Merge fad516e22e into 4d05f84325

* no space after @ in decorator

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;}')
-        # jsi = JSInterpreter('function f(){return 2 + - + - - 2;}')
+        self.assertEqual(jsi.call_function('f'), 0)
-        # 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/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/tver.py

@@ -4,58 +4,229 @@ from __future__ import unicode_literals
 import re
 
 from .common import InfoExtractor
-from ..compat import compat_str
+from ..compat import (
-from ..utils import (
+    compat_kwargs,
-    int_or_none,
+    compat_str,
-    remove_start,
-    smuggle_url,
-    try_get,
 )
+from ..utils import (
+    ExtractorError,
+    int_or_none,
+    smuggle_url,
+    str_or_none,
+)
+
+try:
+    from ..utils import traverse_obj
+except ImportError:
+    from ..compat import compat_collections_abc
+
+    def traverse_obj(obj, *path_list, **kw):
+        ''' Traverse nested list/dict/tuple'''
+
+        # parameter defaults
+        default = kw.get('default')
+        expected_type = kw.get('expected_type')
+        get_all = kw.get('get_all', True)
+        casesense = kw.get('casesense', True)
+        is_user_input = kw.get('is_user_input', False)
+        traverse_string = kw.get('traverse_string', False)
+
+        def variadic(x, allowed_types=(compat_str, bytes)):
+            return x if isinstance(x, compat_collections_abc.Iterable) and not isinstance(x, allowed_types) else (x,)
+
+        def listish(l):
+            return isinstance(l, (list, tuple))
+
+        def from_iterable(iterables):
+            # chain.from_iterable(['ABC', 'DEF']) --> A B C D E F
+            for it in iterables:
+                for element in it:
+                    yield element
+
+        class Nonlocal:
+            pass
+        nl = Nonlocal()
+
+        if not casesense:
+            _lower = lambda k: (k.lower() if isinstance(k, compat_str) else k)
+            path_list = (map(_lower, variadic(path)) for path in path_list)
+
+        def _traverse_obj(obj, path, _current_depth=0):
+            path = tuple(variadic(path))
+            for i, key in enumerate(path):
+                if obj is None:
+                    return None
+                if listish(key):
+                    obj = [_traverse_obj(obj, sub_key, _current_depth) for sub_key in key]
+                    key = Ellipsis
+                if key is Ellipsis:
+                    obj = (obj.values() if isinstance(obj, dict)
+                           else obj if listish(obj)
+                           else compat_str(obj) if traverse_string else [])
+                    _current_depth += 1
+                    nl.depth = max(nl.depth, _current_depth)
+                    return [_traverse_obj(inner_obj, path[i + 1:], _current_depth) for inner_obj in obj]
+                elif callable(key):
+                    if listish(obj):
+                        obj = enumerate(obj)
+                    elif isinstance(obj, dict):
+                        obj = obj.items()
+                    else:
+                        if not traverse_string:
+                            return None
+                        obj = str(obj)
+                    _current_depth += 1
+                    nl.depth = max(nl.depth, _current_depth)
+                    return [_traverse_obj(v, path[i + 1:], _current_depth) for k, v in obj if key(k)]
+                elif isinstance(obj, dict) and not (is_user_input and key == ':'):
+                    obj = (obj.get(key) if casesense or (key in obj)
+                           else next((v for k, v in obj.items() if _lower(k) == key), None))
+                else:
+                    if is_user_input:
+                        key = (int_or_none(key) if ':' not in key
+                               else slice(*map(int_or_none, key.split(':'))))
+                        if key == slice(None):
+                            return _traverse_obj(obj, tuple([Ellipsis] + list(path[i + 1:])), _current_depth)
+                    if not isinstance(key, (int, slice)):
+                        return None
+                    if not listish(obj):
+                        if not traverse_string:
+                            return None
+                        obj = compat_str(obj)
+                    try:
+                        obj = obj[key]
+                    except IndexError:
+                        return None
+            return obj
+
+        if isinstance(expected_type, type):
+            type_test = lambda val: val if isinstance(val, expected_type) else None
+        elif expected_type is not None:
+            type_test = expected_type
+        else:
+            type_test = lambda val: val
+
+        for path in path_list:
+            nl.depth = 0
+            val = _traverse_obj(obj, path)
+            if val is not None:
+                if nl.depth:
+                    for _ in range(nl.depth - 1):
+                        val = from_iterable(v for v in val if v is not None)
+                    val = [v for v in map(type_test, val) if v is not None]
+                    if val:
+                        return val if get_all else val[0]
+                else:
+                    val = type_test(val)
+                    if val is not None:
+                        return val
+        return default
 
 
 class TVerIE(InfoExtractor):
-    _VALID_URL = r'https?://(?:www\.)?tver\.jp/(?P<path>(?:corner|episode|feature)/(?P<id>f?\d+))'
+    _VALID_URL = r'https?://(?:www\.)?tver\.jp/(?:(?P<type>lp|corner|series|episodes?|feature|tokyo2020/video)/)+(?P<id>[a-zA-Z0-9]+)'
     # videos are only available for 7 days
     _TESTS = [{
-        'url': 'https://tver.jp/corner/f0062178',
+        'skip': 'videos are only available for 7 days',
+        'url': 'https://tver.jp/episodes/ephss8yveb',
+        'info_dict': {
+            'id': 'ref:af03de03-21ac-4b98-a53b-9ffd9b102e92',
+            'ext': 'mp4',
+            'title': '#44　料理と値段と店主にびっくり　オモてなしすぎウマい店　2時間SP',
+            'description': 'md5:66985373a66fed8ad3cd595a3cfebb13',
+            'upload_date': '20220329',
+            'uploader_id': '4394098882001',
+            'timestamp': 1648527032,
+        },
+        'add_ie': ['BrightcoveNew'],
+    }, {
+        'skip': 'videos are only available for 7 days',
+        'url': 'https://tver.jp/lp/episodes/ep6f16g26p',
+        'info_dict': {
+            'id': '6302378806001',
+            'ext': 'mp4',
+            # "April 11 (Mon) 23:06-Broadcast scheduled": sorry but this is "correct"
+            'title': '4月11日(月)23時06分 ~ 放送予定',
+            'description': 'md5:4029cc5f4b1e8090dfc5b7bd2bc5cd0b',
+            'upload_date': '20220331',
+            'uploader_id': '3971130137001',
+            'timestamp': 1648696456,
+        },
+        'add_ie': ['BrightcoveNew'],
+    }, {
+        'url': 'https://tver.jp/corner/f0103888',
         'only_matching': True,
     }, {
-        'url': 'https://tver.jp/feature/f0062413',
+        'url': 'https://tver.jp/lp/f0033031',
-        'only_matching': True,
-    }, {
-        'url': 'https://tver.jp/episode/79622438',
-        'only_matching': True,
-    }, {
-        # subtitle = ' '
-        'url': 'https://tver.jp/corner/f0068870',
         'only_matching': True,
     }]
-    _TOKEN = None
     BRIGHTCOVE_URL_TEMPLATE = 'http://players.brightcove.net/%s/default_default/index.html?videoId=%s'
+    _PLATFORM_UID = None
+    _PLATFORM_TOKEN = None
+
+    def _download_json(self, url_or_request, video_id, **kwargs):
+
+        headers = {
+            'Origin': 'https://s.tver.jp',
+            'Referer': 'https://s.tver.jp/',
+        }
+        headers.update(kwargs.get('headers', {}))
+        kwargs.update(compat_kwargs({'headers': headers, }))
+
+        return super(TVerIE, self)._download_json(url_or_request, video_id, **kwargs)
 
     def _real_initialize(self):
-        self._TOKEN = self._download_json(
+        create_response = self._download_json(
-            'https://tver.jp/api/access_token.php', None)['token']
+            'https://platform-api.tver.jp/v2/api/platform_users/browser/create', None,
+            note='Creating session', data=b'device_type=pc', headers={
+                'Content-Type': 'application/x-www-form-urlencoded',
+            })
+        self._PLATFORM_UID = traverse_obj(create_response, ('result', 'platform_uid'), expected_type=compat_str)
+        self._PLATFORM_TOKEN = traverse_obj(create_response, ('result', 'platform_token'), expected_type=compat_str)
 
     def _real_extract(self, url):
-        path, video_id = re.match(self._VALID_URL, url).groups()
+        video_id = self._match_id(url)
-        main = self._download_json(
-            'https://api.tver.jp/v4/' + path, video_id,
-            query={'token': self._TOKEN})['main']
-        p_id = main['publisher_id']
-        service = remove_start(main['service'], 'ts_')
 
-        r_id = main['reference_id']
+        video_id, video_type = re.match(self._VALID_URL, url).group('id', 'type')
-        if service not in ('tx', 'russia2018', 'sebare2018live', 'gorin'):
+        if video_type not in ('episode', 'episodes', 'series'):
+            webpage = self._download_webpage(url, video_id, note='Resolving to new URL')
+            video_id = self._match_id(self._search_regex(
+                (r'''canonical['"]\s+href\s*=\s*(?P<q>'|")(?P<url>https?://tver\.jp/(?!(?P=q)).+?)(?P=q)''',
+                 r'&link=(?P<url>https?://tver\.jp/(?!(?P=q)).+?)[?&]'),
+                webpage, 'url regex', group='url'))
+        video_info = self._download_json(
+            'https://statics.tver.jp/content/episode/{0}.json'.format(video_id), video_id,
+            query={'v': '5'})
+        p_id = traverse_obj(video_info, ('video', 'accountID'), expected_type=compat_str)
+        r_id = traverse_obj(video_info, ('video', ('videoRefID', 'videoID')), get_all=False, expected_type=compat_str)
+        if None in (p_id, r_id):
+            raise ExtractorError(
+                'Failed to extract '
+                + ', '.join(
+                    (x[0] for x in (('accountID', p_id), ('videoRefID', r_id), )
+                     if x[1] is None)),
+                expected=False)
+        if not r_id.isdigit():
             r_id = 'ref:' + r_id
-        bc_url = smuggle_url(
+
-            self.BRIGHTCOVE_URL_TEMPLATE % (p_id, r_id),
+        additional_info = self._download_json(
-            {'geo_countries': ['JP']})
+            'https://platform-api.tver.jp/service/api/v1/callEpisode/{0}?require_data=mylist,later[epefy106ur],good[epefy106ur],resume[epefy106ur]'.format(video_id),
+            video_id, fatal=False,
+            query={
+                'platform_uid': self._PLATFORM_UID,
+                'platform_token': self._PLATFORM_TOKEN,
+            }, headers={
+                'x-tver-platform-type': 'web'
+            })
 
         return {
             '_type': 'url_transparent',
-            'description': try_get(main, lambda x: x['note'][0]['text'], compat_str),
+            'title': str_or_none(video_info.get('title')),
-            'episode_number': int_or_none(try_get(main, lambda x: x['ext']['episode_number'])),
+            'description': str_or_none(video_info.get('description')),
-            'url': bc_url,
+            'url': smuggle_url(
+                self.BRIGHTCOVE_URL_TEMPLATE % (p_id, r_id), {'geo_countries': ['JP']}),
+            'series': traverse_obj(
+                additional_info, ('result', ('episode', 'series'), 'content', ('seriesTitle', 'title')),
+                get_all=False, expected_type=compat_str),
             'ie_key': 'BrightcoveNew',
         }

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.update(op[0])
-                op_chars.add(c)
         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
+        in_quote, escaping, after_op, in_regex_char_group = None, False, True, False
-        after_op, in_regex_char_group = 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():
+    def _all_operators(_cached=[]):
-        return itertools.chain(
+        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)
+                _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
@@ -760,17 +799,27 @@ 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
+                        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:
@@ -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: