Merge b6bfdb25f2 into 4d05f84325

* no space after @ in decorator

docs/supportedsites.md

@@ -910,6 +910,9 @@
  - **SVTPage**
  - **SVTPlay**: SVT Play and Öppet arkiv
  - **SVTSeries**
+ - **switchtube**
+ - **switchtube:channel**
+ - **switchtube:profile**
  - **SWRMediathek**
  - **Syfy**
  - **SztvHu**

test/test_jsinterp.py

@@ -577,9 +577,11 @@ class TestJSInterpreter(unittest.TestCase):
     def test_unary_operators(self):
         jsi = JSInterpreter('function f(){return 2  -  - - 2;}')
         self.assertEqual(jsi.call_function('f'), 0)
-        # fails
-        # jsi = JSInterpreter('function f(){return 2 + - + - - 2;}')
-        # self.assertEqual(jsi.call_function('f'), 0)
+        jsi = JSInterpreter('function f(){return 2 + - + - - 2;}')
+        self.assertEqual(jsi.call_function('f'), 0)
+        # https://github.com/ytdl-org/youtube-dl/issues/32815
+        jsi = JSInterpreter('function f(){return 0  - 7 * - 6;}')
+        self.assertEqual(jsi.call_function('f'), 42)
 
     """ # fails so far
     def test_packed(self):

test/test_youtube_signature.py

@@ -158,6 +158,10 @@ _NSIG_TESTS = [
         'https://www.youtube.com/s/player/b7910ca8/player_ias.vflset/en_US/base.js',
         '_hXMCwMt9qE310D', 'LoZMgkkofRMCZQ',
     ),
+    (
+        'https://www.youtube.com/s/player/590f65a6/player_ias.vflset/en_US/base.js',
+        '1tm7-g_A9zsI8_Lay_', 'xI4Vem4Put_rOg',
+    ),
 ]
 
 

youtube_dl/extractor/common.py

@@ -3033,7 +3033,6 @@ class InfoExtractor(object):
             transform_source=transform_source, default=None)
 
     def _extract_jwplayer_data(self, webpage, video_id, *args, **kwargs):
-
         # allow passing `transform_source` through to _find_jwplayer_data()
         transform_source = kwargs.pop('transform_source', None)
         kwfind = compat_kwargs({'transform_source': transform_source}) if transform_source else {}

youtube_dl/extractor/extractors.py

@@ -1235,6 +1235,11 @@ from .svt import (
     SVTPlayIE,
     SVTSeriesIE,
 )
+from .switchtube import (
+  SwitchTubeIE,
+  SwitchTubeProfileIE,
+  SwitchTubeChannelIE,
+)
 from .swrmediathek import SWRMediathekIE
 from .syfy import SyfyIE
 from .sztvhu import SztvHuIE

youtube_dl/extractor/palcomp3.py

@@ -8,7 +8,7 @@ from ..compat import compat_str
 from ..utils import (
     int_or_none,
     str_or_none,
-    try_get,
+    traverse_obj,
 )
 
 
@@ -109,7 +109,7 @@ class PalcoMP3ArtistIE(PalcoMP3BaseIE):
     }
     name'''
 
-    @ classmethod
+    @classmethod
     def suitable(cls, url):
         return False if re.match(PalcoMP3IE._VALID_URL, url) else super(PalcoMP3ArtistIE, cls).suitable(url)
 
@@ -118,7 +118,8 @@ class PalcoMP3ArtistIE(PalcoMP3BaseIE):
         artist = self._call_api(artist_slug, self._ARTIST_FIELDS_TMPL)['artist']
 
         def entries():
-            for music in (try_get(artist, lambda x: x['musics']['nodes'], list) or []):
+            for music in traverse_obj(artist, (
+                    'musics', 'nodes', lambda _, m: m['musicID'])):
                 yield self._parse_music(music)
 
         return self.playlist_result(
@@ -137,7 +138,7 @@ class PalcoMP3VideoIE(PalcoMP3BaseIE):
             'title': 'Maiara e Maraisa - Você Faz Falta Aqui - DVD Ao Vivo Em Campo Grande',
             'description': 'md5:7043342c09a224598e93546e98e49282',
             'upload_date': '20161107',
-            'uploader_id': 'maiaramaraisaoficial',
+            'uploader_id': '@maiaramaraisaoficial',
             'uploader': 'Maiara e Maraisa',
         }
     }]

