Merge a53b4004cf into 0153b387e5

* [VidLii] Add HD support  (yt-dlp backport-ish)

* Also fix a bug with the view count

---------

Co-authored-by: dirkf <fieldhouse@gmx.net>

test/helper.py

@@ -5,9 +5,9 @@ import hashlib
 import json
 import os.path
 import re
-import types
 import ssl
 import sys
+import types
 import unittest
 
 import youtube_dl.extractor
@@ -181,18 +181,18 @@ def expect_value(self, got, expected, field):
             op, _, expected_num = expected.partition(':')
             expected_num = int(expected_num)
             if op == 'mincount':
-                assert_func = assertGreaterEqual
+                assert_func = self.assertGreaterEqual
                 msg_tmpl = 'Expected %d items in field %s, but only got %d'
             elif op == 'maxcount':
-                assert_func = assertLessEqual
+                assert_func = self.assertLessEqual
                 msg_tmpl = 'Expected maximum %d items in field %s, but got %d'
             elif op == 'count':
-                assert_func = assertEqual
+                assert_func = self.assertEqual
                 msg_tmpl = 'Expected exactly %d items in field %s, but got %d'
             else:
                 assert False
             assert_func(
-                self, len(got), expected_num,
+                len(got), expected_num,
                 msg_tmpl % (expected_num, field, len(got)))
             return
         self.assertEqual(
@@ -262,27 +262,6 @@ def assertRegexpMatches(self, text, regexp, msg=None):
             self.assertTrue(m, msg)
 
 
-def assertGreaterEqual(self, got, expected, msg=None):
-    if not (got >= expected):
-        if msg is None:
-            msg = '%r not greater than or equal to %r' % (got, expected)
-        self.assertTrue(got >= expected, msg)
-
-
-def assertLessEqual(self, got, expected, msg=None):
-    if not (got <= expected):
-        if msg is None:
-            msg = '%r not less than or equal to %r' % (got, expected)
-        self.assertTrue(got <= expected, msg)
-
-
-def assertEqual(self, got, expected, msg=None):
-    if not (got == expected):
-        if msg is None:
-            msg = '%r not equal to %r' % (got, expected)
-        self.assertTrue(got == expected, msg)
-
-
 def expect_warnings(ydl, warnings_re):
     real_warning = ydl.report_warning
 

test/test_download.py

@@ -9,8 +9,6 @@ import unittest
 sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
 
 from test.helper import (
-    assertGreaterEqual,
-    assertLessEqual,
     expect_warnings,
     get_params,
     gettestcases,
@@ -36,12 +34,20 @@ from youtube_dl.utils import (
     ExtractorError,
     error_to_compat_str,
     format_bytes,
+    IDENTITY,
+    preferredencoding,
     UnavailableVideoError,
 )
 from youtube_dl.extractor import get_info_extractor
 
 RETRIES = 3
 
+# Some unittest APIs require actual str
+if not isinstance('TEST', str):
+    _encode_str = lambda s: s.encode(preferredencoding())
+else:
+    _encode_str = IDENTITY
+
 
 class YoutubeDL(youtube_dl.YoutubeDL):
     def __init__(self, *args, **kwargs):
@@ -102,7 +108,7 @@ def generator(test_case, tname):
 
         def print_skipping(reason):
             print('Skipping %s: %s' % (test_case['name'], reason))
-            self.skipTest(reason)
+            self.skipTest(_encode_str(reason))
 
         if not ie.working():
             print_skipping('IE marked as not _WORKING')
@@ -187,16 +193,14 @@ def generator(test_case, tname):
                 expect_info_dict(self, res_dict, test_case.get('info_dict', {}))
 
             if 'playlist_mincount' in test_case:
-                assertGreaterEqual(
-                    self,
+                self.assertGreaterEqual(
                     len(res_dict['entries']),
                     test_case['playlist_mincount'],
                     'Expected at least %d in playlist %s, but got only %d' % (
                         test_case['playlist_mincount'], test_case['url'],
                         len(res_dict['entries'])))
             if 'playlist_maxcount' in test_case:
-                assertLessEqual(
-                    self,
+                self.assertLessEqual(
                     len(res_dict['entries']),
                     test_case['playlist_maxcount'],
                     'Expected at most %d in playlist %s, but got %d' % (
@@ -243,8 +247,8 @@ def generator(test_case, tname):
                         if params.get('test'):
                             expected_minsize = max(expected_minsize, 10000)
                         got_fsize = os.path.getsize(tc_filename)
-                        assertGreaterEqual(
-                            self, got_fsize, expected_minsize,
+                        self.assertGreaterEqual(
+                            got_fsize, expected_minsize,
                             'Expected %s to be at least %s, but it\'s only %s ' %
                             (tc_filename, format_bytes(expected_minsize),
                                 format_bytes(got_fsize)))

youtube_dl/YoutubeDL.py

@@ -1039,8 +1039,8 @@ class YoutubeDL(object):
         elif result_type in ('playlist', 'multi_video'):
             # Protect from infinite recursion due to recursively nested playlists
             # (see https://github.com/ytdl-org/youtube-dl/issues/27833)
-            webpage_url = ie_result['webpage_url']
-            if webpage_url in self._playlist_urls:
+            webpage_url = ie_result.get('webpage_url')  # not all pl/mv have this
+            if webpage_url and webpage_url in self._playlist_urls:
                 self.to_screen(
                     '[download] Skipping already downloaded playlist: %s'
                     % ie_result.get('title') or ie_result.get('id'))
@@ -1048,6 +1048,10 @@ class YoutubeDL(object):
 
