?P<id>.+)' % '|'.join(_SITES.keys())
-    IE_DESC = u'Condé Nast media group: %s' % ', '.join(sorted(_SITES.values()))
-
-    _TEST = {
-        u'url': u'http://video.wired.com/watch/3d-printed-speakers-lit-with-led',
-        u'file': u'5171b343c2b4c00dd0c1ccb3.mp4',
-        u'md5': u'1921f713ed48aabd715691f774c451f7',
-        u'info_dict': {
-            u'title': u'3D Printed Speakers Lit With LED',
-            u'description': u'Check out these beautiful 3D printed LED speakers.  You can\'t actually buy them, but LumiGeek is working on a board that will let you make you\'re own.',
-        }
+    _SITES = {
+        'allure': 'Allure',
+        'architecturaldigest': 'Architectural Digest',
+        'arstechnica': 'Ars Technica',
+        'bonappetit': 'Bon Appétit',
+        'brides': 'Brides',
+        'cnevids': 'Condé Nast',
+        'cntraveler': 'Condé Nast Traveler',
+        'details': 'Details',
+        'epicurious': 'Epicurious',
+        'glamour': 'Glamour',
+        'golfdigest': 'Golf Digest',
+        'gq': 'GQ',
+        'newyorker': 'The New Yorker',
+        'self': 'SELF',
+        'teenvogue': 'Teen Vogue',
+        'vanityfair': 'Vanity Fair',
+        'vogue': 'Vogue',
+        'wired': 'WIRED',
+        'wmagazine': 'W Magazine',
     }
 
+    _VALID_URL = r'''(?x)https?://(?:video|www|player(?:-backend)?)\.(?:%s)\.com/
+        (?:
+            (?:
+                embed(?:js)?|
+                (?:script|inline)/video
+            )/(?P<id>[0-9a-f]{24})(?:/(?P<player_id>[0-9a-f]{24}))?(?:.+?\btarget=(?P<target>[^&]+))?|
+            (?P<type>watch|series|video)/(?P<display_id>[^/?#]+)
+        )''' % '|'.join(_SITES.keys())
+    IE_DESC = 'Condé Nast media group: %s' % ', '.join(sorted(_SITES.values()))
+
+    EMBED_URL = r'(?:https?:)?//player(?:-backend)?\.(?:%s)\.com/(?:embed(?:js)?|(?:script|inline)/video)/.+?' % '|'.join(_SITES.keys())
+
+    _TESTS = [{
+        'url': 'http://video.wired.com/watch/3d-printed-speakers-lit-with-led',
+        'md5': '1921f713ed48aabd715691f774c451f7',
+        'info_dict': {
+            'id': '5171b343c2b4c00dd0c1ccb3',
+            'ext': 'mp4',
+            'title': '3D Printed Speakers Lit With LED',
+            'description': 'Check out these beautiful 3D printed LED speakers.  You can\'t actually buy them, but LumiGeek is working on a board that will let you make you\'re own.',
+            'uploader': 'wired',
+            'upload_date': '20130314',
+            'timestamp': 1363219200,
+        }
+    }, {
+        'url': 'http://video.gq.com/watch/the-closer-with-keith-olbermann-the-only-true-surprise-trump-s-an-idiot?c=series',
+        'info_dict': {
+            'id': '58d1865bfd2e6126e2000015',
+            'ext': 'mp4',
+            'title': 'The Only True Surprise? Trump’s an Idiot',
+            'uploader': 'gq',
+            'upload_date': '20170321',
+            'timestamp': 1490126427,
+        },
+    }, {
+        # JS embed
+        'url': 'http://player.cnevids.com/embedjs/55f9cf8b61646d1acf00000c/5511d76261646d5566020000.js',
+        'md5': 'f1a6f9cafb7083bab74a710f65d08999',
+        'info_dict': {
+            'id': '55f9cf8b61646d1acf00000c',
+            'ext': 'mp4',
+            'title': '3D printed TSA Travel Sentry keys really do open TSA locks',
+            'uploader': 'arstechnica',
+            'upload_date': '20150916',
+            'timestamp': 1442434955,
+        }
+    }, {
+        'url': 'https://player.cnevids.com/inline/video/59138decb57ac36b83000005.js?target=js-cne-player',
+        'only_matching': True,
+    }, {
+        'url': 'http://player-backend.cnevids.com/script/video/59138decb57ac36b83000005.js',
+        'only_matching': True,
+    }]
+
     def _extract_series(self, url, webpage):
-        title = self._html_search_regex(r'<div class="cne-series-info">.*?<h1>(.+?)</h1>',
-                                        webpage, u'series title', flags=re.DOTALL)
+        title = self._html_search_regex(
+            r'(?s)<div class="cne-series-info">.*?<h1>(.+?)</h1>',
+            webpage, 'series title')
         url_object = compat_urllib_parse_urlparse(url)
         base_url = '%s://%s' % (url_object.scheme, url_object.netloc)
-        m_paths = re.finditer(r'<p class="cne-thumb-title">.*?<a href="(/watch/.+?)["\?]',
-                              webpage, flags=re.DOTALL)
+        m_paths = re.finditer(
+            r'(?s)<p class="cne-thumb-title">.*?<a href="(/watch/.+?)["\?]', webpage)
         paths = orderedSet(m.group(1) for m in m_paths)
         build_url = lambda path: compat_urlparse.urljoin(base_url, path)
         entries = [self.url_result(build_url(path), 'CondeNast') for path in paths]
         return self.playlist_result(entries, playlist_title=title)
 