youtube_dl/extractor/switchtube.py

@@ -0,0 +1,219 @@
+# coding: utf-8
+from __future__ import unicode_literals
+
+import itertools
+import re
+
+from .common import InfoExtractor
+from ..utils import (
+    clean_html,
+    extract_attributes,
+    float_or_none,
+    get_element_by_attribute,
+    int_or_none,
+    parse_iso8601,
+    strip_or_none,
+    urljoin,
+)
+
+
+class SwitchTubeIE(InfoExtractor):
+    _VALID_URL = r'https?://tube\.switch\.ch/videos/(?P<id>[\da-zA-Z]+)'
+    IE_NAME = 'switchtube'
+    _TESTS = [{
+        'url': 'https://tube.switch.ch/videos/0T1XfaIFSX',
+        'info_dict': {
+            'id': '0T1XfaIFSX',
+            'title': '2016_ASE_sqC03-Entretien',
+            'channel': 'ASE Assistant-e socio-éducatif-ve CFC',
+            'channel_url': 'https://tube.switch.ch/channels/bsaer76yoL',
+            'channel_id': 'bsaer76yoL',
+            'ext': 'mp4',
+            'description': None,
+            'thumbnail': r're:^https?://tube.switch.ch/image/representations/[\w-]+$',
+            'license': 'All rights reserved',
+            'creator': 'Jean-Marc Pouly from Eidgenössische Hochschule für Berufsbildung',
+            'uploader': 'Jean-Marc Pouly from Eidgenössische Hochschule für Berufsbildung',
+            'uploader_url': 'https://tube.switch.ch/profiles/42481',
+            'uploader_id': '42481',
+            'upload_date': '20220309',
+            'timestamp': 1646839068,
+        }
+    }, {
+        'url': 'https://tube.switch.ch/videos/0cf3886d',
+        'info_dict': {
+            'id': '0cf3886d',
+            'ext': 'mp4',
+            'title': 'Introduction: Mini-Batches in On- and Off-Policy Deep Reinforcement Learning',
+            'license': 'All rights reserved',
+            'description': 'One of the challenges in Deep Reinforcement Learning is to decorrelate the data. How this is possible with replay buffers is explained here.',
+            'thumbnail': r're:^https?://tube.switch.ch/image/representations/[\w-]+$',
+            'channel': 'CS-456 Artificial Neural Networks',
+            'channel_url': 'https://tube.switch.ch/channels/1deb03e0',
+            'channel_id': '1deb03e0',
+            'timestamp': 1590733406,
+            'upload_date': '20200529',
+            'creator': 'Wulfram Gerstner from École polytechnique fédérale de Lausanne (EPFL)',
+            'uploader': 'Wulfram Gerstner from École polytechnique fédérale de Lausanne (EPFL)',
+            'uploader_url': 'https://tube.switch.ch/profiles/94248',
+            'uploader_id': '94248',
+        }
+    }]
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url)
+        webpage = self._download_webpage(url, video_id)
+        title = self._html_search_regex(r'(?s)<title\b[^>]*>(.+?)</title>', webpage, 'title')
+
+        info = {
+            'id': video_id,
+            'title': title,
+            'is_live': False,
+        }
+
+        info['view_count'] = int_or_none(self._html_search_regex(r'(?s)<p\b[^>]*>\s*(\d+)\s+views?\s*</p>', webpage,
+                                                                 'view_count', default=None))
+        info['license'] = self._html_search_regex(r'''(?s)<span\b[^>]*?\bproperty\s*=\s*["'](?:\b[cd]c:license\b\s*){2,}[^>]+>(.+)</span>''',
+                                                  webpage, 'license', default=None)
+        info['description'] = strip_or_none(clean_html(
+            get_element_by_attribute('property', 'dc:description', webpage)))
+        info['duration'] = float_or_none(
+            self._search_regex(r'''\bdata-duration\s*=\s*["']([\d.]+)''',