             self._playlist_level += 1
             self._playlist_urls.add(webpage_url)
+            new_result = dict((k, v) for k, v in extra_info.items() if k not in ie_result)
+            if new_result:
+                new_result.update(ie_result)
+                ie_result = new_result
             try:
                 return self.__process_playlist(ie_result, download)
             finally:
@@ -1593,6 +1597,28 @@ class YoutubeDL(object):
         self.cookiejar.add_cookie_header(pr)
         return pr.get_header('Cookie')
 
+    def _fill_common_fields(self, info_dict, final=True):
+
+        for ts_key, date_key in (
+                ('timestamp', 'upload_date'),
+                ('release_timestamp', 'release_date'),
+        ):
+            if info_dict.get(date_key) is None and info_dict.get(ts_key) is not None:
+                # Working around out-of-range timestamp values (e.g. negative ones on Windows,
+                # see http://bugs.python.org/issue1646728)
+                try:
+                    upload_date = datetime.datetime.utcfromtimestamp(info_dict[ts_key])
+                    info_dict[date_key] = compat_str(upload_date.strftime('%Y%m%d'))
+                except (ValueError, OverflowError, OSError):
+                    pass
+
+        # Auto generate title fields corresponding to the *_number fields when missing
+        # in order to always have clean titles. This is very common for TV series.
+        if final:
+            for field in ('chapter', 'season', 'episode'):
+                if info_dict.get('%s_number' % field) is not None and not info_dict.get(field):
+                    info_dict[field] = '%s %d' % (field.capitalize(), info_dict['%s_number' % field])
+
     def process_video_result(self, info_dict, download=True):
         assert info_dict.get('_type', 'video') == 'video'
 
@@ -1660,24 +1686,7 @@ class YoutubeDL(object):
         if 'display_id' not in info_dict and 'id' in info_dict:
             info_dict['display_id'] = info_dict['id']
 
-        for ts_key, date_key in (
-                ('timestamp', 'upload_date'),
-                ('release_timestamp', 'release_date'),
-        ):
-            if info_dict.get(date_key) is None and info_dict.get(ts_key) is not None:
-                # Working around out-of-range timestamp values (e.g. negative ones on Windows,
-                # see http://bugs.python.org/issue1646728)
-                try:
-                    upload_date = datetime.datetime.utcfromtimestamp(info_dict[ts_key])
-                    info_dict[date_key] = compat_str(upload_date.strftime('%Y%m%d'))
-                except (ValueError, OverflowError, OSError):
-                    pass
-
-        # Auto generate title fields corresponding to the *_number fields when missing
-        # in order to always have clean titles. This is very common for TV series.
-        for field in ('chapter', 'season', 'episode'):
-            if info_dict.get('%s_number' % field) is not None and not info_dict.get(field):
-                info_dict[field] = '%s %d' % (field.capitalize(), info_dict['%s_number' % field])
+        self._fill_common_fields(info_dict)
 
         for cc_kind in ('subtitles', 'automatic_captions'):
             cc = info_dict.get(cc_kind)

youtube_dl/extractor/extractors.py

@@ -898,21 +898,13 @@ from .ooyala import (
 )
 from .ora import OraTVIE
 from .orf import (
-    ORFTVthekIE,
-    ORFFM4IE,
+    ORFONIE,
+    ORFONLiveIE,
     ORFFM4StoryIE,
-    ORFOE1IE,
-    ORFOE3IE,
-    ORFNOEIE,
-    ORFWIEIE,
-    ORFBGLIE,
-    ORFOOEIE,
-    ORFSTMIE,
-    ORFKTNIE,
-    ORFSBGIE,
-    ORFTIRIE,
-    ORFVBGIE,
     ORFIPTVIE,
+    ORFPodcastIE,
+    ORFRadioIE,
+    ORFRadioCollectionIE,
 )
 from .outsidetv import OutsideTVIE
 from .packtpub import (
@@ -1246,10 +1238,7 @@ from .svt import (
 from .swrmediathek import SWRMediathekIE
 from .syfy import SyfyIE
 from .sztvhu import SztvHuIE
-from .tagesschau import (
-    TagesschauPlayerIE,
-    TagesschauIE,
-)
+from .tagesschau import TagesschauIE
 from .tass import TassIE
 from .tbs import TBSIE
 from .tdslifeway import TDSLifewayIE

