[vvvvid] imporove info extraction

closes #18130

youtube_dl/extractor/extractors.py

@@ -1425,7 +1425,10 @@ from .vshare import VShareIE
 from .medialaan import MedialaanIE
 from .vube import VubeIE
 from .vuclip import VuClipIE
-from .vvvvid import VVVVIDIE
+from .vvvvid import (
+    VVVVIDIE,
+    VVVVIDShowIE,
+)
 from .vyborymos import VyboryMosIE
 from .vzaar import VzaarIE
 from .wakanim import WakanimIE

youtube_dl/extractor/vvvvid.py

@@ -12,7 +12,8 @@ from ..utils import (
 
 
 class VVVVIDIE(InfoExtractor):
-    _VALID_URL = r'https?://(?:www\.)?vvvvid\.it/(?:#!)?(?:show|anime|film|series)/(?P<show_id>\d+)/[^/]+/(?P<season_id>\d+)/(?P<id>[0-9]+)'
+    _VALID_URL_BASE = r'https?://(?:www\.)?vvvvid\.it/(?:#!)?(?:show|anime|film|series)/'
+    _VALID_URL = r'%s(?P<show_id>\d+)/[^/]+/(?P<season_id>\d+)/(?P<id>[0-9]+)' % _VALID_URL_BASE
     _TESTS = [{
         # video_type == 'video/vvvvid'
         'url': 'https://www.vvvvid.it/#!show/434/perche-dovrei-guardarlo-di-dario-moccia/437/489048/ping-pong',
@@ -21,6 +22,16 @@ class VVVVIDIE(InfoExtractor):
             'id': '489048',
             'ext': 'mp4',
             'title': 'Ping Pong',
+            'duration': 239,
+            'series': '"Perché dovrei guardarlo?" di Dario Moccia',
+            'season_id': '437',
+            'season_number': 1,
+            'episode': 'Ping Pong',
+            'episode_number': 1,
+            'episode_id': '3334',
+            'view_count': int,
+            'like_count': int,
+            'repost_count': int,
         },
         'params': {
             'skip_download': True,
@@ -37,6 +48,9 @@ class VVVVIDIE(InfoExtractor):
         'params': {
             'skip_download': True,
         },
+    }, {
+        'url': 'https://www.vvvvid.it/show/434/perche-dovrei-guardarlo-di-dario-moccia/437/489048',
+        'only_matching': True
     }]
     _conn_id = None
 
@@ -45,20 +59,36 @@ class VVVVIDIE(InfoExtractor):
             'https://www.vvvvid.it/user/login',
             None, headers=self.geo_verification_headers())['data']['conn_id']
 
-    def _real_extract(self, url):
-        show_id, season_id, video_id = re.match(self._VALID_URL, url).groups()
+    def _download_info(self, show_id, path, video_id, fatal=True):
         response = self._download_json(
-            'https://www.vvvvid.it/vvvvid/ondemand/%s/season/%s' % (show_id, season_id),
+            'https://www.vvvvid.it/vvvvid/ondemand/%s/%s' % (show_id, path),
             video_id, headers=self.geo_verification_headers(), query={
                 'conn_id': self._conn_id,
-            })
-        if response['result'] == 'error':
+            }, fatal=fatal)
+        if not (response or fatal):
+            return
+        if response.get('result') == 'error':
             raise ExtractorError('%s said: %s' % (
                 self.IE_NAME, response['message']), expected=True)
+        return response['data']
+
+    def _extract_common_video_info(self, video_data):
+        return {
+            'thumbnail': video_data.get('thumbnail'),
+            'episode_number': int_or_none(video_data.get('number')),
+            'episode_id': str_or_none(video_data.get('id')),
+        }
+
+    def _real_extract(self, url):
+        show_id, season_id, video_id = re.match(self._VALID_URL, url).groups()
+
+        response = self._download_info(
+            show_id, 'season/%s' % season_id, video_id)
 
         vid = int(video_id)
         video_data = list(filter(
-            lambda episode: episode.get('video_id') == vid, response['data']))[0]
+            lambda episode: episode.get('video_id') == vid, response))[0]
+        title = video_data['title']
         formats = []
 
         # vvvvid embed_info decryption algorithm is reverse engineered from function $ds(h) at vvvvid.js
@@ -141,18 +171,67 @@ class VVVVIDIE(InfoExtractor):
                     'http://sb.top-ix.org/videomg/_definst_/mp4:%s/playlist.m3u8' % embed_code, video_id))
         self._sort_formats(formats)
 
-        return {
+        info = self._extract_common_video_info(video_data)
+        info.update({
             'id': video_id,
-            'title': video_data['title'],
+            'title': title,
             'formats': formats,
-            'thumbnail': video_data.get('thumbnail'),
             'duration': int_or_none(video_data.get('length')),
             'series': video_data.get('show_title'),
             'season_id': season_id,
             'season_number': video_data.get('season_number'),
-            'episode_id': str_or_none(video_data.get('id')),
-            'episode_number': int_or_none(video_data.get('number')),
-            'episode_title': video_data['title'],
+            'episode': title,
             'view_count': int_or_none(video_data.get('views')),
             'like_count': int_or_none(video_data.get('video_likes')),
-        }
+            'repost_count': int_or_none(video_data.get('video_shares')),
+        })
+        return info
+
+
+class VVVVIDShowIE(VVVVIDIE):
+    _VALID_URL = r'(?P<base_url>%s(?P<id>\d+)(?:/(?P<show_title>[^/?&#]+))?)/?(?:[?#&]|$)' % VVVVIDIE._VALID_URL_BASE
+    _TESTS = [{
+        'url': 'https://www.vvvvid.it/show/156/psyco-pass',
+        'info_dict': {
+            'id': '156',
+            'title': 'Psycho-Pass',
+            'description': 'md5:94d572c0bd85894b193b8aebc9a3a806',
+        },
+        'playlist_count': 46,
+    }, {
+        'url': 'https://www.vvvvid.it/show/156',
+        'only_matching': True,
+    }]
+
+    def _real_extract(self, url):
+        base_url, show_id, show_title = re.match(self._VALID_URL, url).groups()
+
+        seasons = self._download_info(
+            show_id, 'seasons/', show_title)
+
+        show_info = self._download_info(
+            show_id, 'info/', show_title, fatal=False)
+
+        entries = []
+        for season in (seasons or []):
+            season_number = int_or_none(season.get('number'))
+            episodes = season.get('episodes') or []
+            for episode in episodes:
+                season_id = str_or_none(episode.get('season_id'))
+                video_id = str_or_none(episode.get('video_id'))
+                if not (season_id and video_id):
+                    continue
+                info = self._extract_common_video_info(episode)
+                info.update({
+                    '_type': 'url',
+                    'ie_key': VVVVIDIE.ie_key(),
+                    'url': '/'.join([base_url, season_id, video_id]),
+                    'title': episode.get('title'),
+                    'description': episode.get('description'),
+                    'season_number': season_number,
+                    'season_id': season_id,
+                })
+                entries.append(info)
+
+        return self.playlist_result(
+            entries, show_id, show_info.get('title'), show_info.get('description'))

youtube_dl/extractor/yandexdisk.py

@@ -1,23 +1,43 @@
 # coding: utf-8
 from __future__ import unicode_literals
 
+import json
+import re
+
 from .common import InfoExtractor
-from ..compat import compat_str
 from ..utils import (
     determine_ext,
     float_or_none,
     int_or_none,
+    mimetype2ext,
     try_get,
-    urlencode_postdata,
+    urljoin,
 )
 
 
 class YandexDiskIE(InfoExtractor):
-    _VALID_URL = r'https?://yadi\.sk/[di]/(?P<id>[^/?#&]+)'
+    _VALID_URL = r'''(?x)https?://
+        (?P<domain>
+            yadi\.sk|
+            disk\.yandex\.
+                (?:
+                    az|
+                    by|
+                    co(?:m(?:\.(?:am|ge|tr))?|\.il)|
+                    ee|
+                    fr|
+                    k[gz]|
+                    l[tv]|
+                    md|
+                    t[jm]|
+                    u[az]|
+                    ru
+                )
+        )/(?:[di]/|public.*?\bhash=)(?P<id>[^/?#&]+)'''
 
