Merge f5dd875a02 into 4d05f84325

* no space after @ in decorator

test/test_jsinterp.py

@@ -577,9 +577,11 @@ class TestJSInterpreter(unittest.TestCase):
     def test_unary_operators(self):
         jsi = JSInterpreter('function f(){return 2  -  - - 2;}')
         self.assertEqual(jsi.call_function('f'), 0)
-        # fails
+        jsi = JSInterpreter('function f(){return 2 + - + - - 2;}')
-        # jsi = JSInterpreter('function f(){return 2 + - + - - 2;}')
+        self.assertEqual(jsi.call_function('f'), 0)
-        # self.assertEqual(jsi.call_function('f'), 0)
+        # https://github.com/ytdl-org/youtube-dl/issues/32815
+        jsi = JSInterpreter('function f(){return 0  - 7 * - 6;}')
+        self.assertEqual(jsi.call_function('f'), 42)
 
     """ # fails so far
     def test_packed(self):

test/test_youtube_signature.py

@@ -158,6 +158,10 @@ _NSIG_TESTS = [
         'https://www.youtube.com/s/player/b7910ca8/player_ias.vflset/en_US/base.js',
         '_hXMCwMt9qE310D', 'LoZMgkkofRMCZQ',
     ),
+    (
+        'https://www.youtube.com/s/player/590f65a6/player_ias.vflset/en_US/base.js',
+        '1tm7-g_A9zsI8_Lay_', 'xI4Vem4Put_rOg',
+    ),
 ]
 
 

youtube_dl/extractor/common.py

@@ -3033,7 +3033,6 @@ class InfoExtractor(object):
             transform_source=transform_source, default=None)
 
     def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
-
         # allow passing `transform_source` through to _find_jwplayer_data()
         transform_source = kwargs.pop('transform_source', None)
         kwfind = compat_kwargs({'transform_source': transform_source}) if transform_source else {}

youtube_dl/extractor/extractors.py

@@ -1384,7 +1384,9 @@ from .tvnow import (
 from .tvp import (
     TVPEmbedIE,
     TVPIE,
-    TVPWebsiteIE,
+    TVPStreamIE,
+    TVPVODSeriesIE,
+    TVPVODVideoIE,
 )
 from .tvplay import (
     TVPlayIE,

youtube_dl/extractor/palcomp3.py

@@ -8,7 +8,7 @@ from ..compat import compat_str
 from ..utils import (
     int_or_none,
     str_or_none,
-    try_get,
+    traverse_obj,
 )
 
 
@@ -109,7 +109,7 @@ class PalcoMP3ArtistIE(PalcoMP3BaseIE):
     }
     name'''
 
-    @ classmethod
+    @classmethod
     def suitable(cls, url):
         return False if re.match(PalcoMP3IE._VALID_URL, url) else super(PalcoMP3ArtistIE, cls).suitable(url)
 
@@ -118,7 +118,8 @@ class PalcoMP3ArtistIE(PalcoMP3BaseIE):
         artist = self._call_api(artist_slug, self._ARTIST_FIELDS_TMPL)['artist']
 
         def entries():
-            for music in (try_get(artist, lambda x: x['musics']['nodes'], list) or []):
+            for music in traverse_obj(artist, (
+                    'musics', 'nodes', lambda _, m: m['musicID'])):
                 yield self._parse_music(music)
 
         return self.playlist_result(
@@ -137,7 +138,7 @@ class PalcoMP3VideoIE(PalcoMP3BaseIE):
             'title': 'Maiara e Maraisa - Você Faz Falta Aqui - DVD Ao Vivo Em Campo Grande',
             'description': 'md5:7043342c09a224598e93546e98e49282',
             'upload_date': '20161107',
-            'uploader_id': 'maiaramaraisaoficial',
+            'uploader_id': '@maiaramaraisaoficial',
             'uploader': 'Maiara e Maraisa',
         }
     }]

youtube_dl/extractor/tvp.py

@@ -2,52 +2,274 @@
 from __future__ import unicode_literals
 
 import itertools
+import random
 import re
 
 from .common import InfoExtractor
+from ..compat import (
+    compat_HTTPError,
+    compat_str,
+)
 from ..utils import (
     clean_html,
     determine_ext,
     ExtractorError,
-    get_element_by_attribute,
+    int_or_none,
-    orderedSet,
+    js_to_json,
+    traverse_obj,
+    url_or_none,
 )
 
