[srgssr] improve extraction
authorAlexander Seiler <seileralex@gmail.com>
Sat, 11 Nov 2017 18:30:10 +0000 (19:30 +0100)
committerRemita Amine <remitamine@gmail.com>
Thu, 25 Feb 2021 14:50:49 +0000 (15:50 +0100)
- extract subtitle
- fix extraction for new videos
- update srf download domains

closes #14717
closes #14725
closes #27231
closes #28238

youtube_dl/extractor/rts.py
youtube_dl/extractor/srgssr.py

index 48f17b828c1a8c6b0494ad1edb2bc3f914b3417a..aed35f8a9802b1d98664ddf6ada0499059b5625f 100644 (file)
@@ -6,11 +6,12 @@ import re
 from .srgssr import SRGSSRIE
 from ..compat import compat_str
 from ..utils import (
+    determine_ext,
     int_or_none,
     parse_duration,
     parse_iso8601,
     unescapeHTML,
-    determine_ext,
+    urljoin,
 )
 
 
@@ -21,7 +22,7 @@ class RTSIE(SRGSSRIE):
     _TESTS = [
         {
             'url': 'http://www.rts.ch/archives/tv/divers/3449373-les-enfants-terribles.html',
-            'md5': 'ff7f8450a90cf58dacb64e29707b4a8e',
+            'md5': '753b877968ad8afaeddccc374d4256a5',
             'info_dict': {
                 'id': '3449373',
                 'display_id': 'les-enfants-terribles',
@@ -35,6 +36,7 @@ class RTSIE(SRGSSRIE):
                 'thumbnail': r're:^https?://.*\.image',
                 'view_count': int,
             },
+            'expected_warnings': ['Unable to download f4m manifest', 'Failed to download m3u8 information'],
         },
         {
             'url': 'http://www.rts.ch/emissions/passe-moi-les-jumelles/5624067-entre-ciel-et-mer.html',
@@ -63,11 +65,12 @@ class RTSIE(SRGSSRIE):
                 # m3u8 download
                 'skip_download': True,
             },
+            'expected_warnings': ['Unable to download f4m manifest', 'Failed to download m3u8 information'],
             'skip': 'Blocked outside Switzerland',
         },
         {
             'url': 'http://www.rts.ch/video/info/journal-continu/5745356-londres-cachee-par-un-epais-smog.html',
-            'md5': '1bae984fe7b1f78e94abc74e802ed99f',
+            'md5': '9bb06503773c07ce83d3cbd793cebb91',
             'info_dict': {
                 'id': '5745356',
                 'display_id': 'londres-cachee-par-un-epais-smog',
@@ -81,6 +84,7 @@ class RTSIE(SRGSSRIE):
                 'thumbnail': r're:^https?://.*\.image',
                 'view_count': int,
             },
+            'expected_warnings': ['Unable to download f4m manifest', 'Failed to download m3u8 information'],
         },
         {
             'url': 'http://www.rts.ch/audio/couleur3/programmes/la-belle-video-de-stephane-laurenceau/5706148-urban-hippie-de-damien-krisl-03-04-2014.html',
@@ -160,7 +164,7 @@ class RTSIE(SRGSSRIE):
         media_type = 'video' if 'video' in all_info else 'audio'
 
         # check for errors
-        self.get_media_data('rts', media_type, media_id)
+        self._get_media_data('rts', media_type, media_id)
 
         info = all_info['video']['JSONinfo'] if 'video' in all_info else all_info['audio']
 
@@ -194,6 +198,7 @@ class RTSIE(SRGSSRIE):
                     'tbr': extract_bitrate(format_url),
                 })
 
+        download_base = 'http://rtsww%s-d.rts.ch/' % ('-a' if media_type == 'audio' else '')
         for media in info.get('media', []):
             media_url = media.get('url')
             if not media_url or re.match(r'https?://', media_url):
@@ -205,7 +210,7 @@ class RTSIE(SRGSSRIE):
                 format_id += '-%dk' % rate
             formats.append({
                 'format_id': format_id,
-                'url': 'http://download-video.rts.ch/' + media_url,
+                'url': urljoin(download_base, media_url),
                 'tbr': rate or extract_bitrate(media_url),
             })
 
index f63a1359aed6620af8c42cd9d4652f5ec379c7c0..ac018e7405f966aa19e6962715bf6ddaa1feced8 100644 (file)
@@ -4,16 +4,32 @@ from __future__ import unicode_literals
 import re
 
 from .common import InfoExtractor
-from ..compat import compat_urllib_parse_urlparse
 from ..utils import (
     ExtractorError,
+    float_or_none,
+    int_or_none,
     parse_iso8601,
     qualities,
+    try_get,
 )
 
 
 class SRGSSRIE(InfoExtractor):
-    _VALID_URL = r'(?:https?://tp\.srgssr\.ch/p(?:/[^/]+)+\?urn=urn|srgssr):(?P<bu>srf|rts|rsi|rtr|swi):(?:[^:]+:)?(?P<type>video|audio):(?P<id>[0-9a-f\-]{36}|\d+)'
+    _VALID_URL = r'''(?x)
+                    (?:
+                        https?://tp\.srgssr\.ch/p(?:/[^/]+)+\?urn=urn|
+                        srgssr
+                    ):
+                    (?P<bu>
+                        srf|rts|rsi|rtr|swi
+                    ):(?:[^:]+:)?
+                    (?P<type>
+                        video|audio
+                    ):
+                    (?P<id>
+                        [0-9a-f\-]{36}|\d+
+                    )
+                    '''
     _GEO_BYPASS = False
     _GEO_COUNTRIES = ['CH']
 
@@ -25,25 +41,39 @@ class SRGSSRIE(InfoExtractor):
         'LEGAL': 'The video cannot be transmitted for legal reasons.',
         'STARTDATE': 'This video is not yet available. Please try again later.',
     }
+    _DEFAULT_LANGUAGE_CODES = {
+        'srf': 'de',
+        'rts': 'fr',
+        'rsi': 'it',
+        'rtr': 'rm',
+        'swi': 'en',
+    }
 
     def _get_tokenized_src(self, url, video_id, format_id):
-        sp = compat_urllib_parse_urlparse(url).path.split('/')
         token = self._download_json(
-            'http://tp.srgssr.ch/akahd/token?acl=/%s/%s/*' % (sp[1], sp[2]),
+            'http://tp.srgssr.ch/akahd/token?acl=*',
             video_id, 'Downloading %s token' % format_id, fatal=False) or {}
-        auth_params = token.get('token', {}).get('authparams')
+        auth_params = try_get(token, lambda x: x['token']['authparams'])
         if auth_params:
-            url += '?' + auth_params
+            url += ('?' if '?' not in url else '&') + auth_params
         return url
 
-    def get_media_data(self, bu, media_type, media_id):
-        media_data = self._download_json(
-            'http://il.srgssr.ch/integrationlayer/1.0/ue/%s/%s/play/%s.json' % (bu, media_type, media_id),
-            media_id)[media_type.capitalize()]
-
-        if media_data.get('block') and media_data['block'] in self._ERRORS:
-            message = self._ERRORS[media_data['block']]
-            if media_data['block'] == 'GEOBLOCK':
+    def _get_media_data(self, bu, media_type, media_id):
+        query = {'onlyChapters': True} if media_type == 'video' else {}
+        full_media_data = self._download_json(
+            'https://il.srgssr.ch/integrationlayer/2.0/%s/mediaComposition/%s/%s.json'
+            % (bu, media_type, media_id),
+            media_id, query=query)['chapterList']
+        try:
+            media_data = next(
+                x for x in full_media_data if x.get('id') == media_id)
+        except StopIteration:
+            raise ExtractorError('No media information found')
+
+        block_reason = media_data.get('blockReason')
+        if block_reason and block_reason in self._ERRORS:
+            message = self._ERRORS[block_reason]
+            if block_reason == 'GEOBLOCK':
                 self.raise_geo_restricted(
                     msg=message, countries=self._GEO_COUNTRIES)
             raise ExtractorError(
@@ -53,53 +83,75 @@ class SRGSSRIE(InfoExtractor):
 
     def _real_extract(self, url):
         bu, media_type, media_id = re.match(self._VALID_URL, url).groups()
+        media_data = self._get_media_data(bu, media_type, media_id)
+        title = media_data['title']
 