     _TESTS = [{
         'url': 'https://yadi.sk/i/VdOeDou8eZs6Y',
-        'md5': '33955d7ae052f15853dc41f35f17581c',
+        'md5': 'a4a8d52958c8fddcf9845935070402ae',
         'info_dict': {
             'id': 'VdOeDou8eZs6Y',
             'ext': 'mp4',
@@ -27,92 +47,101 @@ class YandexDiskIE(InfoExtractor):
             'uploader_id': '300043621',
             'view_count': int,
         },
+        'expected_warnings': ['Unable to download JSON metadata'],
     }, {
         'url': 'https://yadi.sk/d/h3WAXvDS3Li3Ce',
         'only_matching': True,
+    }, {
+        'url': 'https://yadi.sk/public?hash=5DZ296JK9GWCLp02f6jrObjnctjRxMs8L6%2B%2FuhNqk38%3D',
+        'only_matching': True,
     }]
 
     def _real_extract(self, url):
-        video_id = self._match_id(url)
-
-        status = self._download_webpage(
-            'https://disk.yandex.com/auth/status', video_id, query={
-                'urlOrigin': url,
-                'source': 'public',
-                'md5': 'false',
-            })
-
-        sk = self._search_regex(
-            r'(["\'])sk(?:External)?\1\s*:\s*(["\'])(?P<value>(?:(?!\2).)+)\2',
-            status, 'sk', group='value')
+        domain, video_id = re.match(self._VALID_URL, url).groups()
 
         webpage = self._download_webpage(url, video_id)
+        store = self._parse_json(self._search_regex(
+            r'<script[^>]+id="store-prefetch"[^>]*>\s*({.+?})\s*</script>',
+            webpage, 'store'), video_id)
+        resource = store['resources'][store['rootResourceId']]
 
-        models = self._parse_json(
-            self._search_regex(
-                r'<script[^>]+id=["\']models-client[^>]+>\s*(\[.+?\])\s*</script',
-                webpage, 'video JSON'),
-            video_id)
+        title = resource['name']
+        meta = resource.get('meta') or {}
 
-        data = next(
-            model['data'] for model in models
-            if model.get('model') == 'resource')
+        public_url = meta.get('short_url')
+        if public_url:
+            video_id = self._match_id(public_url)
 
-        video_hash = data['id']
-        title = data['name']
+        source_url = (self._download_json(
+            'https://cloud-api.yandex.net/v1/disk/public/resources/download',
+            video_id, query={'public_key': url}, fatal=False) or {}).get('href')
+        video_streams = resource.get('videoStreams') or {}
+        video_hash = resource.get('hash') or url
+        environment = store.get('environment') or {}
+        sk = environment.get('sk')
+        yandexuid = environment.get('yandexuid')
+        if sk and yandexuid and not (source_url and video_streams):
+            self._set_cookie(domain, 'yandexuid', yandexuid)
 
-        models = self._download_json(
-            'https://disk.yandex.com/models/', video_id,
-            data=urlencode_postdata({
-                '_model.0': 'videoInfo',
-                'id.0': video_hash,
-                '_model.1': 'do-get-resource-url',
-                'id.1': video_hash,
-                'version': '13.6',
-                'sk': sk,
-            }), query={'_m': 'videoInfo'})['models']
-
-        videos = try_get(models, lambda x: x[0]['data']['videos'], list) or []
-        source_url = try_get(
-            models, lambda x: x[1]['data']['file'], compat_str)
+            def call_api(action):
+                return (self._download_json(
+                    urljoin(url, '/public/api/') + action, video_id, data=json.dumps({
+                        'hash': video_hash,
+                        'sk': sk,
+                    }).encode(), headers={
+                        'Content-Type': 'text/plain',
+                    }, fatal=False) or {}).get('data') or {}
+            if not source_url:
+                # TODO: figure out how to detect if download limit has
+                # been reached and then avoid unnecessary source format
+                # extraction requests
+                source_url = call_api('download-url').get('url')
+            if not video_streams:
+                video_streams = call_api('get-video-streams')
 
         formats = []
         if source_url:
             formats.append({
                 'url': source_url,
                 'format_id': 'source',
-                'ext': determine_ext(title, 'mp4'),
+                'ext': determine_ext(title, meta.get('ext') or mimetype2ext(meta.get('mime_type')) or 'mp4'),
                 'quality': 1,
+                'filesize': int_or_none(meta.get('size'))
             })
-        for video in videos:
+
+        for video in (video_streams.get('videos') or []):
             format_url = video.get('url')
             if not format_url:
                 continue
-            if determine_ext(format_url) == 'm3u8':
+            if video.get('dimension') == 'adaptive':
                 formats.extend(self._extract_m3u8_formats(
-                    format_url, video_id, 'mp4', entry_protocol='m3u8_native',
+                    format_url, video_id, 'mp4', 'm3u8_native',
                     m3u8_id='hls', fatal=False))
             else:
+                size = video.get('size') or {}
+                height = int_or_none(size.get('height'))
+                format_id = 'hls'
+                if height:
+                    format_id += '-%dp' % height
                 formats.append({
+                    'ext': 'mp4',
+                    'format_id': format_id,
+                    'height': height,
+                    'protocol': 'm3u8_native',
                     'url': format_url,
+                    'width': int_or_none(size.get('width')),
                 })
         self._sort_formats(formats)
 
-        duration = float_or_none(try_get(
-            models, lambda x: x[0]['data']['duration']), 1000)
-        uploader = try_get(
-            data, lambda x: x['user']['display_name'], compat_str)
-        uploader_id = try_get(
-            data, lambda x: x['user']['uid'], compat_str)
-        view_count = int_or_none(try_get(
-            data, lambda x: x['meta']['views_counter']))
+        uid = resource.get('uid')
+        display_name = try_get(store, lambda x: x['users'][uid]['displayName'])
 
         return {
             'id': video_id,
             'title': title,
-            'duration': duration,
-            'uploader': uploader,
-            'uploader_id': uploader_id,
-            'view_count': view_count,
+            'duration': float_or_none(video_streams.get('duration'), 1000),
+            'uploader': display_name,
+            'uploader_id': uid,
+            'view_count': int_or_none(meta.get('views_counter')),
             'formats': formats,
         }

youtube_dl/extractor/yandexvideo.py

@@ -5,6 +5,7 @@ from .common import InfoExtractor
 from ..utils import (
     determine_ext,
     int_or_none,
+    try_get,
     url_or_none,
 )
 
@@ -64,12 +65,7 @@ class YandexVideoIE(InfoExtractor):
     def _real_extract(self, url):
         video_id = self._match_id(url)
 
-        content = self._download_json(
-            # 'https://frontend.vh.yandex.ru/v23/player/%s.json' % video_id,
-            # video_id, query={
-            #     'stream_options': 'hires',
-            #     'disable_trackings': 1,
-            # })['content']
+        player = try_get((self._download_json(
             'https://frontend.vh.yandex.ru/graphql', video_id, data=b'''{
   player(content_id: "%s") {
     computed_title
@@ -90,7 +86,15 @@ class YandexVideoIE(InfoExtractor):
     title
     views_count
   }
-}''' % video_id.encode())['player']['content']['content']
+}''' % video_id.encode(), fatal=False)), lambda x: x['player']['content'])
+        if not player or player.get('error'):
+            player = self._download_json(
+                'https://frontend.vh.yandex.ru/v23/player/%s.json' % video_id,
+                video_id, query={
+                    'stream_options': 'hires',
+                    'disable_trackings': 1,
+                })
+        content = player['content']
 
         title = content.get('title') or content['computed_title']