 
+def txt_or_none(v, default=None):
+    return default if v is None else (compat_str(v).strip() or default)
+
+if not hasattr(InfoExtractor, '_match_valid_url'):
+
+    import sys
+    from ..compat import (
+        compat_os_name,
+        compat_re_Pattern as compiled_regex_type,
+    )
+    from ..utils import (
+        bug_reports_message,
+        error_to_compat_str,
+        NO_DEFAULT,
+        RegexNotFoundError,
+    )
+
+    BaseIE = InfoExtractor
+
+    class InfoExtractor(BaseIE):
+        def _match_valid_url(self, url):
+            return re.match(self._VALID_URL, url)
+
+        def _search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
+            """
+            Perform a regex search on the given string, using a single or a list of
+            patterns returning the first matching group.
+            In case of failure return a default value or raise a WARNING or a
+            RegexNotFoundError, depending on fatal, specifying the field name.
+            """
+            if isinstance(pattern, (str, compat_str, compiled_regex_type)):
+                mobj = re.search(pattern, string, flags)
+            else:
+                for p in pattern:
+                    mobj = re.search(p, string, flags)
+                    if mobj:
+                        break
+
+            if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
+                _name = '\033[0;34m%s\033[0m' % name
+            else:
+                _name = name
+
+            if mobj:
+                if group is None:
+                    # return the first matching group
+                    return next(g for g in mobj.groups() if g is not None)
+                elif isinstance(group, (list, tuple)):
+                    return tuple(mobj.group(g) for g in group)
+                else:
+                    return mobj.group(group)
+            elif default is not NO_DEFAULT:
+                return default
+            elif fatal:
+                raise RegexNotFoundError('Unable to extract %s' % _name)
+            else:
+                self.report_warning('unable to extract %s' % _name + bug_reports_message())
+                return None
+
+        def _html_search_regex(self, pattern, string, name, default=NO_DEFAULT, fatal=True, flags=0, group=None):
+            """
+            Like _search_regex, but strips HTML tags and unescapes entities.
+            """
+            res = self._search_regex(pattern, string, name, default, fatal, flags, group)
+            if isinstance(res, tuple):
+                return tuple(map(clean_html, res))
+            return clean_html(res or None)
+
+        def _search_json(self, start_pattern, string, name, video_id, **kwargs):
+            """Searches string for the JSON object specified by start_pattern"""
+
+            # self, start_pattern, string, name, video_id, *, end_pattern='',
+            # contains_pattern=r'{(?s:.+)}', fatal=True, default=NO_DEFAULT
+            end_pattern = kwargs.pop('end_pattern', '')
+            contains_pattern = kwargs.pop('contains_pattern', r'{(?:[\s\S]+)}')
+            fatal = kwargs.get('fatal', True)
+            default = kwargs.get('default', NO_DEFAULT)
+
+            # NB: end_pattern is only used to reduce the size of the initial match
+            if default is NO_DEFAULT:
+                default, has_default = {}, False
+            else:
+                fatal, has_default = False, True
+
+            json_string = self._search_regex(
+                r'(?:{0})\s*(?P<json>{1})\s*(?:{2})'.format(
+                    start_pattern, contains_pattern, end_pattern),
+                string, name, group='json', fatal=fatal, default=None if has_default else NO_DEFAULT)
+            if not json_string:
+                return default
+
+            try:
+                # return self._parse_json(json_string, video_id, ignore_extra=True, **kwargs)
+                return self._parse_json(json_string, video_id, **kwargs)
+            except ExtractorError as e:
+                if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty():
+                    _name = '\033[0;34m%s\033[0m' % name
+                else:
+                    _name = name
+                msg = 'Unable to extract {0} - Failed to parse JSON'.format(_name)
+                if fatal:
+                    raise ExtractorError(msg, cause=e.cause, video_id=video_id)
+                elif not has_default:
+                    self.report_warning(
+                        '{0}: {1}'.format(msg, error_to_compat_str(e)), video_id=video_id)
+            return default
+
+
 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|tvpworld\.com|swipeto\.pl)/(?:(?:(?!\d+/)[^/]+/)*|(?:video|website)/[^/]+,)(?P<id>\d+)'
-
     _TESTS = [{
+        # TVPlayer 2 in js wrapper
+        'url': 'https://swipeto.pl/64095316/uliczny-foxtrot-wypozyczalnia-kaset-kto-pamieta-dvdvideo',
+        'info_dict': {
+            'id': '64095316',
+            'ext': 'mp4',
+            'title': 'Uliczny Foxtrot — Wypożyczalnia kaset. Kto pamięta DVD-Video?',
+            'age_limit': 0,
+            'duration': 374,
+            'thumbnail': r're:https://.+',
+        },
+        'expected_warnings': [
+            'Failed to download ISM manifest: HTTP Error 404: Not Found',
+            'Failed to download m3u8 information: HTTP Error 404: Not Found',
+        ],
+        'skip': 'Video gone: 404 Nie znaleziono obiektu',
+    }, {
+        # TVPlayer 2 in js wrapper (redirect to VodVideo)
         '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',
+            'description': 'md5:76649d2014f65c99477be17f23a4dead',
+            'age_limit': 12,
         },
+        'add_ie': ['Generic', 'TVPEmbed'],
     }, {
-        'url': 'http://www.tvp.pl/there-can-be-anything-so-i-shortened-it/17916176',
+        # film (old format)
-        'md5': 'b0005b542e5b4de643a9690326ab1257',
+        'url': 'https://vod.tvp.pl/website/krzysztof-krawczyk-cale-moje-zycie,51374466',
         'info_dict': {
-            'id': '17916176',
+            'id': '51374509',
             'ext': 'mp4',
-            'title': 'TVP Gorzów pokaże filmy studentów z podroży dookoła świata',
+            'title': 'Krzysztof Krawczyk – całe moje życie, Krzysztof Krawczyk – całe moje życie',
-            'description': 'TVP Gorzów pokaże filmy studentów z podroży dookoła świata',
+            'description': 'md5:2e80823f00f5fc263555482f76f8fa42',
+            'age_limit': 12,
         },
+        'params': {
+            'skip_download': True,
+        },
+        'add_ie': ['TVPEmbed'],
+        'skip': 'This video is not available from your location due to geo restriction',
     }, {
+        # TVPlayer legacy
+        'url': 'https://www.tvp.pl/polska-press-video-uploader/wideo/62042351',
+        'info_dict': {
+            'id': '62042351',
+            'ext': 'mp4',
+            'title': 'Wideo',
+            'description': 'Wideo Kamera',
+            'duration': 24,
+            'age_limit': 0,
+            'thumbnail': r're:https://.+',
+        },
+        'add_ie': ['TVPEmbed'],
+    }, {
+        # TVPlayer 2 in iframe
         # page id is not the same as video id(#7799)
-        'url': 'https://wiadomosci.tvp.pl/33908820/28092017-1930',
+        'url': 'https://wiadomosci.tvp.pl/50725617/dzieci-na-sprzedaz-dla-homoseksualistow',
-        'md5': '84cd3c8aec4840046e5ab712416b73d0',
+        'md5': 'd35fb45103802488fcb7470e411b9ed4',
         'info_dict': {
-            'id': '33908820',
+            'id': '50725617',
             'ext': 'mp4',
-            'title': 'Wiadomości, 28.09.2017, 19:30',
+            'title': 'Dzieci na sprzedaż dla homoseksualistów',
-            'description': 'Wydanie główne codziennego serwisu informacyjnego.'
+            'description': 'md5:7d318eef04e55ddd9f87a8488ac7d590',
+            'age_limit': 12,
+            'duration': 259,
+            'thumbnail': r're:https://.+',
         },
-        'skip': 'HTTP Error 404: Not Found',
+        'add_ie': ['TVPEmbed'],
+    }, {
+        # 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,
+            'age_limit': 0,
+            'duration': 20,
+            'thumbnail': r're:https://.+',
+        },
+        'add_ie': ['TVPEmbed'],
+        'skip': 'Video is geo restricted',
+    }, {
+        # 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',
+            'age_limit': 0,
+            'thumbnail': r're:https://.+',
+        },
+        'add_ie': ['TVPEmbed'],
+        'skip': 'Video is geo restricted',
+    }, {
+        # 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': 'Video gone: Nie znaleziono obiektu',
+    }, {
+        # yet another vue page
+        'url': 'https://jp2.tvp.pl/46925618/filmy',
+        'info_dict': {
+            'id': '46925618',
+            'title': 'Filmy',
+        },
+        'playlist_mincount': 19,
+    }, {
+        # redirect
+        'url': 'https://vod.tvp.pl/48463890/wadowickie-spotkania-z-janem-pawlem-ii',
+        'info_dict': {
+            'id': '295157',
+            'title': 'Wadowickie spotkania z Janem Pawłem II',
+        },
+        'playlist_mincount': 12,
+        'add_ie': ['TVPEmbed', 'TVPVODSeries'],
     }, {
         'url': 'http://vod.tvp.pl/seriale/obyczajowe/na-sygnale/sezon-2-27-/odc-39/17834272',
         'only_matching': True,
@@ -66,31 +288,212 @@ 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://tvpworld.com/48583640/tescos-polish-business-bought-by-danish-chain-netto',
+        '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\s*\.\s*__(?:website|directory)Data\s*=\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, compat_str):
+            video_data = self._parse_json(video_data, page_id, transform_source=js_to_json)
+        video_id = txt_or_none(video_data.get('_id')) or page_id
+        if not video_id:
+            return
+        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 = 'https://vod.tvp.pl/website/' + ','.join(fucked_up_url_parts.group(2, 1))
+        else:
+            url = 'tvp:' + video_id
+        return {
+            '_type': 'url_transparent',
+            'id': video_id,
+            'url': url,
+            'ie_key': (TVPIE if is_website else TVPEmbedIE).ie_key(),
+            'title': txt_or_none(video_data.get('title')),
+            'description': txt_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': traverse_obj(video_data, ('image', (None, Ellipsis), 'url'), expected_type=url_or_none) or None,
+        }
+
+    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:
+            video_data = self._extract_vue_video(video_data, page_id=page_id)
+            if video_data:
+                return self._extract_vue_video(video_data, page_id=page_id)
+        else:
+            # 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': txt_or_none(website_data.get('title')),
+                    'description': txt_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):
+            for video in traverse_obj(wd, (None, ('latestVideo', (('videos', 'items'), Ellipsis)))):
+                video = self._extract_vue_video(video)
+                if video:
+                    yield video
+
+        for from_ in extract_videos(website_data):
+            yield from_
+
+        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
+                for from_ in extract_videos(page_website_data):
+                    yield from_
+
     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, expected_status=404)
-        video_id = self._search_regex([
+
+        # The URL may redirect to a VOD
+        # example: https://vod.tvp.pl/48463890/wadowickie-spotkania-z-janem-pawlem-ii
+        for ie_cls in (TVPVODSeriesIE, TVPVODVideoIE):
+            if ie_cls.suitable(urlh.url):
+                return self.url_result(urlh.url, ie=ie_cls.ie_key(), video_id=page_id)
+
+        if urlh.getcode() == 404:
+            raise compat_HTTPError(url, 404, 'HTTP Error 404: Not Found', urlh.headers, urlh)
+
+        if re.search(
+                r'window\s*\.\s*__(?: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(
+                webpage, default=None) or (self._html_search_meta(
-                'description', webpage, default=None),
+                    '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',
         }
 
