Merge 1683735ea5 into c5098961b0

[Youtube] Rework n function extraction pattern
Now also succeeds with player b12cc44b
2025-01-25 12:50:10 +09:00 · 2024-08-21 22:26:23 -04:00 · 2024-08-06 20:59:09 +01:00 · 2024-08-06 20:51:38 +01:00 · 2024-08-01 19:18:34 +01:00 · 2020-09-08 09:15:03 +03:00
6 changed files with 223 additions and 13 deletions
--- a/test/test_jsinterp.py
+++ b/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()
--- a/test/test_youtube_signature.py
+++ b/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',
+    ),
 ]


--- a/youtube_dl/extractor/elonet.py
+++ b/youtube_dl/extractor/elonet.py
@ -0,0 +1,137 @@
+# coding: utf-8
+from __future__ import unicode_literals
+
+import os
+import re
+import tempfile
+
+from .common import InfoExtractor
+from ..utils import (
+    base_url,
+    ExtractorError,
+    try_get,
+)
+from ..compat import compat_str
+from ..downloader.hls import HlsFD
+
+
+class ElonetIE(InfoExtractor):
+    _VALID_URL = r'https?://elonet\.finna\.fi/Record/kavi\.elonet_elokuva_(?P<id>[0-9]+)'
+    _TEST = {
+        'url': 'https://elonet.finna.fi/Record/kavi.elonet_elokuva_107867',
+        'md5': '8efc954b96c543711707f87de757caea',
+        'info_dict': {
+            'id': '107867',
+            'ext': 'mp4',
+            'title': 'Valkoinen peura',
+            'description': 'Valkoinen peura (1952) on Erik Blombergin ohjaama ja yhdessä Mirjami Kuosmasen kanssa käsikirjoittama tarunomainen kertomus valkoisen peuran hahmossa lii...',
+            'thumbnail': 'https://elonet.finna.fi/Cover/Show?id=kavi.elonet_elokuva_107867&index=0&size=large',
+        },
+    }
+
+    def _download_m3u8_chunked_subtitle(self, chunklist_url):
+        """
+        Download VTT subtitles from pieces in manifest URL.
+        Return a string containing joined chunks with extra headers removed.
+        """
+        with tempfile.NamedTemporaryFile(delete=True) as outfile:
+            fname = outfile.name
+        hlsdl = HlsFD(self._downloader, {})
+        hlsdl.download(compat_str(fname), {"url": chunklist_url})
+        with open(fname, 'r') as fin:
+            # Remove (some) headers
+            fdata = re.sub(r'X-TIMESTAMP-MAP.*\n+|WEBVTT\n+', '', fin.read())
+        os.remove(fname)
+        return "WEBVTT\n\n" + fdata
+
+    def _parse_m3u8_subtitles(self, m3u8_doc, m3u8_url):
+        """
+        Parse subtitles from HLS / m3u8 manifest.
+        """
+        subtitles = {}
+        baseurl = m3u8_url[:m3u8_url.rindex('/') + 1]
+        for line in m3u8_doc.split('\n'):
+            if 'EXT-X-MEDIA:TYPE=SUBTITLES' in line:
+                lang = self._search_regex(
+                    r'LANGUAGE="(.+?)"', line, 'lang', default=False)
+                uri = self._search_regex(
+                    r'URI="(.+?)"', line, 'uri', default=False)
+                if lang and uri:
+                    data = self._download_m3u8_chunked_subtitle(baseurl + uri)
+                    subtitles[lang] = [{'ext': 'vtt', 'data': data}]
+        return subtitles
+
+    def _parse_mpd_subtitles(self, mpd_doc):
+        """
+        Parse subtitles from MPD manifest.
+        """
+        ns = '{urn:mpeg:dash:schema:mpd:2011}'
+        subtitles = {}
+        for aset in mpd_doc.findall(".//%sAdaptationSet[@mimeType='text/vtt']" % (ns)):
+            lang = aset.attrib.get('lang', 'unk')
+            url = aset.find("./%sRepresentation/%sBaseURL" % (ns, ns)).text
+            subtitles[lang] = [{'ext': 'vtt', 'url': url}]
+        return subtitles
+
+    def _get_subtitles(self, fmt, doc, url):
+        if fmt == 'm3u8':
+            subs = self._parse_m3u8_subtitles(doc, url)
+        elif fmt == 'mpd':
+            subs = self._parse_mpd_subtitles(doc)
+        else:
+            self._downloader.report_warning(
+                "Cannot download subtitles from '%s' streams." % (fmt))
+            subs = {}
+        return subs
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
+        webpage = self._download_webpage(url, video_id)
+
+        title = self._html_search_regex(
+            r'<meta .*property="og&#x3A;title" .*content="(.+?)"', webpage, 'title')
+        description = self._html_search_regex(
+            r'<meta .*property="og&#x3A;description" .*content="(.+?)"', webpage, 'description')
+        thumbnail = self._html_search_regex(
+            r'<meta .*property="og&#x3A;image" .*content="(.+?)"', webpage, 'thumbnail')
+
+        json_s = self._html_search_regex(
+            r'data-video-sources="(.+?)"', webpage, 'json')
+        src = try_get(
+            self._parse_json(json_s, video_id),
+            lambda x: x[0]["src"], compat_str)
+        formats = []
+        if re.search(r'\.m3u8\??', src):
+            fmt = 'm3u8'
+            res = self._download_webpage_handle(
+                # elonet servers have certificate problems
+                src.replace('https:', 'http:'), video_id,
+                note='Downloading m3u8 information',
+                errnote='Failed to download m3u8 information')
+            if res:
+                doc, urlh = res
+                url = urlh.geturl()
+                formats = self._parse_m3u8_formats(doc, url)
+                for f in formats:
+                    f['ext'] = 'mp4'
+        elif re.search(r'\.mpd\??', src):
+            fmt = 'mpd'
+            res = self._download_xml_handle(
+                src, video_id,
+                note='Downloading MPD manifest',
+                errnote='Failed to download MPD manifest')
+            if res:
+                doc, urlh = res
+                url = base_url(urlh.geturl())
+                formats = self._parse_mpd_formats(doc, mpd_base_url=url)
+        else:
+            raise ExtractorError("Unknown streaming format")
+
+        return {
+            'id': video_id,
+            'title': title,
+            'description': description,
+            'thumbnail': thumbnail,
+            'formats': formats,
+            'subtitles': self.extract_subtitles(fmt, doc, url),
+        }
--- a/youtube_dl/extractor/extractors.py
+++ b/youtube_dl/extractor/extractors.py
@ -355,6 +355,7 @@ from .ellentube import (
    EllenTubeVideoIE,
    EllenTubePlaylistIE,
 )
+from .elonet import ElonetIE
 from .elpais import ElPaisIE
 from .embedly import EmbedlyIE
 from .engadget import EngadgetIE
--- a/youtube_dl/extractor/youtube.py
+++ b/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

--- a/youtube_dl/jsinterp.py
+++ b/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')
Author	SHA1	Message	Date
Teemu Ikonen	caf2e20d22	Merge `1683735ea5` into `c5098961b0`	2024-08-21 22:26:23 -04:00
dirkf	c5098961b0	[Youtube] Rework n function extraction pattern Now also succeeds with player b12cc44b	2024-08-06 20:59:09 +01:00
dirkf	dbc08fba83	[jsinterp] Improve slice implementation for player b12cc44b Partly taken from yt-dlp/yt-dlp#10664, thx seproDev Fixes #32896	2024-08-06 20:51:38 +01:00
Aiur Adept	71223bff39	[Youtube] Fix nsig extraction for player 20dfca59 (#32891 ) * dirkf's patch for nsig extraction * add generic search per yt-dlp/yt-dlp/pull/10611 - thx bashonly --------- Co-authored-by: dirkf <fieldhouse@gmx.net>	2024-08-01 19:18:34 +01:00
Teemu Ikonen	1683735ea5	[elonet] Add extractor (closes #24774 )	2020-09-08 09:15:03 +03:00