-        media_data = self.get_media_data(bu, media_type, media_id)
-
-        metadata = media_data['AssetMetadatas']['AssetMetadata'][0]
-        title = metadata['title']
-        description = metadata.get('description')
-        created_date = media_data.get('createdDate') or metadata.get('createdDate')
-        timestamp = parse_iso8601(created_date)
-
-        thumbnails = [{
-            'id': image.get('id'),
-            'url': image['url'],
-        } for image in media_data.get('Image', {}).get('ImageRepresentations', {}).get('ImageRepresentation', [])]
-
-        preference = qualities(['LQ', 'MQ', 'SD', 'HQ', 'HD'])
         formats = []
-        for source in media_data.get('Playlists', {}).get('Playlist', []) + media_data.get('Downloads', {}).get('Download', []):
-            protocol = source.get('@protocol')
-            for asset in source['url']:
-                asset_url = asset['text']
-                quality = asset['@quality']
-                format_id = '%s-%s' % (protocol, quality)
-                if protocol.startswith('HTTP-HDS') or protocol.startswith('HTTP-HLS'):
-                    asset_url = self._get_tokenized_src(asset_url, media_id, format_id)
-                    if protocol.startswith('HTTP-HDS'):
-                        formats.extend(self._extract_f4m_formats(
-                            asset_url + ('?' if '?' not in asset_url else '&') + 'hdcore=3.4.0',
-                            media_id, f4m_id=format_id, fatal=False))
-                    elif protocol.startswith('HTTP-HLS'):
-                        formats.extend(self._extract_m3u8_formats(
-                            asset_url, media_id, 'mp4', 'm3u8_native',
-                            m3u8_id=format_id, fatal=False))
-                else:
-                    formats.append({
-                        'format_id': format_id,
-                        'url': asset_url,
-                        'preference': preference(quality),
-                        'ext': 'flv' if protocol == 'RTMP' else None,
-                    })
+        q = qualities(['SD', 'HD'])
+        for source in (media_data.get('resourceList') or []):
+            format_url = source.get('url')
+            if not format_url:
+                continue
+            protocol = source.get('protocol')
+            quality = source.get('quality')
+            format_id = []
+            for e in (protocol, source.get('encoding'), quality):
+                if e:
+                    format_id.append(e)
+            format_id = '-'.join(format_id)
+
+            if protocol in ('HDS', 'HLS'):
+                if source.get('tokenType') == 'AKAMAI':
+                    format_url = self._get_tokenized_src(
+                        format_url, media_id, format_id)
+                    formats.extend(self._extract_akamai_formats(
+                        format_url, media_id))
+                elif protocol == 'HLS':
+                    formats.extend(self._extract_m3u8_formats(
+                        format_url, media_id, 'mp4', 'm3u8_native',
+                        m3u8_id=format_id, fatal=False))
+            elif protocol in ('HTTP', 'HTTPS'):
+                formats.append({
+                    'format_id': format_id,
+                    'url': format_url,
+                    'quality': q(quality),
+                })
+
+        # This is needed because for audio medias the podcast url is usually
+        # always included, even if is only an audio segment and not the
+        # whole episode.
+        if int_or_none(media_data.get('position')) == 0:
+            for p in ('S', 'H'):
+                podcast_url = media_data.get('podcast%sdUrl' % p)
+                if not podcast_url:
+                    continue
+                quality = p + 'D'
+                formats.append({
+                    'format_id': 'PODCAST-' + quality,
+                    'url': podcast_url,
+                    'quality': q(quality),
+                })
         self._sort_formats(formats)
 
+        subtitles = {}
+        if media_type == 'video':
+            for sub in (media_data.get('subtitleList') or []):
+                sub_url = sub.get('url')
+                if not sub_url:
+                    continue
+                lang = sub.get('locale') or self._DEFAULT_LANGUAGE_CODES[bu]
+                subtitles.setdefault(lang, []).append({
+                    'url': sub_url,
+                })
+
         return {
             'id': media_id,
             'title': title,
-            'description': description,
-            'timestamp': timestamp,
-            'thumbnails': thumbnails,
+            'description': media_data.get('description'),
+            'timestamp': parse_iso8601(media_data.get('date')),
+            'thumbnail': media_data.get('imageUrl'),
+            'duration': float_or_none(media_data.get('duration'), 1000),
+            'subtitles': subtitles,
             'formats': formats,
         }
 
@@ -119,26 +171,17 @@ class SRGSSRPlayIE(InfoExtractor):
 
     _TESTS = [{
         'url': 'http://www.srf.ch/play/tv/10vor10/video/snowden-beantragt-asyl-in-russland?id=28e1a57d-5b76-4399-8ab3-9097f071e6c5',
-        'md5': 'da6b5b3ac9fa4761a942331cef20fcb3',
+        'md5': '6db2226ba97f62ad42ce09783680046c',
         'info_dict': {
             'id': '28e1a57d-5b76-4399-8ab3-9097f071e6c5',
             'ext': 'mp4',
             'upload_date': '20130701',
             'title': 'Snowden beantragt Asyl in Russland',
-            'timestamp': 1372713995,
-        }
-    }, {
-        # No Speichern (Save) button
-        'url': 'http://www.srf.ch/play/tv/top-gear/video/jaguar-xk120-shadow-und-tornado-dampflokomotive?id=677f5829-e473-4823-ac83-a1087fe97faa',
-        'md5': '0a274ce38fda48c53c01890651985bc6',
-        'info_dict': {
-            'id': '677f5829-e473-4823-ac83-a1087fe97faa',
-            'ext': 'flv',
-            'upload_date': '20130710',
-            'title': 'Jaguar XK120, Shadow und Tornado-Dampflokomotive',
-            'description': 'md5:88604432b60d5a38787f152dec89cd56',
-            'timestamp': 1373493600,
+            'timestamp': 1372708215,
+            'duration': 113.827,
+            'thumbnail': r're:^https?://.*1383719781\.png$',
         },
+        'expected_warnings': ['Unable to download f4m manifest'],
     }, {
         'url': 'http://www.rtr.ch/play/radio/actualitad/audio/saira-tujetsch-tuttina-cuntinuar-cun-sedrun-muster-turissem?id=63cb0778-27f8-49af-9284-8c7a8c6d15fc',
         'info_dict': {
@@ -146,7 +189,8 @@ class SRGSSRPlayIE(InfoExtractor):
             'ext': 'mp3',
             'upload_date': '20151013',
             'title': 'Saira: Tujetsch - tuttina cuntinuar cun Sedrun Mustér Turissem',
-            'timestamp': 1444750398,
+            'timestamp': 1444709160,
+            'duration': 336.816,
         },
         'params': {
             # rtmp download
@@ -159,19 +203,32 @@ class SRGSSRPlayIE(InfoExtractor):
             'id': '6348260',
             'display_id': '6348260',
             'ext': 'mp4',
-            'duration': 1796,
+            'duration': 1796.76,
             'title': 'Le 19h30',
-            'description': '',
-            'uploader': '19h30',
             'upload_date': '20141201',
             'timestamp': 1417458600,
             'thumbnail': r're:^https?://.*\.image',
-            'view_count': int,
         },
         'params': {
             # m3u8 download
             'skip_download': True,
         }
+    }, {
+        'url': 'http://play.swissinfo.ch/play/tv/business/video/why-people-were-against-tax-reforms?id=42960270',
+        'info_dict': {
+            'id': '42960270',
+            'ext': 'mp4',
+            'title': 'Why people were against tax reforms',
+            'description': 'md5:7ac442c558e9630e947427469c4b824d',
+            'duration': 94.0,
+            'upload_date': '20170215',
+            'timestamp': 1487173560,
+            'thumbnail': r're:https?://www\.swissinfo\.ch/srgscalableimage/42961964',
+            'subtitles': 'count:9',
+        },
+        'params': {
+            'skip_download': True,
+        }
     }, {
         'url': 'https://www.srf.ch/play/tv/popupvideoplayer?id=c4dba0ca-e75b-43b2-a34f-f708a4932e01',
         'only_matching': True,
@@ -181,6 +238,10 @@ class SRGSSRPlayIE(InfoExtractor):
     }, {
         'url': 'https://www.rts.ch/play/tv/19h30/video/le-19h30?urn=urn:rts:video:6348260',
         'only_matching': True,
+    }, {
+        # audio segment, has podcastSdUrl of the full episode
+        'url': 'https://www.srf.ch/play/radio/popupaudioplayer?id=50b20dc8-f05b-4972-bf03-e438ff2833eb',
+        'only_matching': True,
     }]
 
     def _real_extract(self, url):
@@ -188,5 +249,4 @@ class SRGSSRPlayIE(InfoExtractor):
         bu = mobj.group('bu')
         media_type = mobj.group('type') or mobj.group('type_2')
         media_id = mobj.group('id')
-        # other info can be extracted from url + '&layout=json'
         return self.url_result('srgssr:%s:%s:%s' % (bu[:3], media_type, media_id), 'SRGSSR')