youtube_dl/extractor/mixcloud.py

@@ -1,3 +1,4 @@
+# coding: utf-8
 from __future__ import unicode_literals
 
 import itertools
@@ -10,7 +11,7 @@ from ..compat import (
     compat_ord,
     compat_str,
     compat_urllib_parse_unquote,
-    compat_zip
+    compat_zip as zip,
 )
 from ..utils import (
     int_or_none,
@@ -24,7 +25,7 @@ class MixcloudBaseIE(InfoExtractor):
     def _call_api(self, object_type, object_fields, display_id, username, slug=None):
         lookup_key = object_type + 'Lookup'
         return self._download_json(
-            'https://www.mixcloud.com/graphql', display_id, query={
+            'https://app.mixcloud.com/graphql', display_id, query={
                 'query': '''{
   %s(lookup: {username: "%s"%s}) {
     %s
@@ -44,7 +45,7 @@ class MixcloudIE(MixcloudBaseIE):
             'ext': 'm4a',
             'title': 'Cryptkeeper',
             'description': 'After quite a long silence from myself, finally another Drum\'n\'Bass mix with my favourite current dance floor bangers.',
-            'uploader': 'Daniel Holbach',
+            'uploader': 'dholbach',  # was: 'Daniel Holbach',
             'uploader_id': 'dholbach',
             'thumbnail': r're:https?://.*\.jpg',
             'view_count': int,
@@ -57,7 +58,7 @@ class MixcloudIE(MixcloudBaseIE):
             'id': 'gillespeterson_caribou-7-inch-vinyl-mix-chat',
             'ext': 'mp3',
             'title': 'Caribou 7 inch Vinyl Mix & Chat',
-            'description': 'md5:2b8aec6adce69f9d41724647c65875e8',
+            'description': r're:Last week Dan Snaith aka Caribou swung by the Brownswood.{136}',
             'uploader': 'Gilles Peterson Worldwide',
             'uploader_id': 'gillespeterson',
             'thumbnail': 're:https?://.*',
@@ -65,6 +66,23 @@ class MixcloudIE(MixcloudBaseIE):
             'timestamp': 1422987057,
             'upload_date': '20150203',
         },
+        'params': {
+            'skip_download': '404 not found',
+        },
+    }, {
+        'url': 'https://www.mixcloud.com/gillespeterson/carnival-m%C3%BAsica-popular-brasileira-mix/',
+        'info_dict': {
+            'id': 'gillespeterson_carnival-música-popular-brasileira-mix',
+            'ext': 'm4a',
+            'title': 'Carnival Música Popular Brasileira Mix',
+            'description': r're:Gilles was recently in Brazil to play at Boiler Room.{208}',
+            'timestamp': 1454347174,
+            'upload_date': '20160201',
+            'uploader': 'Gilles Peterson Worldwide',
+            'uploader_id': 'gillespeterson',
+            'thumbnail': 're:https?://.*',
+            'view_count': int,
+        },
     }, {
         'url': 'https://beta.mixcloud.com/RedLightRadio/nosedrip-15-red-light-radio-01-18-2016/',
         'only_matching': True,
@@ -76,10 +94,10 @@ class MixcloudIE(MixcloudBaseIE):
         """Encrypt/Decrypt XOR cipher. Both ways are possible because it's XOR."""
         return ''.join([
             compat_chr(compat_ord(ch) ^ compat_ord(k))
-            for ch, k in compat_zip(ciphertext, itertools.cycle(key))])
+            for ch, k in zip(ciphertext, itertools.cycle(key))])
 
     def _real_extract(self, url):
-        username, slug = re.match(self._VALID_URL, url).groups()
+        username, slug = self._match_valid_url(url).groups()
         username, slug = compat_urllib_parse_unquote(username), compat_urllib_parse_unquote(slug)
         track_id = '%s_%s' % (username, slug)
 

youtube_dl/extractor/tagesschau.py

@@ -5,127 +5,54 @@ import re
 
 from .common import InfoExtractor
 from ..utils import (
+    bool_or_none,
+    compat_str,
     determine_ext,
-    js_to_json,
-    parse_iso8601,
+    ExtractorError,
+    parse_duration,
     parse_filesize,
+    remove_quotes,
+    strip_or_none,
+    try_get,
+    unescapeHTML,
+    unified_timestamp,
+    url_or_none,
 )
 
+# Note that there are tagesschau.de/api and tagesschau.de/api2 endpoints, which
+# may be useful, but not all pages and not all formats can be easily accessed
+# by API.
 
-class TagesschauPlayerIE(InfoExtractor):
-    IE_NAME = 'tagesschau:player'
-    _VALID_URL = r'https?://(?:www\.)?tagesschau\.de/multimedia/(?P<kind>audio|video)/(?P=kind)-(?P<id>\d+)~player(?:_[^/?#&]+)?\.html'
 
-    _TESTS = [{
-        'url': 'http://www.tagesschau.de/multimedia/video/video-179517~player.html',
-        'md5': '8d09548d5c15debad38bee3a4d15ca21',
-        'info_dict': {
-            'id': '179517',
-            'ext': 'mp4',
-            'title': 'Marie Kristin Boese, ARD Berlin, über den zukünftigen Kurs der AfD',
-            'thumbnail': r're:^https?:.*\.jpg$',
-            'formats': 'mincount:6',
-        },
-    }, {
-        'url': 'https://www.tagesschau.de/multimedia/audio/audio-29417~player.html',
-        'md5': '76e6eec6ebd40740671cf0a2c88617e5',
-        'info_dict': {
-            'id': '29417',
-            'ext': 'mp3',
-            'title': 'Trabi - Bye, bye Rennpappe',
-            'thumbnail': r're:^https?:.*\.jpg$',
-            'formats': 'mincount:2',
-        },
-    }, {
-        'url': 'http://www.tagesschau.de/multimedia/audio/audio-29417~player_autoplay-true.html',
-        'only_matching': True,
-    }]
+_FORMATS = {
+    'xs': {'quality': 0},
+    's': {'width': 320, 'height': 180, 'quality': 1},
+    'sm': {'width': 480, 'height': 270, 'quality': 1},
+    'm': {'width': 512, 'height': 288, 'quality': 2},
+    'ml': {'width': 640, 'height': 360, 'quality': 2},
+    'l': {'width': 960, 'height': 540, 'quality': 3},
+    'xl': {'width': 1280, 'height': 720, 'quality': 4},
+    'xxl': {'quality': 5},
+    'mp3': {'abr': 64, 'vcodec': 'none', 'quality': 0},
+    'hi.mp3': {'abr': 192, 'vcodec': 'none', 'quality': 1},
+}
 
-    _FORMATS = {
-        'xs': {'quality': 0},
-        's': {'width': 320, 'height': 180, 'quality': 1},
-        'm': {'width': 512, 'height': 288, 'quality': 2},
-        'l': {'width': 960, 'height': 540, 'quality': 3},
-        'xl': {'width': 1280, 'height': 720, 'quality': 4},
-        'xxl': {'quality': 5},
-    }
+_FIELD_PREFERENCE = ('height', 'width', 'vbr', 'abr')
 
-    def _extract_via_api(self, kind, video_id):
-        info = self._download_json(
-            'https://www.tagesschau.de/api/multimedia/{0}/{0}-{1}.json'.format(kind, video_id),
-            video_id)
-        title = info['headline']
-        formats = []
-        for media in info['mediadata']:
-            for format_id, format_url in media.items():
-                if determine_ext(format_url) == 'm3u8':
-                    formats.extend(self._extract_m3u8_formats(
-                        format_url, video_id, 'mp4',
-                        entry_protocol='m3u8_native', m3u8_id='hls'))
-                else:
-                    formats.append({
-                        'url': format_url,
-                        'format_id': format_id,
-                        'vcodec': 'none' if kind == 'audio' else None,
-                    })
-        self._sort_formats(formats)
-        timestamp = parse_iso8601(info.get('date'))
-        return {
-            'id': video_id,
-            'title': title,
-            'timestamp': timestamp,
-            'formats': formats,
-        }
 
-    def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
-        video_id = mobj.group('id')
-
-        # kind = mobj.group('kind').lower()
-        # if kind == 'video':
-        #     return self._extract_via_api(kind, video_id)
-
-        # JSON api does not provide some audio formats (e.g. ogg) thus
-        # extracting audio via webpage
-
-        webpage = self._download_webpage(url, video_id)
-
-        title = self._og_search_title(webpage).strip()
-        formats = []
-
-        for media_json in re.findall(r'({src\s*:\s*["\']http[^}]+type\s*:[^}]+})', webpage):
-            media = self._parse_json(js_to_json(media_json), video_id, fatal=False)
-            if not media:
-                continue
-            src = media.get('src')
-            if not src:
-                return
-            quality = media.get('quality')
-            kind = media.get('type', '').split('/')[0]
-            ext = determine_ext(src)
-            f = {
-                'url': src,
-                'format_id': '%s_%s' % (quality, ext) if quality else ext,
-                'ext': ext,
-                'vcodec': 'none' if kind == 'audio' else None,
-            }
-            f.update(self._FORMATS.get(quality, {}))
-            formats.append(f)
-
-        self._sort_formats(formats)
-
-        thumbnail = self._og_search_thumbnail(webpage)
-
-        return {
-            'id': video_id,
-            'title': title,
-            'thumbnail': thumbnail,
-            'formats': formats,
-        }
+def _normalize_format_id(format_id, ext):
+    if format_id:
+        m = re.match(r"web([^.]+)\.[^.]+$", format_id)
+        if m:
+            format_id = m.group(1)
+        if format_id == 'hi' and ext:
+            # high-quality audio files
+            format_id = '%s.%s' % (format_id, ext)
+    return format_id
 
 
 class TagesschauIE(InfoExtractor):
-    _VALID_URL = r'https?://(?:www\.)?tagesschau\.de/(?P<path>[^/]+/(?:[^/]+/)*?(?P<id>[^/#?]+?(?:-?[0-9]+)?))(?:~_?[^/#?]+?)?\.html'
+    _VALID_URL = r'https?://(?:www\.)?tagesschau\.de(?:/?|/(?P<path>[^?#]+?(?:/(?P<id>[^/#?]+?(?:-?[0-9]+))(?:~_?[^/#?]+?)?(?:\.html)?)?))(?:[#?].*)?$'
 
     _TESTS = [{
         'url': 'http://www.tagesschau.de/multimedia/video/video-102143.html',
@@ -134,48 +61,111 @@ class TagesschauIE(InfoExtractor):
             'id': 'video-102143',
             'ext': 'mp4',
             'title': 'Regierungsumbildung in Athen: Neue Minister in Griechenland vereidigt',
-            'description': '18.07.2015 20:10 Uhr',
+            'description': '18.07.2015 20:10',
             'thumbnail': r're:^https?:.*\.jpg$',
+            'upload_date': '20150718',
+            'duration': 138,
+            'timestamp': 1437250200,
+            'uploader': 'ARD',
+        },
+    }, {
+        # with player
+        'url': 'http://www.tagesschau.de/multimedia/video/video-102143~player.html',
+        'md5': 'f7c27a0eff3bfe8c7727e65f8fe1b1e6',
+        'info_dict': {
+            'id': 'video-102143',
+            'ext': 'mp4',
+            'title': 'Regierungsumbildung in Athen: Neue Minister in Griechenland vereidigt',
+            'description': '18.07.2015 20:10',
+            'thumbnail': r're:^https?:.*\.jpg$',
+            'upload_date': '20150718',
+            'timestamp': 1437250200,
+            'uploader': 'ARD',
         },
     }, {
         'url': 'http://www.tagesschau.de/multimedia/sendung/ts-5727.html',
         'md5': '3c54c1f6243d279b706bde660ceec633',
         'info_dict': {
-            'id': 'ts-5727',
+            'id': 'video-45741',
             'ext': 'mp4',
-            'title': 'Sendung: tagesschau \t04.12.2014 20:00 Uhr',
-            'description': 'md5:695c01bfd98b7e313c501386327aea59',
+            'title': 'tagesschau 20 Uhr - 04.12.14 20:00',
+            'description': '04.12.2014 20:00',
             'thumbnail': r're:^https?:.*\.jpg$',
+            'uploader': 'tagesschau',
+            'timestamp': 1417723200,
+            'upload_date': '20141204',
+            'subtitles': dict,
         },
     }, {
         # exclusive audio
-        'url': 'http://www.tagesschau.de/multimedia/audio/audio-29417.html',
-        'md5': '76e6eec6ebd40740671cf0a2c88617e5',
+        'url': 'https://www.tagesschau.de/multimedia/audio/audio-103205.html',
+        'md5': 'c8e7b72aeca664031db0ba198519b09a',
         'info_dict': {
-            'id': 'audio-29417',
+            'id': 'audio-103205',
             'ext': 'mp3',
-            'title': 'Trabi - Bye, bye Rennpappe',
-            'description': 'md5:8687dda862cbbe2cfb2df09b56341317',
+            'title': 'Die USA: ein Impfwunder?',
+            'description': '06.03.2021 06:07',
+            'timestamp': 1615010820,
+            'upload_date': '20210306',
             'thumbnail': r're:^https?:.*\.jpg$',
+            'uploader': 'Jule Käppel, ARD Washington',
+            'creator': 'ARD',
+            'channel': 'tagesschau.de',
+            'is_live': False,
         },
     }, {
         # audio in article
-        'url': 'http://www.tagesschau.de/inland/bnd-303.html',
-        'md5': 'e0916c623e85fc1d2b26b78f299d3958',
+        'url': 'https://www.tagesschau.de/ausland/amerika/biden-versoehnung-101.html',
+        'md5': '4c46b0283719d97aa976037e1ecb7b73',
         'info_dict': {
-            'id': 'bnd-303',
+            'id': 'audio-103429',
+            'title': 'Bidens Versöhnungswerk kommt nicht voran',
             'ext': 'mp3',
-            'title': 'Viele Baustellen für neuen BND-Chef',
-            'description': 'md5:1e69a54be3e1255b2b07cdbce5bcd8b4',
-            'thumbnail': r're:^https?:.*\.jpg$',
+            'timestamp': 1615444860,
+            'uploader': 'Sebastian Hesse, ARD Washington',
+            'description': '11.03.2021 06:41',
+            'upload_date': '20210311',
+            'creator': 'ARD',
+            'channel': 'tagesschau.de',
         },
     }, {
-        'url': 'http://www.tagesschau.de/inland/afd-parteitag-135.html',
+        # playlist in article
+        'url': 'https://www.tagesschau.de/ausland/impfungen-coronavirus-usa-101.html',
         'info_dict': {
-            'id': 'afd-parteitag-135',
-            'title': 'Möchtegern-Underdog mit Machtanspruch',
+            'id': 'impfungen-coronavirus-usa-101',
+            'title': 'Kampf gegen das Coronavirus: Impfwunder USA?',
         },
         'playlist_count': 2,
+    }, {
+        # article without videos
+        'url': 'https://www.tagesschau.de/wirtschaft/ukraine-russland-kredit-101.html',
+        'info_dict': {
+            'id': 'ukraine-russland-kredit-101',
+            'title': 'Ukraine stoppt Rückzahlung russischer Kredite',
+        },
+        'playlist_count': 0,
+    }, {
+        # legacy website
+        'url': 'https://www.tagesschau.de/multimedia/video/video-102303~_bab-sendung-211.html',
+        'md5': 'ab6d190c8147560d6429a467566affe6',
+        'info_dict': {
+            'id': 'video-102303',
+            'ext': 'mp4',
+            'title': 'Bericht aus Berlin: Sommerinterview mit Angela Merkel',
+            'description': '19.07.2015 19:05 Uhr',
+        }
+    }, {
+        # handling of generic title
+        'url': 'https://www.tagesschau.de/multimedia/video/video-835681.html',
+        'info_dict': {
+            'id': 'video-835681',
+            'ext': 'mp4',
+            'title': 'Tagesschau in 100 Sekunden - 13.03.21 17:35',
+            'upload_date': '20210313',
+            'uploader': 'Tagesschau24',
+            'description': '13.03.2021 17:35',
+            'timestamp': 1615656900,
+        }
     }, {
         'url': 'http://www.tagesschau.de/multimedia/sendung/tsg-3771.html',
         'only_matching': True,
@@ -204,13 +194,176 @@ class TagesschauIE(InfoExtractor):
         # playlist article with collapsing sections
         'url': 'http://www.tagesschau.de/wirtschaft/faq-freihandelszone-eu-usa-101.html',
         'only_matching': True,
+    }, {
+        'url': 'https://www.tagesschau.de/',
+        'only_matching': True,
     }]
 
-    @classmethod
-    def suitable(cls, url):
-        return False if TagesschauPlayerIE.suitable(url) else super(TagesschauIE, cls).suitable(url)
+    def _video_id_from_url(self, url):
+        if url:
+            mobj = re.match(self._VALID_URL, url)
+            if mobj:
+                return mobj.group('id')
 
-    def _extract_formats(self, download_text, media_kind):
+    def _handle_generic_titles(self, title, pixelConf):
+        if strip_or_none(title, '').lower() not in ('ganze sendung', '100 sekunden',
+                                                    'tagesschau in 100 sekunden'):
+            return title
+        # otherwise find more meaningful title than the generic Ganze Sendung/100 Sekunden
+        for item in pixelConf:
+            if item.get('tracker') == 'AGFdebug':
+                s = try_get(item, lambda x: x['clipData']['program'], compat_str)
+                if s:
+                    # extract date and time
+                    parts = (try_get(item, lambda x: x['clipData']['title'], compat_str)
+                             or '').split('_')[-2:]
+                    if len(parts) == 2:
+                        title = "%s - %s" % (s, ' '.join(parts))
+                    else:
+                        title = s
+                break
+        return title
+
+    def _extract_from_player(self, player_div, video_id_fallback, title_fallback):
+        player_data = unescapeHTML(self._search_regex(
+            r'data-config=(?P<quote>["\'])(?P<data>[^"\']*)(?P=quote)',
+            player_div, 'data-config', group='data'))
+
+        meta = self._parse_json(player_data, video_id_fallback, fatal=False)
+        mc = try_get(meta, lambda x: x['mc'], dict)
+        if not mc:
+            # fallback if parsing json fails, as tagesschau API sometimes sends
+            # invalid json
+            stream_hls = remove_quotes(self._search_regex(
+                r'"http[^"]+?\.m3u8"', player_data, '.m3u8-url', group=0))
+            formats = self._extract_m3u8_formats(stream_hls, video_id_fallback,
+                                                 ext='mp4', m3u8_id='hls',
+                                                 entry_protocol='m3u8_native')
+            self._sort_formats(formats, field_preference=_FIELD_PREFERENCE)
+            return {
+                'id': video_id_fallback,
+                'title': title_fallback,
+                'formats': formats,
+            }
+
+        # this url is more permanent than the original link
+        webpage_url = url_or_none(try_get(mc, lambda x: x['_sharing']['link']))
+
+        video_id = self._video_id_from_url(webpage_url)
+        duration = None
+        pixelConf = try_get(meta, lambda x: x['pc']['_pixelConfig'], list) or []
+        for item in pixelConf:
+            video_id = (video_id or try_get(item,
+                        [lambda x: x['playerID'],
+                         lambda x: x['clipData']['playerId']], compat_str))
+            duration = (duration or parse_duration(try_get(item,
+                        [lambda x: x['clipData']['length'],
+                         lambda x: x['clipData']['duration']])))
+        if not video_id:
+            video_id = video_id_fallback
+
+        formats = []
+        for elem in mc.get('_mediaArray', []):
+            for d in elem.get('_mediaStreamArray', []):
+                link_url = url_or_none(d.get('_stream'))
+                if not link_url:
+                    continue
+                ext = determine_ext(link_url)
+                if ext == "m3u8":
+                    formats.extend(self._extract_m3u8_formats(
+                        link_url, video_id_fallback, ext='mp4',
+                        entry_protocol='m3u8_native',
+                        m3u8_id='hls', fatal=False))
+                elif ext == "f4m":
+                    formats.extend(self._extract_f4m_formats(
+                        link_url, video_id_fallback, f4m_id='hds', fatal=False))
+                else:
+                    format_id = _normalize_format_id(self._search_regex(
+                        r'.*/[^/.]+\.([^/]+)\.[^/.]+$', link_url, 'format ID',
+                        default=ext, fatal=False),
+                        ext)
+                    fmt = {
+                        'format_id': format_id,
+                        'url': link_url,
+                        'format_name': ext,
+                    }
+                    fmt.update(_FORMATS.get(format_id, {}))
+                    formats.append(fmt)
+        self._sort_formats(formats, field_preference=_FIELD_PREFERENCE)
+        if not formats:
+            raise ExtractorError("could not extract formats from json")
+
+        # note that mc['_title'] can be very different from actual title,
+        # such as an image description in case of audio files
+        title = (try_get(mc, [lambda x: x['_info']['clipTitle'],
+                              lambda x: x['_download']['title']], compat_str)
+                 or title_fallback)
+        title = self._handle_generic_titles(title, pixelConf)
+
+        sub_url = url_or_none(mc.get('_subtitleUrl'))
+        subs = {'de': [{'ext': 'ttml', 'url': sub_url}]} if sub_url else None
+
+        images = try_get(mc, lambda x: x['_previewImage'], dict) or {}
+        thumbnails = [{
+            'url': url_or_none('https://www.tagesschau.de/%s'
+                               % (images[format_id],)),
+            'preference': _FORMATS.get(format_id, {}).get('quality'),
+        } for format_id in images] or None
+
+        return {
+            'id': video_id,
+            'title': title,
+            'formats': formats,
+            'webpage_url': webpage_url,
+            'subtitles': subs,
+            'thumbnails': thumbnails,
+            'duration': duration,
+            'timestamp': unified_timestamp(try_get(mc, [lambda x: x['_download']['date'],
+                                                        lambda x: x['_info']['clipDate']])),
+            'is_live': bool_or_none(mc.get('_isLive')),
+            'channel': try_get(mc, lambda x: x['_download']['channel'], compat_str),
+            'uploader': try_get(mc, lambda x: x['_info']['channelTitle'], compat_str),
+            'creator': try_get(mc, lambda x: x['_info']['clipContentSrc'], compat_str),
+            'description': try_get(mc, lambda x: x['_info']['clipDate'], compat_str),
+        }
+
+    def _real_extract(self, url):
+        mobj = re.match(self._VALID_URL, url)
+        video_id = mobj.group('id') or mobj.group('path')
+        display_id = video_id.lstrip('-') if video_id else 'tagesschau.de'
+
+        webpage = self._download_webpage(url, display_id)
+
+        title = (self._og_search_title(webpage, default=None)
+                 or self._html_search_regex(
+                     [r'<span[^>]*class="headline"[^>]*>(.+?)</span>',
+                      r'<title[^>]*>(.+?)</title>'],
+                     webpage, 'title'))
+
+        webpage_type = self._og_search_property('type', webpage, default=None)
+
+        player_pattern = r'<div[^>]+data-ts_component=(?P<quote>["\'])ts-mediaplayer(?P=quote)[^>]*>'
+        players = [m.group(0) for m in re.finditer(player_pattern, webpage)]
+        if not players:
+            # assume old website format
+            return self._legacy_extract(webpage, display_id, title, webpage_type)
+        elif (len(players) > 1
+              and not self._downloader.params.get('noplaylist')
+              and (webpage_type == 'website' or not mobj.group('id'))):
+            # article or playlist
+            entries = []
+            seen = set()
+            for player in players:
+                entry = self._extract_from_player(player, video_id, title)
+                if entry['id'] not in seen:
+                    entries.append(entry)
+                    seen.add(entry['id'])
+            return self.playlist_result(entries, display_id, title)
+        else:
+            # single video/audio
+            return self._extract_from_player(players[0], video_id, title)
+
+    def _legacy_extract_formats(self, download_text, media_kind):
         links = re.finditer(
             r'<div class="button" title="(?P<title>[^"]*)"><a href="(?P<url>[^"]+)">(?P<name>.+?)</a></div>',
             download_text)
@@ -219,9 +372,10 @@ class TagesschauIE(InfoExtractor):
             link_url = l.group('url')
             if not link_url:
                 continue
-            format_id = self._search_regex(
+            ext = determine_ext(link_url)
+            format_id = _normalize_format_id(self._search_regex(
                 r'.*/[^/.]+\.([^/]+)\.[^/.]+$', link_url, 'format ID',
-                default=determine_ext(link_url))
+                default=ext), ext)
             format = {
                 'format_id': format_id,
                 'url': l.group('url'),
@@ -262,39 +416,30 @@ class TagesschauIE(InfoExtractor):
         self._sort_formats(formats)
         return formats
 
-    def _real_extract(self, url):
-        mobj = re.match(self._VALID_URL, url)
-        video_id = mobj.group('id') or mobj.group('path')
-        display_id = video_id.lstrip('-')
+    # Some old pages still use the old format, so we keep the previous
+    # extractor for now.
+    def _legacy_extract(self, webpage, display_id, title, webpage_type):
+        DOWNLOAD_REGEX = r'<p>Wir bieten dieses (?P<kind>Video|Audio) in folgenden Formaten zum Download an:</p>\s*<div class="controls">(?P<links>.*?)</div>\s*<p>'
 
-        webpage = self._download_webpage(url, display_id)
-
-        title = self._html_search_regex(
-            r'<span[^>]*class="headline"[^>]*>(.+?)</span>',
-            webpage, 'title', default=None) or self._og_search_title(webpage)
-
-        DOWNLOAD_REGEX = r'(?s)<p>Wir bieten dieses (?P<kind>Video|Audio) in folgenden Formaten zum Download an:</p>\s*<div class="controls">(?P<links>.*?)</div>\s*<p>'
-
-        webpage_type = self._og_search_property('type', webpage, default=None)
         if webpage_type == 'website':  # Article
             entries = []
             for num, (entry_title, media_kind, download_text) in enumerate(re.findall(
-                    r'(?s)<p[^>]+class="infotext"[^>]*>\s*(?:<a[^>]+>)?\s*<strong>(.+?)</strong>.*?</p>.*?%s' % DOWNLOAD_REGEX,
-                    webpage), 1):
+                    r'<p[^>]+class="infotext"[^>]*>\s*(?:<a[^>]+>)?\s*<strong>(.+?)</strong>.*?</p>.*?%s' % DOWNLOAD_REGEX,
+                    webpage, flags=re.S), 1):
                 entries.append({
                     'id': '%s-%d' % (display_id, num),
                     'title': '%s' % entry_title,
-                    'formats': self._extract_formats(download_text, media_kind),
+                    'formats': self._legacy_extract_formats(download_text, media_kind),
                 })
-            if len(entries) > 1:
+            if len(entries) != 1:
                 return self.playlist_result(entries, display_id, title)
             formats = entries[0]['formats']
         else:  # Assume single video
             download_text = self._search_regex(
-                DOWNLOAD_REGEX, webpage, 'download links', group='links')
+                DOWNLOAD_REGEX, webpage, 'download links', flags=re.S, group='links')
             media_kind = self._search_regex(
-                DOWNLOAD_REGEX, webpage, 'media kind', default='Video', group='kind')
-            formats = self._extract_formats(download_text, media_kind)
+                DOWNLOAD_REGEX, webpage, 'media kind', default='Video', flags=re.S, group='kind')
+            formats = self._legacy_extract_formats(download_text, media_kind)
         thumbnail = self._og_search_thumbnail(webpage)
         description = self._html_search_regex(
             r'(?s)<p class="teasertext">(.*?)</p>',

youtube_dl/extractor/vidlii.py

@@ -4,6 +4,7 @@ from __future__ import unicode_literals
 import re
 
 from .common import InfoExtractor
+
 from ..utils import (
     float_or_none,
     get_element_by_id,
@@ -11,6 +12,7 @@ from ..utils import (
     strip_or_none,
     unified_strdate,
     urljoin,
+    str_to_int,
 )
 
 
@@ -35,6 +37,26 @@ class VidLiiIE(InfoExtractor):
             'categories': ['News & Politics'],
             'tags': ['Vidlii', 'Jan', 'Videogames'],
         }
+    }, {
+        # HD
+        'url': 'https://www.vidlii.com/watch?v=2Ng8Abj2Fkl',
+        'md5': '450e7da379c884788c3a4fa02a3ce1a4',
+        'info_dict': {
+            'id': '2Ng8Abj2Fkl',
+            'ext': 'mp4',
+            'title': 'test',
+            'description': 'md5:cc55a86032a7b6b3cbfd0f6b155b52e9',
+            'thumbnail': 'https://www.vidlii.com/usfi/thmp/2Ng8Abj2Fkl.jpg',
+            'uploader': 'VidLii',
+            'uploader_url': 'https://www.vidlii.com/user/VidLii',
+            'upload_date': '20200927',
+            'duration': 5,
+            'view_count': int,
+            'comment_count': int,
+            'average_rating': float,
+            'categories': ['Film & Animation'],
+            'tags': list,
+        },
     }, {
         'url': 'https://www.vidlii.com/embed?v=tJluaH4BJ3v&a=0',
         'only_matching': True,
@@ -46,11 +68,32 @@ class VidLiiIE(InfoExtractor):
         webpage = self._download_webpage(
             'https://www.vidlii.com/watch?v=%s' % video_id, video_id)
 
-        video_url = self._search_regex(
-            r'src\s*:\s*(["\'])(?P<url>(?:https?://)?(?:(?!\1).)+)\1', webpage,
-            'video url', group='url')
+        formats = []
 
-        title = self._search_regex(
+        def add_format(format_url, height=None):
+            height = int(self._search_regex(r'(\d+)\.mp4',
+                         format_url, 'height', default=360))
+
+            formats.append({
+                'url': format_url,
+                'format_id': '%dp' % height if height else None,
+                'height': height,
+            })
+
+        sources = re.findall(
+            r'src\s*:\s*(["\'])(?P<url>(?:https?://)?(?:(?!\1).)+)\1',
+            webpage)
+
+        formats = []
+        if len(sources) > 1:
+            add_format(sources[1][1])
+            self._check_formats(formats, video_id)
+        if len(sources) > 0:
+            add_format(sources[0][1])
+
+        self._sort_formats(formats)
+
+        title = self._html_search_regex(
             (r'<h1>([^<]+)</h1>', r'<title>([^<]+) - VidLii<'), webpage,
             'title')
 
@@ -82,9 +125,9 @@ class VidLiiIE(InfoExtractor):
             default=None) or self._search_regex(
             r'duration\s*:\s*(\d+)', webpage, 'duration', fatal=False))
 
-        view_count = int_or_none(self._search_regex(
-            (r'<strong>(\d+)</strong> views',
-             r'Views\s*:\s*<strong>(\d+)</strong>'),
+        view_count = str_to_int(self._html_search_regex(
+            (r'<strong>([\d,.]+)</strong> views',
+             r'Views\s*:\s*<strong>([\d,.]+)</strong>'),
             webpage, 'view count', fatal=False))
 
         comment_count = int_or_none(self._search_regex(
@@ -109,7 +152,7 @@ class VidLiiIE(InfoExtractor):
 
         return {
             'id': video_id,
-            'url': video_url,
+            'formats': formats,
             'title': title,
             'description': description,
             'thumbnail': thumbnail,