 
+class TVPStreamIE(InfoExtractor):
+    IE_NAME = 'tvp:stream'
+    _VALID_URL = r'(?:tvpstream:|https?://(?:tvpstream\.vod|stream)\.tvp\.pl/(?:\?(?:[^&]+[&;])*channel_id=)?)(?P<id>\d*)'
+    _TESTS = [{
+        'url': 'https://stream.tvp.pl/?channel_id=56969941',
+        'only_matching': True,
+    }, {
+        'url': 'https://tvpstream.vod.tvp.pl/?channel_id=1455',
+        'info_dict': {
+            'id': r're:\d+',
+            'title': r're:\S.*',
+            'ext': 'mp4',
+        },
+        'params': {
+            'skip_download': 'm3u8',
+        },
+        'add_ie': ['TVPEmbed'],
+    }, {
+        'url': 'tvpstream:39821455',
+        'only_matching': True,
+    }, {
+        # the default stream when you provide no channel_id, most probably TVP Info
+        'url': 'tvpstream:',
+        'only_matching': True,
+    }, {
+        'url': 'https://tvpstream.vod.tvp.pl/',
+        'only_matching': True,
+    }]
+
+    def _real_extract(self, url):
+        channel_id = self._match_id(url)
+        channel_url = self._proto_relative_url('//stream.tvp.pl/?channel_id=%s' % channel_id or 'default')
+        webpage = self._download_webpage(channel_url, channel_id or 'default', 'Downloading channel webpage')
+        channels = self._search_json(
+            r'window\s*\.\s*__channels\s*=', webpage, 'channel list', channel_id,
+            contains_pattern=r'\[\s*\{[\s\S]+}\s*]')
+        channel = traverse_obj(channels, (lambda _, v: channel_id == compat_str(v['id'])), get_all=False) if channel_id else channels[0]
+        audition = traverse_obj(channel, ('items', lambda _, v: v['is_live'] is True), get_all=False)
+        return {
+            '_type': 'url_transparent',
+            'id': channel_id or channel['id'],
+            'url': 'tvp:%s' % (audition['video_id'], ),
+            'title': audition.get('title'),
+            'alt_title': channel.get('title'),
+            'is_live': True,
+            'ie_key': 'TVPEmbed',
+        }
+
+
 class TVPEmbedIE(InfoExtractor):
     IE_NAME = 'tvp:embed'
     IE_DESC = 'Telewizja Polska'
-    _VALID_URL = r'(?:tvp:|https?://[^/]+\.tvp\.(?:pl|info)/sess/tvplayer\.php\?.*?object_id=)(?P<id>\d+)'
+    # XFF is not effective
-
+    _GEO_BYPASS = False
+    _VALID_URL_PAT = (
+        r'''
+            (?:
+                tvp:
+                |https?://
+                    (?:[^/]+\.)?
+                    (?:tvp(?:parlament)?\.pl|tvp\.info|tvpworld\.com|swipeto\.pl)/
+                    (?:sess/
+                            (?:tvplayer\.php\?.*?object_id
+                            |TVPlayer2/(?:embed|api)\.php\?.*[Ii][Dd])
+                        |shared/details\.php\?.*?object_id)
+                    =)
+            (?P<id>\d+)
+        ''')
+    _VALID_URL = '(?x)' + _VALID_URL_PAT
+    _EMBED_REGEX = [r'(?x)<iframe[^>]+?src=(["\'])(?P<url>{0})'.format(_VALID_URL_PAT)]
     _TESTS = [{
         'url': 'tvp:194536',
         'md5': 'a21eb0aa862f25414430f15fdfb9e76c',
@@ -98,9 +501,16 @@ class TVPEmbedIE(InfoExtractor):
             'id': '194536',
             'ext': 'mp4',
             'title': 'Czas honoru, odc. 13 – Władek',
+            'description': 'md5:76649d2014f65c99477be17f23a4dead',
+            'age_limit': 12,
+            'duration': 2652,
+            'series': 'Czas honoru',
+            'episode': 'Episode 13',
+            'episode_number': 13,
+            'season': 'sezon 1',
+            'thumbnail': r're:https://.+',
         },
     }, {
-        # not available
         'url': 'http://www.tvp.pl/sess/tvplayer.php?object_id=22670268',
         'md5': '8c9cd59d16edabf39331f93bf8a766c7',
         'info_dict': {
@@ -108,7 +518,28 @@ class TVPEmbedIE(InfoExtractor):
             'ext': 'mp4',
             'title': 'Panorama, 07.12.2015, 15:40',
         },
-        'skip': 'Transmisja została zakończona lub materiał niedostępny',
+        'skip': 'Nie znaleziono obiektu',
+    }, {
+        'url': 'https://www.tvp.pl/sess/tvplayer.php?object_id=51247504&amp;autoplay=false',
+        'info_dict': {
+            'id': '51247504',
+            'ext': 'mp4',
+            'title': 'Razmova 091220',
+            'duration': 876,
+            'age_limit': 0,
+            'thumbnail': r're:https://.+',
+        },
+    }, {
+        # TVPlayer2 embed URL
+        'url': 'https://tvp.info/sess/TVPlayer2/embed.php?ID=50595757',
+        'only_matching': True,
+    }, {
+        'url': 'https://wiadomosci.tvp.pl/sess/TVPlayer2/api.php?id=51233452',
+        'only_matching': True,
+    }, {
+        # pulsembed on dziennik.pl
+        'url': 'https://www.tvp.pl/shared/details.php?copy_id=52205981&object_id=52204505&autoplay=false&is_muted=false&allowfullscreen=true&template=external-embed/video/iframe-video.html',
+        'only_matching': True,
     }, {
         'url': 'tvp:22670268',
         'only_matching': True,
@@ -117,136 +548,272 @@ class TVPEmbedIE(InfoExtractor):
     def _real_extract(self, url):
         video_id = self._match_id(url)
 