+                               webpage, 'duration', default=None))
+
+        channel_groups = self._search_regex(r'''(?s)<a\b[^>]+?\bhref\s*=\s*["']/channels/(.{4,}?)</a>''',
+                                            webpage, 'channel groups', default='')
+        channel_groups = re.split(r'''(?s)["'][^>]*>\s*''', channel_groups, 1)
+        if len(channel_groups) == 2:
+            for i, ch in enumerate(('channel_id', 'channel')):
+                info[ch] = strip_or_none(channel_groups[i])
+            if info['channel_id']:
+                info['channel_url'] = 'https://tube.switch.ch/channels/' + info['channel_id']
+
+        def outer_elements_by_attribute(attr, value, html, tag=None, escape_value=True, include_inner=False):
+            """generate matching HTML element strings
+               if include_inner, tuples of (element, content)"""
+            pattern = r'''(?s)(?P<element><(%s)\b[^>]+?\b%s\s*=\s*("|')%s\b[^>]+>)(?P<inner>.*?)</\2>''' % \
+                (re.escape(tag) if tag is not None else r'\w+', attr, re.escape(value) if escape_value else value)
+            matches = re.finditer(pattern, html)
+            for m in matches:
+                yield m.group('element', 'inner') if include_inner else m.group('element')
+
+        for dt in outer_elements_by_attribute('property', 'dc:date', webpage, tag='span'):
+            dt = extract_attributes(dt)
+            if dt.get('class') == 'dt-published':
+                info['timestamp'] = parse_iso8601(dt.get('content'))
+                break
+
+        creator_groups = self._search_regex(
+            r'''(?s)<span\b[^>]+?\bclass\s*=\s*("|')(?:(?!\1).)*\bp-author\b(?:(?!\1).)*\1\s*property\s*=\s*["']dc:creator\b[^>]+>\s*(.*?<span\b[^>]+?\bclass\s*=\s*["']p-name\b.*</span>).*?</span>''',
+            webpage, 'creator groups', default='', group=2)
+        creator_groups = re.match(r'''(?s)<a\b[^>]+?\bhref\s*=\s*("|')/profiles/(?P<profile_id>.+?)\1[^>]*>\s*<span\b[^>]+>\s*(?P<creator_name>.+?)\s*</span>[,\s]*<span\b[^>]+\bclass\s*=\s*["']p-organization-name\b[^>]+>\s*(?P<organization_name>.+?)\s*</span>''', creator_groups)
+        if creator_groups:
+            creator_groups = creator_groups.groupdict()
+            info['uploader'] = info['creator'] = ' from '.join((creator_groups['creator_name'], creator_groups['organization_name']))
+            info['uploader_id'] = creator_groups['profile_id']
+            info['uploader_url'] = 'https://tube.switch.ch/profiles/' + info['uploader_id']
+
+        parsed_media_entries = self._parse_html5_media_entries(url, webpage, video_id)[0]
+        info['thumbnail'] = parsed_media_entries['thumbnail']
+        info['formats'] = parsed_media_entries['formats']
+        self._sort_formats(info['formats'])
+
+        return info
+
+
+class SwitchTubeProfileIE(InfoExtractor):
+    _VALID_URL = r'https?://tube\.switch\.ch/profiles/(?P<id>[\da-zA-Z]+)'
+    IE_NAME = 'switchtube:profile'
+    _TESTS = [{
+        'url': 'https://tube.switch.ch/profiles/94248',
+        'info_dict': {
+            'id': '94248',
+            'title': 'Wulfram Gerstner',
+            'description': None,
+        },
+        'playlist_mincount': 94,
+    }]
+
+    @classmethod
+    def suitable(cls, url):
+        return False if SwitchTubeIE.suitable(url) else super(SwitchTubeProfileIE, cls).suitable(url)
+
+    def _real_extract(self, url):
+        channel_id = self._match_id(url)
+        webpage = self._download_webpage(url, channel_id)
+        channel_title = self._html_search_regex(r'(?s)<title\b[^>]*>(.+?)</title>', webpage, 'title', default=None)
+
