Merge 260e0c132a into c5098961b0

Now also succeeds with player b12cc44b

test/test_jsinterp.py

@@ -425,6 +425,34 @@ class TestJSInterpreter(unittest.TestCase):
             self._test(jsi, [''], args=['', '-'])
             self._test(jsi, [], args=['', ''])
 
+    def test_slice(self):
+        self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice()}', [0, 1, 2, 3, 4, 5, 6, 7, 8])
+        self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(0)}', [0, 1, 2, 3, 4, 5, 6, 7, 8])
+        self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(5)}', [5, 6, 7, 8])
+        self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(99)}', [])
+        self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(-2)}', [7, 8])
+        self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(-99)}', [0, 1, 2, 3, 4, 5, 6, 7, 8])
+        self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(0, 0)}', [])
+        self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(1, 0)}', [])
+        self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(0, 1)}', [0])
+        self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(3, 6)}', [3, 4, 5])
+        self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(1, -1)}', [1, 2, 3, 4, 5, 6, 7])
+        self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(-1, 1)}', [])
+        self._test('function f(){return [0, 1, 2, 3, 4, 5, 6, 7, 8].slice(-3, -1)}', [6, 7])
+        self._test('function f(){return "012345678".slice()}', '012345678')
+        self._test('function f(){return "012345678".slice(0)}', '012345678')
+        self._test('function f(){return "012345678".slice(5)}', '5678')
+        self._test('function f(){return "012345678".slice(99)}', '')
+        self._test('function f(){return "012345678".slice(-2)}', '78')
+        self._test('function f(){return "012345678".slice(-99)}', '012345678')
+        self._test('function f(){return "012345678".slice(0, 0)}', '')
+        self._test('function f(){return "012345678".slice(1, 0)}', '')
+        self._test('function f(){return "012345678".slice(0, 1)}', '0')
+        self._test('function f(){return "012345678".slice(3, 6)}', '345')
+        self._test('function f(){return "012345678".slice(1, -1)}', '1234567')
+        self._test('function f(){return "012345678".slice(-1, 1)}', '')
+        self._test('function f(){return "012345678".slice(-3, -1)}', '67')
+
 
 if __name__ == '__main__':
     unittest.main()

test/test_youtube_signature.py

@@ -174,6 +174,14 @@ _NSIG_TESTS = [
         'https://www.youtube.com/s/player/5604538d/player_ias.vflset/en_US/base.js',
         '7X-he4jjvMx7BCX', 'sViSydX8IHtdWA',
     ),
+    (
+        'https://www.youtube.com/s/player/20dfca59/player_ias.vflset/en_US/base.js',
+        '-fLCxedkAk4LUTK2', 'O8kfRq1y1eyHGw',
+    ),
+    (
+        'https://www.youtube.com/s/player/b12cc44b/player_ias.vflset/en_US/base.js',
+        'keLa5R2U00sR9SQK', 'N1OGyujjEwMnLw',
+    ),
 ]
 
 

youtube_dl/extractor/extractors.py

@@ -989,6 +989,7 @@ from .puhutv import (
     PuhuTVIE,
     PuhuTVSerieIE,
 )
+from .premierone import PremierOneIE
 from .presstv import PressTVIE
 from .prosiebensat1 import ProSiebenSat1IE
 from .puls4 import Puls4IE

youtube_dl/extractor/premierone.py

@@ -0,0 +1,127 @@
+# coding: utf-8
+from __future__ import unicode_literals
+import re
+
+from .common import InfoExtractor
+from ..utils import (
+    parse_duration,
+    parse_iso8601,
+)
+
+
+class PremierOneIE(InfoExtractor):
+    _VALID_URL = r'https?://premier\.one/show/(?P<show_id>[0-9]+)(/season/(?P<season_number>\d+))?(/video/(?P<id>[0-9a-f]+))?'
+    _TEST = {
+        'url': 'https://premier.one/show/3421/season/6/video/b4766f2eeb90cbb5538729061ac3949b',
+        'info_dict': {
+            'id': 'b4766f2eeb90cbb5538729061ac3949b',
+            'ext': 'mp4',
+            'title': '32 выпуск',
+            'thumbnail': 'https://pic.uma.media/pic/video/06/c2/06c2cb95bb06c03c379d9d20930c5718.jpg',
+            'description': 'В этом выпуске шоу «Где логика?» в интеллектуальном поединке сразятся Екатерина Моргунова и Манижа против Ильи Соболева и Зураба Матуа.',
+            'timestamp': 1611615378,
+            'upload_date': '20210125',
+        },
+        'params': {
+            # m3u8 download
+            'skip_download': True,
+        },
+    }
+
+    def _extract_video(self, video_id):
+        video = self._download_json(
+            'https://premier.one/uma-api/video/' + video_id,
+            video_id, 'Downloading video information ' + video_id,
+            headers={
+                'Accept': 'application/json',
+                'Referer': 'https://premier.one/play/embed/' + video_id + '?controlledFullscreen=true'
+            })
+
+        return self._extract_video_info(video)
+
+    def _extract_video_info(self, item):
+        video_id = item.get('id')
+        info = {
+            'id': video_id,
+            'title': item.get('title'),
+            'description': item.get('description'),
+            'thumbnail': item.get('thumbnail_url'),
+            'duration': parse_duration(item.get('duration')),
+            'timestamp': parse_iso8601(item.get('publication_ts')),
+            'comment_count': item.get('comments_count'),
+            'season_number': item.get('season'),
+            'episode_number': item.get('episode'),
+            'url': item.get('video_url'),
+            'ext': 'mp4',
+        }
+
+        video = self._download_json(
+            'https://premier.one/api/play/options/' + video_id + '/?format=json&no_404=true',
+            video_id, 'Downloading video ' + info.get('title'),
+            headers={
+                'Accept': 'application/json',
+                'Referer': 'https://premier.one/play/embed/' + video_id + '?controlledFullscreen=true'
+            })
+        if video.get('video_balancer') and video.get('video_balancer').get('default'):
+            formats = []
+            m3u8_url = video.get('video_balancer').get('default')
+            m3u8_formats = self._extract_m3u8_formats(
+                m3u8_url, video_id, 'mp4', 'm3u8_native',
+                fatal=False)
+            formats.extend(m3u8_formats)
+
+            info['formats'] = formats
+
+        return info
+
+    def _extract_show_playlist(self, show_id, season_number):
+        seasons = self._download_json(
+            'https://premier.one/uma-api/metainfo/tv/' + show_id + '/season/',
+            show_id, 'Downloading videos JSON',
+            headers={
+                'Accept': 'application/json',
+                'Referer': 'https://premier.one/show/' + show_id
+            })
+
+        entries = []
+        for season in seasons:
+            if season_number is not None and str(season.get('number')) != season_number:
+                continue
+
+            metainfo = self._download_json(
+                'https://premier.one/uma-api/metainfo/tv/' + show_id + '/video/?season=' + str(season.get('number')),
+                str(season.get('number')), 'Downloading season ' + str(season.get('number')),
+                headers={
+                    'Accept': 'application/json',
+                    'Referer': 'https://premier.one/show/' + show_id + '/season/' + str(season.get('number'))
+                })
+            while metainfo and len(metainfo.get('results')) > 0:
+                for item in metainfo.get('results'):
+                    video_id = item.get('id')
+                    if not video_id:
+                        continue
+
+                    info = self._extract_video_info(item)
+                    if info:
+                        entries.append(info)
+
+                if metainfo.get('has_next'):
+                    metainfo = self._download_json(
+                        metainfo.get('next'),
+                        str(season.get('number')), 'Downloading metainfo JSON for page ' + str(metainfo.get('page') + 1),
+                        headers={
+                            'Accept': 'application/json',
+                            'Referer': 'https://premier.one/show/' + str(show_id) + '/season/' + str(season.get('number')) + '/video/' + video_id
+                        })
+                else:
+                    metainfo = None
+
+        return self.playlist_result(entries, show_id)
+
+    def _real_extract(self, url):
+        mobj = re.match(self._VALID_URL, url)
+
+        if mobj.group('id'):
+            return self._extract_video(mobj.group('id'))
+        else:
+            return self._extract_show_playlist(mobj.group('show_id'), mobj.group('season_number'))

youtube_dl/extractor/youtube.py

@@ -1659,17 +1659,46 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
     def _extract_n_function_name(self, jscode):
         func_name, idx = self._search_regex(
             # new: (b=String.fromCharCode(110),c=a.get(b))&&c=nfunc[idx](c)
-            # or:  (b="nn"[+a.D],c=a.get(b))&&(c=nfunc[idx](c)s
-            # old: .get("n"))&&(b=nfunc[idx](b)
-            # older: .get("n"))&&(b=nfunc(b)
+            # or:  (b="nn"[+a.D],c=a.get(b))&&(c=nfunc[idx](c)
+            # or:  (PL(a),b=a.j.n||null)&&(b=nfunc[idx](b)
+            # or:  (b="nn"[+a.D],vL(a),c=a.j[b]||null)&&(c=narray[idx](c),a.set(b,c),narray.length||nfunc("")
+            # old: (b=a.get("n"))&&(b=nfunc[idx](b)(?P<c>[a-z])\s*=\s*[a-z]\s*
+            # older: (b=a.get("n"))&&(b=nfunc(b)
             r'''(?x)
-                (?:\(\s*(?P<b>[a-z])\s*=\s*(?:
-                    String\s*\.\s*fromCharCode\s*\(\s*110\s*\)|
-                    "n+"\[\s*\+?s*[\w$.]+\s*]
-                )\s*,(?P<c>[a-z])\s*=\s*[a-z]\s*)?
-                \.\s*get\s*\(\s*(?(b)(?P=b)|"n{1,2}")(?:\s*\)){2}\s*&&\s*\(\s*(?(c)(?P=c)|b)\s*=\s*
-                (?P<nfunc>[a-zA-Z_$][\w$]*)(?:\s*\[(?P<idx>\d+)\])?\s*\(\s*[\w$]+\s*\)
-            ''', jscode, 'Initial JS player n function name', group=('nfunc', 'idx'))
+                \((?:[\w$()\s]+,)*?\s*      # (
+                (?P<b>[a-z])\s*=\s*         # b=
+                (?:
+                    (?:                     # expect ,c=a.get(b) (etc)
+                        String\s*\.\s*fromCharCode\s*\(\s*110\s*\)|
+                        "n+"\[\s*\+?s*[\w$.]+\s*]
+                    )\s*(?:,[\w$()\s]+(?=,))*|
+                       (?P<old>[\w$]+)      # a (old[er])
+                   )\s*
+                   (?(old)
+                                            # b.get("n")
+                       (?:\.\s*[\w$]+\s*|\[\s*[\w$]+\s*]\s*)*?
+                       (?:\.\s*n|\[\s*"n"\s*]|\.\s*get\s*\(\s*"n"\s*\))
+                       |                    # ,c=a.get(b)
+                       ,\s*(?P<c>[a-z])\s*=\s*[a-z]\s*
+                       (?:\.\s*[\w$]+\s*|\[\s*[\w$]+\s*]\s*)*?
+                       (?:\[\s*(?P=b)\s*]|\.\s*get\s*\(\s*(?P=b)\s*\))
+                   )
+                                            # interstitial junk
+                   \s*(?:\|\|\s*null\s*)?(?:\)\s*)?&&\s*(?:\(\s*)?
+               (?(c)(?P=c)|(?P=b))\s*=\s*   # [c|b]=
+                                            # nfunc|nfunc[idx]
+                   (?P<nfunc>[a-zA-Z_$][\w$]*)(?:\s*\[(?P<idx>\d+)\])?\s*\(\s*[\w$]+\s*\)
+            ''', jscode, 'Initial JS player n function name', group=('nfunc', 'idx'),
+            default=(None, None))
+        # thx bashonly: yt-dlp/yt-dlp/pull/10611
+        if not func_name:
+            self.report_warning('Falling back to generic n function search')
+            return self._search_regex(
+                r'''(?xs)
+                    (?:(?<=[^\w$])|^)       # instead of \b, which ignores $
+                    (?P<name>(?!\d)[a-zA-Z\d_$]+)\s*=\s*function\((?!\d)[a-zA-Z\d_$]+\)
+                    \s*\{(?:(?!};).)+?["']enhanced_except_
+                ''', jscode, 'Initial JS player n function name', group='name')
         if not idx:
             return func_name
 

youtube_dl/jsinterp.py

@@ -925,9 +925,16 @@ class JSInterpreter(object):
                     obj.reverse()
                     return obj
                 elif member == 'slice':
-                    assertion(isinstance(obj, list), 'must be applied on a list')
-                    assertion(len(argvals) == 1, 'takes exactly one argument')
-                    return obj[argvals[0]:]
+                    assertion(isinstance(obj, (list, compat_str)), 'must be applied on a list or string')
+                    # From [1]:
+                    # .slice() - like [:]
+                    # .slice(n) - like [n:] (not [slice(n)]
+                    # .slice(m, n) - like [m:n] or [slice(m, n)]
+                    # [1] https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Array/slice
+                    assertion(len(argvals) <= 2, 'takes between 0 and 2 arguments')
+                    if len(argvals) < 2:
+                        argvals += (None,)
+                    return obj[slice(*argvals)]
                 elif member == 'splice':
                     assertion(isinstance(obj, list), 'must be applied on a list')
                     assertion(argvals, 'takes one or more arguments')