+        # could be anything that is a valid JS function name
+        callback = random.choice((
+            'jebac_pis',
+            'jebacpis',
+            'ziobro',
+            'sasin70',
+            'sasin_przejebal_70_milionow_PLN',
+            'tvp_is_a_state_propaganda_service',
+        ))
         webpage = self._download_webpage(
-            'http://www.tvp.pl/sess/tvplayer.php?object_id=%s' % video_id, video_id)
+            ('https://www.tvp.pl/sess/TVPlayer2/api.php?id=%s'
+             + '&@method=getTvpConfig&@callback=%s') % (video_id, callback), video_id)
 
-        error = self._html_search_regex(
+        # stripping JSONP padding
-            r'(?s)<p[^>]+\bclass=["\']notAvailable__text["\'][^>]*>(.+?)</p>',
+        null, datastr = self._search_regex(
-            webpage, 'error', default=None) or clean_html(
+            r'\s%s\s*\(\s*(?P<null>null\s*,\s*)?(?P<json>(?(null)\[\s*)?\{(?:[\s\S]+)}(?(null)]\s*))\)\s*;' % (re.escape(callback), ),
-            get_element_by_attribute('class', 'msg error', webpage))
+            webpage, 'JSON API result', group=('null', 'json'))
-        if error:
+        data = self._parse_json(datastr, video_id, fatal=False)
-            raise ExtractorError('%s said: %s' % (
+        if null:
-                self.IE_NAME, clean_html(error)), expected=True)
+            error_desc = traverse_obj(data, (0, 'desc'), expected_type=compat_str)
+            if error_desc == 'Obiekt wymaga płatności':
+                error_desc = 'Video requires payment and log-in, but log-in is not implemented'
+            raise ExtractorError(error_desc or 'unexpected JSON error', expected=error_desc)
 
