[tvp] Fix extractor (#1401)

Authored by: selfisekai

yt_dlp/extractor/tvp.py

@@ -2,35 +2,40 @@
 from __future__ import unicode_literals
 
 import itertools
+import random
 import re
 
 from .common import InfoExtractor
 from ..utils import (
-    clean_html,
     determine_ext,
+    dict_get,
     ExtractorError,
-    get_element_by_attribute,
+    int_or_none,
+    js_to_json,
     orderedSet,
+    str_or_none,
+    try_get,
 )
 
 
 class TVPIE(InfoExtractor):
     IE_NAME = 'tvp'
     IE_DESC = 'Telewizja Polska'
-    _VALID_URL = r'https?://[^/]+\.tvp\.(?:pl|info)/(?:video/(?:[^,\s]*,)*|(?:(?!\d+/)[^/]+/)*)(?P<id>\d+)'
+    _VALID_URL = r'https?://(?:[^/]+\.)?(?:tvp(?:parlament)?\.(?:pl|info)|polandin\.com)/(?:video/(?:[^,\s]*,)*|(?:(?!\d+/)[^/]+/)*)(?P<id>\d+)'
 
     _TESTS = [{
+        # TVPlayer 2 in js wrapper
         'url': 'https://vod.tvp.pl/video/czas-honoru,i-seria-odc-13,194536',
-        'md5': 'a21eb0aa862f25414430f15fdfb9e76c',
         'info_dict': {
             'id': '194536',
             'ext': 'mp4',
             'title': 'Czas honoru, odc. 13 – Władek',
             'description': 'md5:437f48b93558370b031740546b696e24',
+            'age_limit': 12,
         },
     }, {
+        # TVPlayer legacy
         'url': 'http://www.tvp.pl/there-can-be-anything-so-i-shortened-it/17916176',
-        'md5': 'b0005b542e5b4de643a9690326ab1257',
         'info_dict': {
             'id': '17916176',
             'ext': 'mp4',
@@ -38,16 +43,63 @@ class TVPIE(InfoExtractor):
             'description': 'TVP Gorzów pokaże filmy studentów z podroży dookoła świata',
         },
     }, {
-        # page id is not the same as video id(#7799)
-        'url': 'https://wiadomosci.tvp.pl/33908820/28092017-1930',
-        'md5': '84cd3c8aec4840046e5ab712416b73d0',
+        # TVPlayer 2 in iframe
+        'url': 'https://wiadomosci.tvp.pl/50725617/dzieci-na-sprzedaz-dla-homoseksualistow',
         'info_dict': {
-            'id': '33908820',
+            'id': '50725617',
             'ext': 'mp4',
-            'title': 'Wiadomości, 28.09.2017, 19:30',
-            'description': 'Wydanie główne codziennego serwisu informacyjnego.'
+            'title': 'Dzieci na sprzedaż dla homoseksualistów',
+            'description': 'md5:7d318eef04e55ddd9f87a8488ac7d590',
+            'age_limit': 12,
         },
-        'skip': 'HTTP Error 404: Not Found',
+    }, {
+        # TVPlayer 2 in client-side rendered website (regional; window.__newsData)
+        'url': 'https://warszawa.tvp.pl/25804446/studio-yayo',
+        'info_dict': {
+            'id': '25804446',
+            'ext': 'mp4',
+            'title': 'Studio Yayo',
+            'upload_date': '20160616',
+            'timestamp': 1466075700,
+        }
+    }, {
+        # TVPlayer 2 in client-side rendered website (tvp.info; window.__videoData)
+        'url': 'https://www.tvp.info/52880236/09042021-0800',
+        'info_dict': {
+            'id': '52880236',
+            'ext': 'mp4',
+            'title': '09.04.2021, 08:00',
+        },
+    }, {
+        # client-side rendered (regional) program (playlist) page
+        'url': 'https://opole.tvp.pl/9660819/rozmowa-dnia',
+        'info_dict': {
+            'id': '9660819',
+            'description': 'Od poniedziałku do piątku o 18:55',
+            'title': 'Rozmowa dnia',
+        },
+        'playlist_mincount': 1800,
+        'params': {
+            'skip_download': True,
+        }
+    }, {
+        # ABC-specific video embeding
+        # moved to https://bajkowakraina.tvp.pl/wideo/50981130,teleranek,51027049,zubr,51116450
+        'url': 'https://abc.tvp.pl/48636269/zubry-odc-124',
+        'info_dict': {
+            'id': '48320456',
+            'ext': 'mp4',
+            'title': 'Teleranek, Żubr',
+        },
+        'skip': 'unavailable',
+    }, {
+        # yet another vue page
+        'url': 'https://jp2.tvp.pl/46925618/filmy',
+        'info_dict': {
+            'id': '46925618',
+            'title': 'Filmy',
+        },
+        'playlist_mincount': 19,
     }, {
         'url': 'http://vod.tvp.pl/seriale/obyczajowe/na-sygnale/sezon-2-27-/odc-39/17834272',
         'only_matching': True,
@@ -66,21 +118,134 @@ class TVPIE(InfoExtractor):
     }, {
         'url': 'http://www.tvp.info/25511919/trwa-rewolucja-wladza-zdecydowala-sie-na-pogwalcenie-konstytucji',
         'only_matching': True,
+    }, {
+        'url': 'https://tvp.info/49193823/teczowe-flagi-na-pomnikach-prokuratura-wszczela-postepowanie-wieszwiecej',
+        'only_matching': True,
+    }, {
+        'url': 'https://www.tvpparlament.pl/retransmisje-vod/inne/wizyta-premiera-mateusza-morawieckiego-w-firmie-berotu-sp-z-oo/48857277',
+        'only_matching': True,
+    }, {
+        'url': 'https://polandin.com/47942651/pln-10-billion-in-subsidies-transferred-to-companies-pm',
+        'only_matching': True,
     }]
 