-    def _extract_video(self, webpage):
-        description = self._html_search_regex([r'<div class="cne-video-description">(.+?)</div>',
-                                               r'<div class="video-post-content">(.+?)</div>',
-                                               ],
-                                              webpage, u'description',
-                                              fatal=False, flags=re.DOTALL)
-        params = self._search_regex(r'var params = {(.+?)}[;,]', webpage,
-                                    u'player params', flags=re.DOTALL)
-        video_id = self._search_regex(r'videoId: [\'"](.+?)[\'"]', params, u'video id')
-        player_id = self._search_regex(r'playerId: [\'"](.+?)[\'"]', params, u'player id')
-        target = self._search_regex(r'target: [\'"](.+?)[\'"]', params, u'target')
-        data = compat_urllib_parse.urlencode({'videoId': video_id,
-                                              'playerId': player_id,
-                                              'target': target,
-                                              })
-        base_info_url = self._search_regex(r'url = [\'"](.+?)[\'"][,;]',
-                                           webpage, u'base info url',
-                                           default='http://player.cnevids.com/player/loader.js?')
-        info_url = base_info_url + data
-        info_page = self._download_webpage(info_url, video_id,
-                                           u'Downloading video info')
-        video_info = self._search_regex(r'var video = ({.+?});', info_page, u'video info')
-        video_info = json.loads(video_info)
-
-        def _formats_sort_key(f):
-            type_ord = 1 if f['type'] == 'video/mp4' else 0
-            quality_ord = 1 if f['quality'] == 'high' else 0
-            return (quality_ord, type_ord)
-        best_format = sorted(video_info['sources'][0], key=_formats_sort_key)[-1]
-
-        return {'id': video_id,
-                'url': best_format['src'],
-                'ext': best_format['type'].split('/')[-1],
-                'title': video_info['title'],
-                'thumbnail': video_info['poster_frame'],
-                'description': description,
-                }
+    def _extract_video_params(self, webpage, display_id):
+        query = self._parse_json(
+            self._search_regex(
+                r'(?s)var\s+params\s*=\s*({.+?})[;,]', webpage, 'player params',
+                default='{}'),
+            display_id, transform_source=js_to_json, fatal=False)
+        if query:
+            query['videoId'] = self._search_regex(
+                r'(?:data-video-id=|currentVideoId\s*=\s*)["\']([\da-f]+)',
+                webpage, 'video id', default=None)
+        else:
+            params = extract_attributes(self._search_regex(
+                r'(<[^>]+data-js="video-player"[^>]+>)',
+                webpage, 'player params element'))
+            query.update({
+                'videoId': params['data-video'],
+                'playerId': params['data-player'],
+                'target': params['id'],
+            })
+        return query
+
+    def _extract_video(self, params):
+        video_id = params['videoId']
+
+        video_info = None
+
+        # New API path
+        query = params.copy()
+        query['embedType'] = 'inline'
+        info_page = self._download_json(
+            'http://player.cnevids.com/embed-api.json', video_id,
+            'Downloading embed info', fatal=False, query=query)
+
+        # Old fallbacks
+        if not info_page:
+            if params.get('playerId'):
+                info_page = self._download_json(
+                    'http://player.cnevids.com/player/video.js', video_id,
+                    'Downloading video info', fatal=False, query=params)
+        if info_page:
+            video_info = info_page.get('video')
+        if not video_info:
+            info_page = self._download_webpage(
+                'http://player.cnevids.com/player/loader.js',
+                video_id, 'Downloading loader info', query=params)
+        if not video_info:
+            info_page = self._download_webpage(
+                'https://player.cnevids.com/inline/video/%s.js' % video_id,
+                video_id, 'Downloading inline info', query={
+                    'target': params.get('target', 'embedplayer')
+                })
+
+        if not video_info:
+            video_info = self._parse_json(
+                self._search_regex(
+                    r'(?s)var\s+config\s*=\s*({.+?});', info_page, 'config'),
+                video_id, transform_source=js_to_json)['video']
+
+        title = video_info['title']
+
+        formats = []
+        for fdata in video_info['sources']:
+            src = fdata.get('src')
+            if not src:
+                continue
+            ext = mimetype2ext(fdata.get('type')) or determine_ext(src)
+            if ext == 'm3u8':
+                formats.extend(self._extract_m3u8_formats(
+                    src, video_id, 'mp4', entry_protocol='m3u8_native',
+                    m3u8_id='hls', fatal=False))
+                continue
+            quality = fdata.get('quality')
+            formats.append({
+                'format_id': ext + ('-%s' % quality if quality else ''),
+                'url': src,
+                'ext': ext,
+                'quality': 1 if quality == 'high' else 0,
+            })
+        self._sort_formats(formats)
+
+        return {
+            'id': video_id,
+            'formats': formats,
+            'title': title,
+            'thumbnail': video_info.get('poster_frame'),
+            'uploader': video_info.get('brand'),
+            'duration': int_or_none(video_info.get('duration')),
+            'tags': video_info.get('tags'),
+            'series': video_info.get('series_title'),
+            'season': video_info.get('season_title'),
+            'timestamp': parse_iso8601(video_info.get('premiere_date')),
+            'categories': video_info.get('categories'),
+        }
 
     def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
-        site = mobj.group('site')
-        url_type = mobj.group('type')
-        id = mobj.group('id')
+        video_id, player_id, target, url_type, display_id = re.match(self._VALID_URL, url).groups()
+
+        if video_id:
+            return self._extract_video({
+                'videoId': video_id,
+                'playerId': player_id,
+                'target': target,
+            })
 
-        self.to_screen(u'Extracting from %s with the Condé Nast extractor' % self._SITES[site])
-        webpage = self._download_webpage(url, id)
+        webpage = self._download_webpage(url, display_id)
 
         if url_type == 'series':
             return self._extract_series(url, webpage)
         else:
-            return self._extract_video(webpage)
+            params = self._extract_video_params(webpage, display_id)
+            info = self._search_json_ld(
+                webpage, display_id, fatal=False)
+            info.update(self._extract_video(params))
+            return info