-        title = self._search_regex(
+        content = data['content']
-            r'name\s*:\s*([\'"])Title\1\s*,\s*value\s*:\s*\1(?P<title>.+?)\1',
+        info = traverse_obj(content, 'info', expected_type=dict)
-            webpage, 'title', group='title')
-        series_title = self._search_regex(
-            r'name\s*:\s*([\'"])SeriesTitle\1\s*,\s*value\s*:\s*\1(?P<series>.+?)\1',
-            webpage, 'series', group='series', default=None)
-        if series_title:
-            title = '%s, %s' % (series_title, title)
 
-        thumbnail = self._search_regex(
+        if traverse_obj(info, 'isGeoBlocked', expected_type=bool):
-            r"poster\s*:\s*'([^']+)'", webpage, 'thumbnail', default=None)
+            # actual country list is not provided, we just assume it's always available in PL
+            self.raise_geo_restricted(countries=['PL'])
 
-        video_url = self._search_regex(
+        is_live = traverse_obj(info, 'isLive', expected_type=bool)
-            r'0:{src:([\'"])(?P<url>.*?)\1', webpage,
-            'formats', group='url', default=None)
-        if not video_url or 'material_niedostepny.mp4' in video_url:
-            video_url = self._download_json(
-                'http://www.tvp.pl/pub/stat/videofileinfo?video_id=%s' % video_id,
-                video_id)['video_url']
 
         formats = []
-        video_url_base = self._search_regex(
+        for file in traverse_obj(content, ('files', Ellipsis), expected_type=dict):
-            r'(https?://.+?/video)(?:\.(?:ism|f4m|m3u8)|-\d+\.mp4)',
+            video_url = url_or_none(file.get('url'))
-            video_url, 'video base url', default=None)
+            if not video_url:
-        if video_url_base:
+                continue
-            # TODO: <Group> found instead of <AdaptationSet> in MPD manifest.
+            ext = determine_ext(video_url, None)
-            # It's not mentioned in MPEG-DASH standard. Figure that out.
+            if ext == 'm3u8':
-            # formats.extend(self._extract_mpd_formats(
+                formats.extend(self._extract_m3u8_formats(
-            #     video_url_base + '.ism/video.mpd',
+                    video_url, video_id, ext='mp4', m3u8_id='hls',
-            #     video_id, mpd_id='dash', fatal=False))
+                    fatal=False, live=is_live))
-            formats.extend(self._extract_ism_formats(
+            elif ext == 'mpd':
-                video_url_base + '.ism/Manifest',
+                if is_live:
-                video_id, 'mss', fatal=False))
+                    # doesn't work with either ffmpeg or native downloader
-            formats.extend(self._extract_f4m_formats(
+                    continue
-                video_url_base + '.ism/video.f4m',
+                formats.extend(self._extract_mpd_formats(video_url, video_id, mpd_id='dash', fatal=False))
-                video_id, f4m_id='hds', fatal=False))
+            elif ext == 'f4m':
-            m3u8_formats = self._extract_m3u8_formats(
+                formats.extend(self._extract_f4m_formats(video_url, video_id, f4m_id='hds', fatal=False))
-                video_url_base + '.ism/video.m3u8', video_id,
+            elif video_url.endswith('.ism/manifest'):
-                'mp4', 'm3u8_native', m3u8_id='hls', fatal=False)
+                formats.extend(self._extract_ism_formats(video_url, video_id, ism_id='mss', fatal=False))
-            self._sort_formats(m3u8_formats)
+            elif ext == 'ism':
-            m3u8_formats = list(filter(
+                if '.ism/manifest' in video_url:
-                lambda f: f.get('vcodec') != 'none', m3u8_formats))
+                    formats.extend(self._extract_ism_formats(video_url, video_id, ism_id='mss', fatal=False))
-            formats.extend(m3u8_formats)
+            else:
-            for i, m3u8_format in enumerate(m3u8_formats, 2):
+                # mp4, wmv or something
-                http_url = '%s-%d.mp4' % (video_url_base, i)
+                quality = traverse_obj(file, 'quality', expected_type=dict) or {}
-                if self._is_valid_url(http_url, video_id):
+                formats.append({
-                    f = m3u8_format.copy()
+                    'format_id': 'direct',
-                    f.update({
+                    'url': video_url,
-                        'url': http_url,
+                    'ext': ext or file.get('type'),
-                        'format_id': f['format_id'].replace('hls', 'http'),
+                    'fps': int_or_none(quality.get('fps')),
-                        'protocol': 'http',
+                    'tbr': int_or_none(quality.get('bitrate'), scale=1000),
-                    })
+                    'width': int_or_none(quality.get('width')),
-                    formats.append(f)
+                    'height': int_or_none(quality.get('height')),
-        else:
+                })
-            formats = [{
-                'format_id': 'direct',
-                'url': video_url,
-                'ext': determine_ext(video_url, 'mp4'),
-            }]
 
         self._sort_formats(formats)
 
-        return {
+        title = traverse_obj(info, 'subtitle', 'title', 'seoTitle', expected_type=txt_or_none)
+        # `seoDescription` may be Falsen
+        description = traverse_obj(info, 'description', 'seoDescription',
+                                   expected_type=lambda x: txt_or_none(x or None))
+        thumbnails = []
+        for thumb in traverse_obj(content, ('posters', Ellipsis), expected_type=dict):
+            thumb_url = thumb.get('src')
+            if not thumb_url or '{width}' in thumb_url or '{height}' in thumb_url:
+                continue
+            thumbnails.append({
+                'url': thumb.get('src'),
+                'width': thumb.get('width'),
+                'height': thumb.get('height'),
+            })
+        age_limit = traverse_obj(info, ('ageGroup', 'minAge'), expected_type=int)
+        if age_limit == 1:
+            age_limit = 0
+        duration = traverse_obj(info, 'duration', expected_type=int) if not is_live else None
+
+        subtitles = {}
+        for sub in traverse_obj(content, ('subtitles', Ellipsis), expected_type=dict):
+            if not (sub.get('url') and sub.get('lang')):
+                continue
+            subtitles.setdefault(sub['lang'], []).append({
+                'url': sub['url'],
+                'ext': sub.get('type'),
+            })
+
+        info_dict = {
             'id': video_id,
             'title': title,
-            'thumbnail': thumbnail,
+            'description': description,
+            'thumbnails': thumbnails,
+            'age_limit': age_limit,
+            'is_live': is_live,
+            'duration': duration,
             'formats': formats,
+            'subtitles': subtitles,
+        }
+
+        # vod.tvp.pl
+        if traverse_obj(info, 'vortalName') == 'vod':
+            info_dict.update({
+                'title': '%s, %s' % (info.get('title'), info.get('subtitle')),
+                'series': info.get('title'),
+                'season': info.get('season'),
+                'episode_number': info.get('episode') or None,
+            })
+
+        return info_dict
+
+
+class TVPVODBaseIE(InfoExtractor):
+    _API_BASE_URL = 'https://vod.tvp.pl/api/products/'
+
+    def _call_api(self, resource, video_id, **kwargs):
+        return self._download_json(
+            self._API_BASE_URL + resource, video_id,
+            query={'lang': 'pl', 'platform': 'BROWSER'}, **kwargs)
+
+    def _parse_video(self, video):
+        video_id = traverse_obj(video, 'externalUid', expected_type=txt_or_none)
+
+        if not video_id:
+            return None
+
+        return {
+            '_type': 'url',
+            'url': 'tvp:' + video_id,
+            'ie_key': TVPEmbedIE.ie_key(),
+            'title': video.get('title'),
+            'description': traverse_obj(video, ('lead', 'description'), expected_type=txt_or_none),
+            'age_limit': int_or_none(video.get('rating')),
+            'duration': int_or_none(video.get('duration')),
         }
 
 
-class TVPWebsiteIE(InfoExtractor):
+class TVPVODVideoIE(TVPVODBaseIE):
-    IE_NAME = 'tvp:series'
+    IE_NAME = 'tvp:vod'
-    _VALID_URL = r'https?://vod\.tvp\.pl/website/(?P<display_id>[^,]+),(?P<id>\d+)'
+    _VALID_URL = r'https?://vod\.tvp\.pl/[a-z\d-]+,\d+/[a-z\d-]+(?<!-odcinki)(?:-odcinki,\d+/odcinek-\d+,S\d+E\d+)?,(?P<id>\d+)(?:\?[^#]+)?(?:#.+)?$'
 
     _TESTS = [{
-        # series
+        'url': 'https://vod.tvp.pl/dla-dzieci,24/laboratorium-alchemika-odcinki,309338/odcinek-24,S01E24,311357',
-        'url': 'https://vod.tvp.pl/website/lzy-cennet,38678312/video',
         'info_dict': {
-            'id': '38678312',
+            'id': '60468609',
-        },
-        'playlist_count': 115,
-    }, {
-        # film
-        'url': 'https://vod.tvp.pl/website/gloria,35139666',
-        'info_dict': {
-            'id': '36637049',
             'ext': 'mp4',
-            'title': 'Gloria, Gloria',
+            'title': 'Laboratorium alchemika, Tusze termiczne. Jak zobaczyć niewidoczne. Odcinek 24',
-        },
+            'description': 'md5:1d4098d3e537092ccbac1abf49b7cd4c',
-        'params': {
+            'duration': 300,
-            'skip_download': True,
+            'episode_number': 24,
+            'episode': 'Episode 24',
+            'age_limit': 0,
+            'series': 'Laboratorium alchemika',
+            'thumbnail': 're:https://.+',
         },
         'add_ie': ['TVPEmbed'],
     }, {
-        'url': 'https://vod.tvp.pl/website/lzy-cennet,38678312',
+        'url': 'https://vod.tvp.pl/filmy-dokumentalne,163/ukrainski-sluga-narodu,339667',
+        'info_dict': {
+            'id': '51640077',
+            'ext': 'mp4',
+            'title': 'Ukraiński sługa narodu, Ukraiński sługa narodu',
+            'series': 'Ukraiński sługa narodu',
+            'description': 'md5:b7940c0a8e439b0c81653a986f544ef3',
+            'age_limit': 12,
+            'duration': 3051,
+            'thumbnail': 're:https://.+',
+        },
+        'add_ie': ['TVPEmbed'],
+    }, {
+        # new URL format
+        'url': 'https://vod.tvp.pl/seriale,18/czas-honoru-odcinki,292065/odcinek-13,S01E13,313867',
+        'md5': 'a21eb0aa862f25414430f15fdfb9e76c',
+        'info_dict': {
+            'id': '194536',
+            'ext': 'mp4',
+            'title': 'Czas honoru, odc. 13 – Władek',
+            'description': 'md5:76649d2014f65c99477be17f23a4dead',
+            'age_limit': 12,
+        },
+        'add_ie': ['TVPEmbed'],
+    }, {
+        'url': 'https://vod.tvp.pl/filmy-fabularne,136/rozlam,390638',
+        'only_matching': True,
+    }]
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
+
+        video = self._parse_video(
+            self._call_api('vods/' + video_id, video_id))
+        if not video:
+            raise ExtractorError('No video data for ' + video_id)
+        return video
+
+
+class TVPVODSeriesIE(TVPVODBaseIE):
+    IE_NAME = 'tvp:vod:series'
+    _VALID_URL = r'''(?x)
+                    https?://vod\.tvp\.pl/
+                        seriale,(?P<cat>\d+)/
+                        (?P<display_id>[^,]+?)(?(cat)-odcinki),(?P<id>\d+)
+                        (?(cat)|(?P<video>/video)?)(?:[#?]|$)
+                '''
+    _VALID_URL = r'https?://vod\.tvp\.pl/(?P<display_id>[a-z\d-]+,\d+)/[a-z\d-]+-odcinki,(?P<id>\d+)(?:\?[^#]+)?(?:#.+)?$'
+
+    _TESTS = [{
+        # series
+        'url': 'https://vod.tvp.pl/seriale,18/ranczo-odcinki,316445',
+        #  series (old) - redirects to home page
+        #  'url': 'https://vod.tvp.pl/website/wspaniale-stulecie,17069012/video',
+        'info_dict': {
+            'id': '316445',
+            'title': 'Ranczo',
+            # 'description': 'md5:a7ccbe1296e6f32425cef17639f1b24b',
+            'age_limit': 12,
+            'categories': ['seriale'],
+        },
+        'playlist_mincount': 129,
+    }, {
+        'url': 'https://vod.tvp.pl/programy,88/rolnik-szuka-zony-odcinki,284514',
+        'only_matching': True,
+    }, {
+        'url': 'https://vod.tvp.pl/dla-dzieci,24/laboratorium-alchemika-odcinki,309338',
         'only_matching': True,
     }]
 
     def _entries(self, display_id, playlist_id):
-        url = 'https://vod.tvp.pl/website/%s,%s/video' % (display_id, playlist_id)
+        season_path = 'vods/serials/%s/seasons' % (playlist_id, )
-        for page_num in itertools.count(1):
+        seasons = self._call_api(
-            page = self._download_webpage(
+            season_path, playlist_id,
-                url, display_id, 'Downloading page %d' % page_num,
+            note='Downloading season list') or []
-                query={'page': page_num})
 
-            video_ids = orderedSet(re.findall(
+        for ii, season in enumerate(seasons, 1):
-                r'<a[^>]+\bhref=["\']/video/%s,[^,]+,(\d+)' % display_id,
+            season_id = traverse_obj(season, 'id', expected_type=txt_or_none)
-                page))
+            if not season_id:
-
+                continue
-            if not video_ids:
+            episodes = self._call_api(
-                break
+                '%s/%s/episodes' % (season_path, season_id), playlist_id,
-
+                note='Downloading episode list (season %d)' % ii)
-            for video_id in video_ids:
+            for episode in episodes or []:
-                yield self.url_result(
+                video_id = traverse_obj(episode, 'externalUid', expected_type=txt_or_none)
-                    'tvp:%s' % video_id, ie=TVPEmbedIE.ie_key(),
+                if video_id:
-                    video_id=video_id)
+                    yield self._parse_video(episode)
 
     def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
+        display_id, playlist_id = self._match_valid_url(url).group('display_id', 'id')
-        display_id, playlist_id = mobj.group('display_id', 'id')
+        metadata = self._call_api(
-        return self.playlist_result(
+            'vods/serials/' + playlist_id, playlist_id,
-            self._entries(display_id, playlist_id), playlist_id)
+            note='Downloading serial metadata') or {}
+
+        pl = self.playlist_result(
+            self._entries(display_id, playlist_id), playlist_id, txt_or_none(metadata.get('title')))
+        pl.update({
+            'description': traverse_obj(metadata, ('description', 'lead'), expected_type=clean_html),
+            'categories': traverse_obj(metadata, ('mainCategory', (None, Ellipsis), 'name'), expected_type=txt_or_none),
+            'age_limit': traverse_obj(metadata, 'rating', expected_type=int),
+        })
+        return pl

youtube_dl/jsinterp.py

@@ -14,6 +14,7 @@ from .utils import (
     remove_quotes,
     unified_timestamp,
     variadic,
+    write_string,
 )
 from .compat import (
     compat_basestring,
@@ -53,15 +54,16 @@ def wraps_op(op):
 
 # NB In principle NaN cannot be checked by membership.
 # Here all NaN values are actually this one, so _NaN is _NaN,
-# although _NaN != _NaN.
+# although _NaN != _NaN. Ditto Infinity.
 
 _NaN = float('nan')
+_Infinity = float('inf')
 
 
 def _js_bit_op(op):
 
     def zeroise(x):
-        return 0 if x in (None, JS_Undefined, _NaN) else x
+        return 0 if x in (None, JS_Undefined, _NaN, _Infinity) else x
 
     @wraps_op(op)
     def wrapped(a, b):
@@ -84,7 +86,7 @@ def _js_arith_op(op):
 def _js_div(a, b):
     if JS_Undefined in (a, b) or not (a or b):
         return _NaN
-    return operator.truediv(a or 0, b) if b else float('inf')
+    return operator.truediv(a or 0, b) if b else _Infinity
 
 
 def _js_mod(a, b):
@@ -220,6 +222,42 @@ class LocalNameSpace(ChainMap):
         return 'LocalNameSpace%s' % (self.maps, )
 
 
+class Debugger(object):
+    ENABLED = False
+
+    @staticmethod
+    def write(*args, **kwargs):
+        level = kwargs.get('level', 100)
+
+        def truncate_string(s, left, right=0):
+            if s is None or len(s) <= left + right:
+                return s
+            return '...'.join((s[:left - 3], s[-right:] if right else ''))
+
+        write_string('[debug] JS: {0}{1}\n'.format(
+            '  ' * (100 - level),
+            ' '.join(truncate_string(compat_str(x), 50, 50) for x in args)))
+
+    @classmethod
+    def wrap_interpreter(cls, f):
+        def interpret_statement(self, stmt, local_vars, allow_recursion, *args, **kwargs):
+            if cls.ENABLED and stmt.strip():
+                cls.write(stmt, level=allow_recursion)
+            try:
+                ret, should_ret = f(self, stmt, local_vars, allow_recursion, *args, **kwargs)
+            except Exception as e:
+                if cls.ENABLED:
+                    if isinstance(e, ExtractorError):
+                        e = e.orig_msg
+                    cls.write('=> Raises:', e, '<-|', stmt, level=allow_recursion)
+                raise
+            if cls.ENABLED and stmt.strip():
+                if should_ret or not repr(ret) == stmt:
+                    cls.write(['->', '=>'][should_ret], repr(ret), '<-|', stmt, level=allow_recursion)
+            return ret, should_ret
+        return interpret_statement
+
+
 class JSInterpreter(object):
     __named_object_counter = 0
 