+        description = self._html_search_regex(r'''(?s)<div\b[^>]+class\s*=\s*("|')p-summary formatted\1[^>]+property\s*=\s*("|')dc:description\2[^>]*>\s*(.+?)\s*</div>''',
+                                              webpage, 'description', default=None, group=3)
+
+        entries = []
+        next_page = None
+        for current_page_number in itertools.count(1):
+            if next_page:
+                webpage = self._download_webpage(next_page, channel_id, note='Downloading page %d' % (current_page_number, ))
+
+            for _, video_path, video_id, title in re.findall(
+                    r'''(?s)<a\b[^>]+\bhref\s*=\s*("|')(/videos/((?:(?!\1).)+?))\1[^>]*>\s*<div\b[^>]+\bclass\s*=\s*["']title\b[^>]+>(.+?)</div>''',
+                    webpage):
+                video_url = urljoin(url, video_path)
+                if video_url:
+                    entries.append(self.url_result(video_url, ie=SwitchTubeIE.ie_key(), video_id=video_id))
+
+            next_page = self._search_regex(
+                r'''<a\b[^>]+?\bhref\s*=\s*("|')(?P<path>/profiles/%s\?(?:(?!\1).)+)\1[^>]*>\s*Next\s*</a>''' % (channel_id,),
+                webpage, 'next page', group='path', default=None)
+            if next_page:
+                next_page = urljoin(url, next_page)
+            if not next_page:
+                break
+
+        return self.playlist_result(entries, channel_id, channel_title,
+                                    description)
+
+
+class SwitchTubeChannelIE(InfoExtractor):
+    _VALID_URL = r'https?://tube\.switch\.ch/channels/(?P<id>[\da-zA-Z]+)'
+    IE_NAME = 'switchtube:channel'
+    _TESTS = [{
+        'url': 'https://tube.switch.ch/channels/1deb03e0',
+        'info_dict': {
+            'id': '1deb03e0',
+            'title': 'CS-456 Artificial Neural Networks',
+            'description': 'Class on Artificial Neural Networks and Reinforcement Learning designed for EPFL master students in CS and related disciplines.'
+        },
+        'playlist_mincount': 94,
+    }]
+
+    @classmethod
+    def suitable(cls, url):
+        return False if SwitchTubeIE.suitable(url) else super(SwitchTubeChannelIE, cls).suitable(url)
+
+    def _real_extract(self, url):
+        channel_id = self._match_id(url)
+
+        entries = []
+        for current_page_number in itertools.count(0):
+            page_url = urljoin(url, '/channels/%s?order=episodes&page=%d' % (channel_id, current_page_number))
+            webpage = self._download_webpage(page_url, channel_id, note='Downloading page %d' % (current_page_number + 1, ))
+
+            if current_page_number == 0:
+                channel_title = self._html_search_regex(r'(?s)<title\b[^>]*>(.+?)</title>', webpage, 'title', default=None)
+                description = self._html_search_regex(r'''(?s)<div\b[^>]+class\s*=\s*("|')description formatted\1[^>]*>\s*<p>\s*(.+?)\s*</p>\s*</div>''',
+                                                      webpage, 'description', default=None, group=2)
+
+            this_page_still_has_something = False
+            for _, video_path, video_id in re.findall(
+                    r'''(?s)<a\b[^>]+\bhref\s*=\s*("|')(/videos/((?:(?!\1).)+?))\1[^>]*>''',
+                    webpage):
+                video_url = urljoin(url, video_path)
+                if video_url:
+                    this_page_still_has_something = True
+                    entries.append(self.url_result(video_url, ie=SwitchTubeIE.ie_key(), video_id=video_id))
+
+            if not this_page_still_has_something:
+                break
+
+        return self.playlist_result(entries, channel_id, channel_title,
+                                    description)

youtube_dl/jsinterp.py

@@ -14,6 +14,7 @@ from .utils import (
     remove_quotes,
     unified_timestamp,
     variadic,
+    write_string,
 )
 from .compat import (
     compat_basestring,
@@ -53,15 +54,16 @@ def wraps_op(op):
 
 # NB In principle NaN cannot be checked by membership.
 # Here all NaN values are actually this one, so _NaN is _NaN,
-# although _NaN != _NaN.
+# although _NaN != _NaN. Ditto Infinity.
 
 _NaN = float('nan')
+_Infinity = float('inf')
 
 
 def _js_bit_op(op):
 
     def zeroise(x):
-        return 0 if x in (None, JS_Undefined, _NaN) else x
+        return 0 if x in (None, JS_Undefined, _NaN, _Infinity) else x
 
     @wraps_op(op)
     def wrapped(a, b):
@@ -84,7 +86,7 @@ def _js_arith_op(op):
 def _js_div(a, b):
     if JS_Undefined in (a, b) or not (a or b):
         return _NaN
-    return operator.truediv(a or 0, b) if b else float('inf')
+    return operator.truediv(a or 0, b) if b else _Infinity
 
 
 def _js_mod(a, b):
@@ -220,6 +222,42 @@ class LocalNameSpace(ChainMap):
         return 'LocalNameSpace%s' % (self.maps, )
 
 
+class Debugger(object):
+    ENABLED = False
+
+    @staticmethod
+    def write(*args, **kwargs):
+        level = kwargs.get('level', 100)
+
+        def truncate_string(s, left, right=0):
+            if s is None or len(s) <= left + right:
+                return s
+            return '...'.join((s[:left - 3], s[-right:] if right else ''))
+
+        write_string('[debug] JS: {0}{1}\n'.format(
+            '  ' * (100 - level),
+            ' '.join(truncate_string(compat_str(x), 50, 50) for x in args)))
+
+    @classmethod
+    def wrap_interpreter(cls, f):
+        def interpret_statement(self, stmt, local_vars, allow_recursion, *args, **kwargs):
+            if cls.ENABLED and stmt.strip():
+                cls.write(stmt, level=allow_recursion)
+            try:
+                ret, should_ret = f(self, stmt, local_vars, allow_recursion, *args, **kwargs)
+            except Exception as e:
+                if cls.ENABLED:
+                    if isinstance(e, ExtractorError):
+                        e = e.orig_msg
+                    cls.write('=> Raises:', e, '<-|', stmt, level=allow_recursion)
+                raise
+            if cls.ENABLED and stmt.strip():
+                if should_ret or not repr(ret) == stmt:
+                    cls.write(['->', '=>'][should_ret], repr(ret), '<-|', stmt, level=allow_recursion)
+            return ret, should_ret
+        return interpret_statement
+
+
 class JSInterpreter(object):
     __named_object_counter = 0
 
@@ -307,8 +345,7 @@ class JSInterpreter(object):
     def __op_chars(cls):
         op_chars = set(';,[')
         for op in cls._all_operators():
-            for c in op[0]:
-                op_chars.add(c)
+            op_chars.update(op[0])
         return op_chars
 
     def _named_object(self, namespace, obj):
@@ -326,9 +363,8 @@ class JSInterpreter(object):
         # collections.Counter() is ~10% slower in both 2.7 and 3.9
         counters = dict((k, 0) for k in _MATCHING_PARENS.values())
         start, splits, pos, delim_len = 0, 0, 0, len(delim) - 1
-        in_quote, escaping, skipping = None, False, 0
-        after_op, in_regex_char_group = True, False
-
+        in_quote, escaping, after_op, in_regex_char_group = None, False, True, False
+        skipping = 0
         for idx, char in enumerate(expr):
             paren_delta = 0
             if not in_quote:
@@ -382,10 +418,12 @@ class JSInterpreter(object):
         return separated[0][1:].strip(), separated[1].strip()
 
     @staticmethod
-    def _all_operators():
-        return itertools.chain(
-            # Ref: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Operator_Precedence
-            _SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS)
+    def _all_operators(_cached=[]):
+        if not _cached:
+            _cached.extend(itertools.chain(
+                # Ref: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Operators/Operator_Precedence
+                _SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS))
+        return _cached
 
     def _operator(self, op, left_val, right_expr, expr, local_vars, allow_recursion):
         if op in ('||', '&&'):
@@ -416,7 +454,7 @@ class JSInterpreter(object):
         except Exception as e:
             if allow_undefined:
                 return JS_Undefined
-            raise self.Exception('Cannot get index {idx:.100}'.format(**locals()), expr=repr(obj), cause=e)
+            raise self.Exception('Cannot get index {idx!r:.100}'.format(**locals()), expr=repr(obj), cause=e)
 
     def _dump(self, obj, namespace):
         try:
@@ -438,6 +476,7 @@ class JSInterpreter(object):
     _FINALLY_RE = re.compile(r'finally\s*\{')
     _SWITCH_RE = re.compile(r'switch\s*\(')
 
+    @Debugger.wrap_interpreter
     def interpret_statement(self, stmt, local_vars, allow_recursion=100):
         if allow_recursion < 0:
             raise self.Exception('Recursion limit reached')
@@ -511,7 +550,6 @@ class JSInterpreter(object):
                 expr = self._dump(inner, local_vars) + outer
 
         if expr.startswith('('):
-
             m = re.match(r'\((?P<d>[a-z])%(?P<e>[a-z])\.length\+(?P=e)\.length\)%(?P=e)\.length', expr)
             if m:
                 # short-cut eval of frequently used `(d%e.length+e.length)%e.length`, worth ~6% on `pytest -k test_nsig`
@@ -693,7 +731,7 @@ class JSInterpreter(object):
                 (?P<op>{_OPERATOR_RE})?
                 =(?!=)(?P<expr>.*)$
             )|(?P<return>
-                (?!if|return|true|false|null|undefined)(?P<name>{_NAME_RE})$
+                (?!if|return|true|false|null|undefined|NaN|Infinity)(?P<name>{_NAME_RE})$
             )|(?P<indexing>
                 (?P<in>{_NAME_RE})\[(?P<idx>.+)\]$
             )|(?P<attribute>
@@ -727,11 +765,12 @@ class JSInterpreter(object):
             raise JS_Break()
         elif expr == 'continue':
             raise JS_Continue()
-
         elif expr == 'undefined':
             return JS_Undefined, should_return
         elif expr == 'NaN':
             return _NaN, should_return
+        elif expr == 'Infinity':
+            return _Infinity, should_return
 
         elif md.get('return'):
             return local_vars[m.group('name')], should_return
@@ -760,18 +799,28 @@ class JSInterpreter(object):
             right_expr = separated.pop()
             # handle operators that are both unary and binary, minimal BODMAS
             if op in ('+', '-'):
+                # simplify/adjust consecutive instances of these operators
                 undone = 0
                 while len(separated) > 1 and not separated[-1].strip():
                     undone += 1
                     separated.pop()
                 if op == '-' and undone % 2 != 0:
                     right_expr = op + right_expr
+                elif op == '+':
+                    while len(separated) > 1 and separated[-1].strip() in self.OP_CHARS:
+                        right_expr = separated.pop() + right_expr
+                # hanging op at end of left => unary + (strip) or - (push right)
                 left_val = separated[-1]
                 for dm_op in ('*', '%', '/', '**'):
                     bodmas = tuple(self._separate(left_val, dm_op, skip_delims=skip_delim))
                     if len(bodmas) > 1 and not bodmas[-1].strip():
                         expr = op.join(separated) + op + right_expr
-                        right_expr = None
+                        if len(separated) > 1:
+                            separated.pop()
+                            right_expr = op.join((left_val, right_expr))
+                        else:
+                            separated = [op.join((left_val, right_expr))]
+                            right_expr = None
                         break
                 if right_expr is None:
                     continue
@@ -797,6 +846,8 @@ class JSInterpreter(object):
 
             def eval_method():
                 if (variable, member) == ('console', 'debug'):
+                    if Debugger.ENABLED:
+                        Debugger.write(self.interpret_expression('[{}]'.format(arg_str), local_vars, allow_recursion))
                     return
                 types = {
                     'String': compat_str,

youtube_dl/utils.py

@@ -2406,7 +2406,7 @@ class ExtractorError(YoutubeDLError):
         """ tb, if given, is the original traceback (so that it can be printed out).
         If expected is set, this is a normal error message and most likely not a bug in youtube-dl.
         """
-
+        self.orig_msg = msg
         if sys.exc_info()[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError):
             expected = True
         if video_id is not None: