Merge 481ad85994 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

@@ -917,6 +917,7 @@ from .palcomp3 import (
     PalcoMP3VideoIE,
 )
 from .pandoratv import PandoraTVIE
+from .parler import ParlerIE
 from .parliamentliveuk import ParliamentLiveUKIE
 from .patreon import PatreonIE
 from .pbs import PBSIE

youtube_dl/extractor/parler.py

@@ -0,0 +1,114 @@
+# coding: utf-8
+from __future__ import unicode_literals
+
+from .common import InfoExtractor
+from .youtube import YoutubeIE
+from ..utils import (
+    clean_html,
+    int_or_none,
+    strip_or_none,
+    try_get,
+    unified_timestamp,
+    urlencode_postdata,
+)
+
+
+class ParlerIE(InfoExtractor):
+    IE_DESC = 'Posts on parler.com'
+    _VALID_URL = r'https://parler\.com/feed/(?P<id>[0-9a-f]{8}-(?:[0-9a-f]{4}-){3}[0-9a-f]{12})'
+    _TESTS = [
+        {
+            'url': 'https://parler.com/feed/df79fdba-07cc-48fe-b085-3293897520d7',
+            'md5': '16e0f447bf186bb3cf64de5bbbf4d22d',
+            'info_dict': {
+                'id': 'df79fdba-07cc-48fe-b085-3293897520d7',
+                'ext': 'mp4',
+                'thumbnail': 'https://bl-images.parler.com/videos/6ce7cdf3-a27a-4d72-bf9c-d3e17ce39a66/thumbnail.jpeg',
+                'title': 'Parler video #df79fdba-07cc-48fe-b085-3293897520d7',
+                'description': 'md5:6f220bde2df4a97cbb89ac11f1fd8197',
+                'timestamp': 1659744000,
+                'upload_date': '20220806',
+                'uploader': 'Tulsi Gabbard',
+                'uploader_id': 'TulsiGabbard',
+                'uploader_url': 'https://parler.com/TulsiGabbard',
+                'view_count': int,
+                'comment_count': int,
+                'repost_count': int,
+            },
+        },
+        {
+            'url': 'https://parler.com/feed/a7406eb4-91e5-4793-b5e3-ade57a24e287',
+            'md5': '11687e2f5bb353682cee338d181422ed',
+            'info_dict': {
+                'id': 'a7406eb4-91e5-4793-b5e3-ade57a24e287',
+                'ext': 'mp4',
+                'thumbnail': 'https://bl-images.parler.com/videos/317827a8-1e48-4cbc-981f-7dd17d4c1183/thumbnail.jpeg',
+                'title': 'Parler video #a7406eb4-91e5-4793-b5e3-ade57a24e287',
+                'description': 'This man should run for office',
+                'timestamp': 1659657600,
+                'upload_date': '20220805',
+                'uploader': 'Benny Johnson',
+                'uploader_id': 'BennyJohnson',
+                'uploader_url': 'https://parler.com/BennyJohnson',
+                'view_count': int,
+                'comment_count': int,
+                'repost_count': int,
+            },
+        },
+        {
+            'url': 'https://parler.com/feed/f23b85c1-6558-470f-b9ff-02c145f28da5',
+            'md5': 'eaba1ff4a10fe281f5ce74e930ab2cb4',
+            'info_dict': {
+                'id': 'r5vkSaz8PxQ',
+                'ext': 'mp4',
+                'thumbnail': 'https://i.ytimg.com/vi_webp/r5vkSaz8PxQ/maxresdefault.webp',
+                'title': 'Tom MacDonald Names Reaction',
+                'description': 'md5:33c21f0d35ae6dc2edf3007d6696baea',
+                'upload_date': '20220716',
+                'duration': 1267,
+                'uploader': 'Mahesh Chookolingo',
+                'uploader_id': 'maheshchookolingo',
+                'uploader_url': 'http://www.youtube.com/user/maheshchookolingo',
+                'channel': 'Mahesh Chookolingo',
+                'channel_id': 'UCox6YeMSY1PQInbCtTaZj_w',
+                'channel_url': 'https://www.youtube.com/channel/UCox6YeMSY1PQInbCtTaZj_w',
+                'categories': ['Entertainment'],
+                'tags': list,
+                'availability': 'public',
+                'live_status': 'not_live',
+                'view_count': int,
+                'comment_count': int,
+                'like_count': int,
+                'channel_follower_count': int,
+                'age_limit': 0,
+                'playable_in_embed': True,
+            },
+            'add_ie': ['Youtube'],
+        },
+    ]
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
+        data = self._download_json(
+            'https://parler.com/open-api/ParleyDetailEndpoint.php', video_id,
+            data=urlencode_postdata({'uuid': video_id}))['data'][0]
+        primary = try_get(data, lambda x: x['primary'], dict) or {}
+
+        embed = self._parse_json(primary.get('V2LINKLONG') or '', video_id, fatal=False)
+        if embed:
+            return self.url_result(embed[0], YoutubeIE.ie_key())
+
+        return {
+            'id': video_id,
+            'url': primary['video_data']['videoSrc'],
+            'thumbnail': primary['video_data']['thumbnailUrl'],
+            'title': "Parler video #%s" % video_id,
+            'description': strip_or_none(clean_html(primary.get('full_body'))) or None,
+            'timestamp': unified_timestamp(primary.get('date_created')),
+            'uploader': strip_or_none(primary.get('name')),
+            'uploader_id': strip_or_none(primary.get('username')),
+            'uploader_url': 'https://parler.com/%s' % strip_or_none(primary.get('username')),
+            'view_count': int_or_none(primary.get('view_count')),
+            'comment_count': int_or_none(data['engagement']['commentCount']),
+            'repost_count': int_or_none(data['engagement']['echoCount']),
+        }

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')