+    def _parse_vue_website_data(self, webpage, page_id):
+        website_data = self._search_regex([
+            # website - regiony, tvp.info
+            # directory - jp2.tvp.pl
+            r'window\.__(?:website|directory)Data\s*=\s*({(?:.|\s)+?});',
+        ], webpage, 'website data')
+        if not website_data:
+            return None
+        return self._parse_json(website_data, page_id, transform_source=js_to_json)
+
+    def _extract_vue_video(self, video_data, page_id=None):
+        if isinstance(video_data, str):
+            video_data = self._parse_json(video_data, page_id, transform_source=js_to_json)
+        thumbnails = []
+        image = video_data.get('image')
+        if image:
+            for thumb in (image if isinstance(image, list) else [image]):
+                thmb_url = str_or_none(thumb.get('url'))
+                if thmb_url:
+                    thumbnails.append({
+                        'url': thmb_url,
+                    })
+        is_website = video_data.get('type') == 'website'
+        if is_website:
+            url = video_data['url']
+            fucked_up_url_parts = re.match(r'https?://vod\.tvp\.pl/(\d+)/([^/?#]+)', url)
+            if fucked_up_url_parts:
+                url = f'https://vod.tvp.pl/website/{fucked_up_url_parts.group(2)},{fucked_up_url_parts.group(1)}'
+        else:
+            url = 'tvp:' + str_or_none(video_data.get('_id') or page_id)
+        return {
+            '_type': 'url_transparent',
+            'id': str_or_none(video_data.get('_id') or page_id),
+            'url': url,
+            'ie_key': 'TVPEmbed' if not is_website else 'TVPWebsite',
+            'title': str_or_none(video_data.get('title')),
+            'description': str_or_none(video_data.get('lead')),
+            'timestamp': int_or_none(video_data.get('release_date_long')),
+            'duration': int_or_none(video_data.get('duration')),
+            'thumbnails': thumbnails,
+        }
+
+    def _handle_vuejs_page(self, url, webpage, page_id):
+        # vue client-side rendered sites (all regional pages + tvp.info)
+        video_data = self._search_regex([
+            r'window\.__(?:news|video)Data\s*=\s*({(?:.|\s)+?})\s*;',
+        ], webpage, 'video data', default=None)
+        if video_data:
+            return self._extract_vue_video(video_data, page_id=page_id)
+        # paged playlists
+        website_data = self._parse_vue_website_data(webpage, page_id)
+        if website_data:
+            entries = self._vuejs_entries(url, website_data, page_id)
+
+            return {
+                '_type': 'playlist',
+                'id': page_id,
+                'title': str_or_none(website_data.get('title')),
+                'description': str_or_none(website_data.get('lead')),
+                'entries': entries,
+            }
+        raise ExtractorError('Could not extract video/website data')
+
+    def _vuejs_entries(self, url, website_data, page_id):
+
+        def extract_videos(wd):
+            if wd.get('latestVideo'):
+                yield self._extract_vue_video(wd['latestVideo'])
+            for video in wd.get('videos') or []:
+                yield self._extract_vue_video(video)
+            for video in wd.get('items') or []:
+                yield self._extract_vue_video(video)
+
+        yield from extract_videos(website_data)
+
+        if website_data.get('items_total_count') > website_data.get('items_per_page'):
+            for page in itertools.count(2):
+                page_website_data = self._parse_vue_website_data(
+                    self._download_webpage(url, page_id, note='Downloading page #%d' % page,
+                                           query={'page': page}),
+                    page_id)
+                if not page_website_data.get('videos') and not page_website_data.get('items'):
+                    break
+                yield from extract_videos(page_website_data)
+
     def _real_extract(self, url):
         page_id = self._match_id(url)
-        webpage = self._download_webpage(url, page_id)
+        webpage, urlh = self._download_webpage_handle(url, page_id)
+
+        # The URL may redirect to a VOD
+        # example: https://vod.tvp.pl/48463890/wadowickie-spotkania-z-janem-pawlem-ii
+        if TVPWebsiteIE.suitable(urlh.url):
+            return self.url_result(urlh.url, ie=TVPWebsiteIE.ie_key(), video_id=page_id)
+
+        if re.search(
+                r'window\.__(?:video|news|website|directory)Data\s*=',
+                webpage):
+            return self._handle_vuejs_page(url, webpage, page_id)
+
+        # classic server-side rendered sites
         video_id = self._search_regex([
+            r'<iframe[^>]+src="[^"]*?embed\.php\?(?:[^&]+&)*ID=(\d+)',
             r'<iframe[^>]+src="[^"]*?object_id=(\d+)',
             r"object_id\s*:\s*'(\d+)'",
-            r'data-video-id="(\d+)"'], webpage, 'video id', default=page_id)
+            r'data-video-id="(\d+)"',
+
+            # abc.tvp.pl - somehow there are more than one video IDs that seem to be the same video?
+            # the first one is referenced to as "copyid", and seems to be unused by the website
+            r'<script>\s*tvpabc\.video\.init\(\s*\d+,\s*(\d+)\s*\)\s*</script>',
+        ], webpage, 'video id', default=page_id)
         return {
             '_type': 'url_transparent',
             'url': 'tvp:' + video_id,
             'description': self._og_search_description(
-                webpage, default=None) or self._html_search_meta(
-                'description', webpage, default=None),
+                webpage, default=None) or (self._html_search_meta(
+                    'description', webpage, default=None)
+                    if '//s.tvp.pl/files/portal/v' in webpage else None),
             'thumbnail': self._og_search_thumbnail(webpage, default=None),
             'ie_key': 'TVPEmbed',
         }
@@ -252,18 +417,20 @@ class TVPWebsiteIE(InfoExtractor):
 
     _TESTS = [{
         # series
-        'url': 'https://vod.tvp.pl/website/lzy-cennet,38678312/video',
+        'url': 'https://vod.tvp.pl/website/wspaniale-stulecie,17069012/video',
         'info_dict': {
-            'id': '38678312',
+            'id': '17069012',
         },
-        'playlist_count': 115,
+        'playlist_count': 312,
     }, {
         # film
-        'url': 'https://vod.tvp.pl/website/gloria,35139666',
+        'url': 'https://vod.tvp.pl/website/krzysztof-krawczyk-cale-moje-zycie,51374466',
         'info_dict': {
-            'id': '36637049',
+            'id': '51374509',
             'ext': 'mp4',
-            'title': 'Gloria, Gloria',
+            'title': 'Krzysztof Krawczyk – całe moje życie, Krzysztof Krawczyk – całe moje życie',
+            'description': 'md5:2e80823f00f5fc263555482f76f8fa42',
+            'age_limit': 12,
         },
         'params': {
             'skip_download': True,