Merge 2730e55b36 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

@@ -810,6 +810,10 @@ from .nhk import (
     NhkVodIE,
     NhkVodProgramIE,
 )
+from .nhkRadio import (
+    NhkRadioIE,
+    NhkRadioProgramIE
+)
 from .nhl import NHLIE
 from .nick import (
     NickIE,

youtube_dl/extractor/nhkRadio.py

@@ -0,0 +1,106 @@
+# coding: utf-8
+from __future__ import unicode_literals
+from .common import InfoExtractor
+from ..utils import ExtractorError, parse_iso8601
+from ..compat import compat_HTTPError
+import re
+
+
+class NhkRadioBase(InfoExtractor):
+    def _get_json_meta(self, program_id, corner_id):
+        program_corner_id = program_id + "_" + corner_id
+        try:
+            data = self._download_json(
+                "https://www.nhk.or.jp/radioondemand/json/"
+                + program_id
+                + "/bangumi_"
+                + program_corner_id
+                + ".json",
+                program_corner_id,
+            )
+        except ExtractorError as e:
+            if isinstance(e.cause, compat_HTTPError) and e.cause.code == 404:
+                raise ExtractorError("The invalid url", expected=True)
+        return data
+
+    def _extract_program(self, info, program_corner_id):
+        id = program_corner_id + "_" + info.get("headline_id")
+        file = info.get("file_list")[0]
+        formats = self._extract_m3u8_formats(file.get("file_name", None), id, "m4a", entry_protocol="m3u8_native", m3u8_id="hls")
+        self._sort_formats(formats)
+        return {
+            "id": id,
+            "title": file.get("file_title"),
+            "formats": formats,
+            "timestamp": parse_iso8601(file.get("close_time")),
+        }
+
+
+class NhkRadioIE(NhkRadioBase):
+    _VALID_URL = r"https?://www\.nhk\.or\.jp/radio/player/ondemand\.html\?p=(?P<program_id>\d+)_(?P<corner_id>\d+)_(?P<headline_id>\d+)"
+
+    _TESTS = [
+        {
+            "url": "https://www.nhk.or.jp/radio/player/ondemand.html?p=4812_01_2898188",
+            "info_dict": {
+                "id": "4812_01_2898188",
+                "ext": "m4a",
+                "title": "世界へ発信！ニュースで英語術　＃２０９▽“首相長男から接待”　総務省１１人を処分",
+                "upload_date": str,
+                "timestamp": int,
+            },
+        },
+        {
+            "url": "https://www.nhk.or.jp/radio/player/ondemand.html?p=0444_01_2890944",
+            "info_dict": {
+                "ext": "m4a",
+                "id": "0444_01_2890944",
+                "title": "歌謡スクランブル　春色コレクション（３）　▽尾崎亜美",
+                "upload_date": str,
+                "timestamp": int,
+            },
+        },
+    ]
+
+    def _real_extract(self, url):
+        program_id, corner_id, headline_id = re.match(self._VALID_URL, url).groups()
+        program_corner_id = program_id + "_" + corner_id
+        data = self._download_json(
+            "https://www.nhk.or.jp/radioondemand/json/"
+            + program_id
+            + "/bangumi_"
+            + program_corner_id
+            + ".json",
+            program_corner_id,
+        )
+        for detail in data["main"]["detail_list"]:
+            if headline_id == detail.get("headline_id"):
+                return self._extract_program(detail, program_corner_id)
+        raise ExtractorError("The program not found", expected=True)
+
+
+class NhkRadioProgramIE(NhkRadioBase):
+    _VALID_URL = r"https?://www\.nhk\.or\.jp/radio/ondemand/detail\.html\?p=(?P<program_id>\d+)_(?P<corner_id>\d+)"
+
+    _TESTS = [
+        {
+            "url": "https://www.nhk.or.jp/radio/ondemand/detail.html?p=0164_01",
+            "info_dict": {"title": "青春アドベンチャー", "id": "0164_01"},
+            "playlist_mincount": 5,
+        },
+        {
+            "url": "https://www.nhk.or.jp/radio/ondemand/detail.html?p=0455_01",
+            "info_dict": {"id": "0455_01", "title": "弾き語りフォーユー"},
+            "playlist_mincount": 5,
+        },
+    ]
+
+    def _real_extract(self, url):
+        program_id, corner_id = re.match(self._VALID_URL, url).groups()
+        data = self._get_json_meta(program_id, corner_id)
+        entries = []
+        for detail in data["main"]["detail_list"]:
+            entries.append(self._extract_program(detail, program_id + "_" + corner_id))
+        return self.playlist_result(
+            entries, program_id + "_" + corner_id, data["main"]["program_name"]
+        )

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*(?:
+                \((?:[\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*,(?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*
+                    )\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'))
+            ''', 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')