Merge 57c3cb420c 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

@@ -559,6 +559,10 @@ from .joj import JojIE
 from .jwplatform import JWPlatformIE
 from .kakao import KakaoIE
 from .kaltura import KalturaIE
+from .kan import (
+    KanEpisodeIE,
+    KanPlaylistIE
+)
 from .kankan import KankanIE
 from .karaoketv import KaraoketvIE
 from .karrierevideos import KarriereVideosIE

youtube_dl/extractor/kan.py

@@ -0,0 +1,116 @@
+# coding: utf-8
+from __future__ import unicode_literals
+from ..compat import compat_str
+import re
+
+from .common import InfoExtractor
+from ..utils import (
+    ExtractorError,
+    parse_duration,
+    try_get,
+    unified_strdate)
+
+
+def get_thumbnail(data):
+    for media in data.get('media_group', []):
+        if media.get('type') == 'image':
+            for item in media.get('media_item', []):
+                thumbnail = item.get('src')
+                if thumbnail:
+                    return thumbnail
+
+
+class KanBaseIE(InfoExtractor):
+    def download_webpage(self, url, video_id):
+        return self._download_webpage(
+            url,
+            video_id)
+
+    def extract_item(self, video_id, webpage):
+        data = self._parse_json(
+            self._search_regex(
+                r'<script[^>]+id="kan_app_search_data"[^>]*>([^<]+)</script>',
+                webpage,
+                'data'),
+            video_id)
+        title = data.get('title') or self._og_search_title(webpage)
+        m3u8_url = try_get(data, lambda x: x['content']['src'], compat_str)
+        formats = self._extract_m3u8_formats(m3u8_url, video_id, ext='mp4')
+        if not formats:
+            raise ExtractorError('Unable to extract video formats')
+        description = data.get('summary') or \
+            self._og_search_description(webpage, fatal=False)
+        creator = try_get(data, lambda x: x['author']['name'], compat_str) or \
+            self._og_search_property('site_name', webpage, fatal=False)
+        thumbnail = get_thumbnail(data)
+
+        return {
+            'id': video_id,
+            'title': title,
+            'thumbnail': thumbnail,
+            'formats': formats,
+            'description': description,
+            'creator': creator,
+            'release_date': unified_strdate(data.get('published')),
+            'duration': parse_duration(
+                try_get(data, lambda x: x['extensions']['duration']))
+        }
+
+
+class KanEpisodeIE(KanBaseIE):
+    _VALID_URL = r'https?://(?:www\.)?kan\.org\.il/[iI]tem/\?item[iI]d=(?P<id>[0-9]+)'
+    _TEST = {
+        'url': 'https://www.kan.org.il/Item/?itemId=74658',
+        'md5': 'c28763bdb61c1bb7823528dd024e6129',
+        'info_dict': {
+            'id': '74658',
+            'ext': 'mp4',
+            'title': 'העד - פרק 2',
+            'thumbnail': r're:^https://.*36805_A\.jpeg$',
+            'description': 'הגופות ממשיכות להיערם, אך איזי עדיין מפקפק בחשדות נגד ברק',
+            'creator': 'מערכת כאן',
+            'release_date': '20200803',
+            'duration': 2393,
+        },
+        'params': {
+            'skip_download': True
+        },
+    }
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
+        return self.extract_item(video_id, self.download_webpage(url, video_id))
+
+
+class KanPlaylistIE(KanBaseIE):
+    _VALID_URL = r'https?://(?:www\.)?kan\.org\.il/program/\?cat[iI]d=(?P<id>[0-9]+)'
+    _TEST = {
+        'url': 'https://www.kan.org.il/program/?catId=1636',
+        'playlist_mincount': 9,
+        'info_dict': {
+            'id': '1636',
+            'title': 'מנאייכ - פרקים מלאים לצפייה ישירה | כאן',
+            'description': 'md5:9dfbd501189d08674d20762464c5301b',
+        },
+    }
+
+    def _real_extract(self, url):
+        list_id = self._match_id(url)
+        webpage = self.download_webpage(url, list_id)
+        video_ids = re.findall(r'onclick="playVideo\(.*,\'([0-9]+)\'\)', webpage)
+        entries = []
+        for video_id in video_ids:
+            video_url = 'https://www.kan.org.il/Item/?itemId=%s' % video_id
+            entries.append(self.extract_item(
+                video_id,
+                self.download_webpage(video_url, video_id)))
+        if not entries:
+            raise ExtractorError('Unable to extract playlist entries')
+
+        return {
+            '_type': 'playlist',
+            'id': list_id,
+            'entries': entries,
+            'title': self._og_search_title(webpage, fatal=False),
+            'description': self._og_search_description(webpage),
+        }

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')