@@ -307,8 +345,7 @@ class JSInterpreter(object):
     def __op_chars(cls):
         op_chars = set(';,[')
         for op in cls._all_operators():
-            for c in op[0]:
+            op_chars.update(op[0])
-                op_chars.add(c)
         return op_chars
 
     def _named_object(self, namespace, obj):
@@ -326,9 +363,8 @@ class JSInterpreter(object):
         # collections.Counter() is ~10% slower in both 2.7 and 3.9
         counters = dict((k, 0) for k in _MATCHING_PARENS.values())
         start, splits, pos, delim_len = 0, 0, 0, len(delim) - 1
-        in_quote, escaping, skipping = None, False, 0
+        in_quote, escaping, after_op, in_regex_char_group = None, False, True, False
-        after_op, in_regex_char_group = True, False
+        skipping = 0
-
         for idx, char in enumerate(expr):
             paren_delta = 0
             if not in_quote:
@@ -382,10 +418,12 @@ class JSInterpreter(object):
         return separated[0][1:].strip(), separated[1].strip()
 
     @staticmethod
-    def _all_operators():
+    def _all_operators(_cached=[]):
-        return itertools.chain(
+        if not _cached:
-            # Ref: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Operator_Precedence
+            _cached.extend(itertools.chain(
-            _SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS)
+                # Ref: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Operator_Precedence
+                _SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS))
+        return _cached
 
     def _operator(self, op, left_val, right_expr, expr, local_vars, allow_recursion):
         if op in ('||', '&&'):
@@ -416,7 +454,7 @@ class JSInterpreter(object):
         except Exception as e:
             if allow_undefined:
                 return JS_Undefined
-            raise self.Exception('Cannot get index {idx:.100}'.format(**locals()), expr=repr(obj), cause=e)
+            raise self.Exception('Cannot get index {idx!r:.100}'.format(**locals()), expr=repr(obj), cause=e)
 
     def _dump(self, obj, namespace):
         try:
@@ -438,6 +476,7 @@ class JSInterpreter(object):
     _FINALLY_RE = re.compile(r'finally\s*\{')
     _SWITCH_RE = re.compile(r'switch\s*\(')
 
+    @Debugger.wrap_interpreter
     def interpret_statement(self, stmt, local_vars, allow_recursion=100):
         if allow_recursion < 0:
             raise self.Exception('Recursion limit reached')
@@ -511,7 +550,6 @@ class JSInterpreter(object):
                 expr = self._dump(inner, local_vars) + outer
 
         if expr.startswith('('):
-
             m = re.match(r'\((?P<d>[a-z])%(?P<e>[a-z])\.length\+(?P=e)\.length\)%(?P=e)\.length', expr)
             if m:
                 # short-cut eval of frequently used `(d%e.length+e.length)%e.length`, worth ~6% on `pytest -k test_nsig`
@@ -693,7 +731,7 @@ class JSInterpreter(object):
                 (?P<op>{_OPERATOR_RE})?
                 =(?!=)(?P<expr>.*)$
             )|(?P<return>
-                (?!if|return|true|false|null|undefined)(?P<name>{_NAME_RE})$
+                (?!if|return|true|false|null|undefined|NaN|Infinity)(?P<name>{_NAME_RE})$
             )|(?P<indexing>
                 (?P<in>{_NAME_RE})\[(?P<idx>.+)\]$
             )|(?P<attribute>
@@ -727,11 +765,12 @@ class JSInterpreter(object):
             raise JS_Break()
         elif expr == 'continue':
             raise JS_Continue()
-
         elif expr == 'undefined':
             return JS_Undefined, should_return
         elif expr == 'NaN':
             return _NaN, should_return
+        elif expr == 'Infinity':
+            return _Infinity, should_return
 
         elif md.get('return'):
             return local_vars[m.group('name')], should_return
@@ -760,18 +799,28 @@ class JSInterpreter(object):
             right_expr = separated.pop()
             # handle operators that are both unary and binary, minimal BODMAS
             if op in ('+', '-'):
+                # simplify/adjust consecutive instances of these operators
                 undone = 0
                 while len(separated) > 1 and not separated[-1].strip():
                     undone += 1
                     separated.pop()
                 if op == '-' and undone % 2 != 0:
                     right_expr = op + right_expr
+                elif op == '+':
+                    while len(separated) > 1 and separated[-1].strip() in self.OP_CHARS:
+                        right_expr = separated.pop() + right_expr
+                # hanging op at end of left => unary + (strip) or - (push right)
                 left_val = separated[-1]
                 for dm_op in ('*', '%', '/', '**'):
                     bodmas = tuple(self._separate(left_val, dm_op, skip_delims=skip_delim))
                     if len(bodmas) > 1 and not bodmas[-1].strip():
                         expr = op.join(separated) + op + right_expr
-                        right_expr = None
+                        if len(separated) > 1:
+                            separated.pop()
+                            right_expr = op.join((left_val, right_expr))
+                        else:
+                            separated = [op.join((left_val, right_expr))]
+                            right_expr = None
                         break
                 if right_expr is None:
                     continue
@@ -797,6 +846,8 @@ class JSInterpreter(object):
 
             def eval_method():
                 if (variable, member) == ('console', 'debug'):
+                    if Debugger.ENABLED:
+                        Debugger.write(self.interpret_expression('[{}]'.format(arg_str), local_vars, allow_recursion))
                     return
                 types = {
                     'String': compat_str,

youtube_dl/utils.py

@@ -2406,7 +2406,7 @@ class ExtractorError(YoutubeDLError):
         """ tb, if given, is the original traceback (so that it can be printed out).
         If expected is set, this is a normal error message and most likely not a bug in youtube-dl.
         """
-
+        self.orig_msg = msg
         if sys.exc_info()[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError):
             expected = True
         if video_id is not None: