Merge fed9c8d17b 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;}')
-        # 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

@@ -1242,6 +1242,10 @@ from .tagesschau import (
     TagesschauPlayerIE,
     TagesschauIE,
 )
+from .talktv import (
+    TalkTVIE,
+    TalkTVSeriesIE,
+)
 from .tass import TassIE
 from .tbs import TBSIE
 from .tdslifeway import TDSLifewayIE

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/talktv.py

@@ -0,0 +1,178 @@
+# coding: utf-8
+from __future__ import unicode_literals
+
+import re
+import calendar
+from datetime import datetime
+import time
+
+from .common import InfoExtractor
+
+from ..compat import compat_str
+from ..utils import (
+    clean_html,
+    determine_ext,
+    extract_attributes,
+    ExtractorError,
+    get_elements_by_class,
+    HEADRequest,
+    parse_duration,
+    try_get,
+    unified_timestamp,
+    url_or_none,
+    urljoin,
+)
+
+
+class TalkTVIE(InfoExtractor):
+    IE_NAME = 'talk.tv'
+    IE_DESC = 'TalkTV UK catch-up and live shows'
+    _VALID_URL = r'https?://watch\.talk\.tv/(?P<id>watch/(?:vod|replay)/\d+|live)'
+    _TESTS = [{
+        'url': 'https://watch.talk.tv/watch/replay/12216792',
+        'md5': 'dc9071f7d26f48ce4057a98425894eb3',
+        'info_dict': {
+            'id': '12216792',
+            'ext': 'mp4',
+            'title': 'Piers Morgan Uncensored',
+            'description': 'The host interviews former US President Donald Trump',
+            'timestamp': 1650917390,
+            'upload_date': '20220425',
+            'duration': float,
+        },
+        'params': {
+            'skip_download': True,  # adaptive download
+        },
+    }, {
+        'url': 'https://watch.talk.tv/live',
+        'info_dict': {
+            'id': 'live',
+            'ext': 'mp4',
+            'title': 'Piers Morgan Uncensored',
+            'description': compat_str,
+            'timestamp': int,
+            # needs core fix to force compat_str type
+            'upload_date': r're:\d{8}',
+            'duration': float,
+        },
+        'params': {
+            'skip_download': True,
+        },
+    },
+    ]
+
+    def _real_extract(self, url):
+        video_id = self._match_id(url).rsplit('/', 1)[-1]
+        is_live = (video_id == 'live')
+        webpage = self._download_webpage(url, video_id)
+        title = self._html_search_regex(r'(?is)<h1\b[^>]+>\s*(.+?)\s*</h1', webpage, 'title')
+        player = self._search_regex(r'''(<[dD][iI][vV]\b[^>]+?\bid\s*=\s*(?P<q>"|')player(?P=q)[^>]*>)''', webpage, video_id)
+        player = extract_attributes(player)
+        expiry = player.get('expiry')
+        if expiry is not None and expiry < time.time():
+            raise ExtractorError('Video has expired', expected=True)
+        api_info = self._download_json(
+            'https://mm-v2.simplestream.com/ssmp/api.php?id=%(data-id)s&env=%(data-env)s' % player,
+            video_id, note='Downloading API info', fatal=False)
+        player['api_url'] = (
+            url_or_none(try_get(api_info, lambda x: x['response']['api_hostname']))
+            or 'https://v2-streams-elb.simplestreamcdn.com')
+        headers = {'Referer': url, }
+        for item in ('uvid', 'token', ('expiry', 'Token-Expiry')):
+            if isinstance(item, compat_str):
+                name = item.capitalize()
+            else:
+                item, name = item
+            val = player.get('data-' + item)
+            if val is not None:
+                headers[name] = val
+
+        stream_info = self._download_json(
+            '%(api_url)s/api/%(data-type)s/stream/%(data-uvid)s?key=%(data-key)s&platform=firefox&cc=%(data-country)s' % player,
+            video_id, headers=headers)
+        error = try_get(stream_info, lambda x: x['response']['error'])
+        if error:
+            raise ExtractorError('Streaming API reported: ' + error, expected=True)
+        fmt_url = (stream_info['response'].get('drm') in (None, False)) and stream_info['response']['stream']
+        formats = []
+        duration = None
+        description = None
+        timestamp = None
+        if fmt_url:
+            ext = determine_ext(fmt_url)
+            if ext == 'm3u8':
+                formats.extend(self._extract_m3u8_formats(
+                    fmt_url, video_id, 'mp4', entry_protocol='m3u8_native',
+                    m3u8_id='hls', live=is_live, fatal=False))
+            elif ext == 'mpd':
+                formats.extend(self._extract_mpd_formats(
+                    fmt_url, video_id, mpd_id='dash', live=is_live, fatal=False))
+            else:
+                formats.append({
+                    'url': fmt_url,
+                })
+            if not is_live:
+                res = self._request_webpage(HEADRequest(fmt_url), video_id, note='Checking date', fatal=False)
+                if res is not False:
+                    timestamp = unified_timestamp(res.info().getheader('last-modified'))
+
+        self._sort_formats(formats)
+
+        text_fields = get_elements_by_class('text-start', webpage)
+        for text in text_fields:
+            text = clean_html(text)
+            if text.startswith('EPISODE'):
+                duration = parse_duration(
+                    self._html_search_regex(r'^EPISODE\b\W*(\w[\w\s]*?)\s*$', text, 'duration', default=None))
+            elif text.startswith('Live'):
+                duration = self._html_search_regex(r'^Live\b(?:<[^>]+>|\W)*([0-2]?\d:\d{2}\s*-\s*[0-2]?\d:\d{2})\s*$', text, 'duration', default=None)
+                duration = list(map(lambda x: datetime.strptime(x, '%H:%M'), re.split(r'\s*-\s*', duration)))
+                if None not in duration and len(duration) == 2:
+                    timestamp = datetime.now().replace(hour=duration[0].hour, minute=duration[0].minute, second=0, microsecond=0)
+                    timestamp = calendar.timegm(timestamp.timetuple())
+                    duration = duration[1] - duration[0]
+                    try:
+                        duration = duration.total_seconds()
+                    except AttributeError:
+                        # Py 2.6
+                        duration = duration.td_seconds
+                    if duration is not None and duration < 0:
+                        duration += 24 * 3600
+            else:
+                description = text
+
+        return {
+            # ensure live has a fixed ID
+            'id': player['data-uvid'] if not is_live else video_id,
+            'title': title,
+            'display_id': video_id if not is_live else player['data-uvid'],
+            'formats': formats,
+            'thumbnail': player.get('data-poster'),
+            'duration': duration,
+            'timestamp': timestamp,
+            'description': description,
+            'is_live': is_live,
+        }
+
+
+class TalkTVSeriesIE(InfoExtractor):
+    IE_NAME = 'talk.tv:series'
+    IE_DESC = 'TalkTV UK series catch-up'
+    _VALID_URL = r'https?://(?:watch\.|www\.)?talk\.tv/shows/(?P<id>[\da-f]{8}-(?:[\da-f]{4}-){3}[\da-f]{12})'
+    _TESTS = [{
+        'url': 'https://watch.talk.tv/shows/86dadc3e-c4d2-11ec-b4c6-0af62ebc70d1',
+        'info_dict': {
+            'id': '86dadc3e-c4d2-11ec-b4c6-0af62ebc70d1',
+        },
+        'playlist_mincount': 4,
+    },
+    ]
+
+    def _real_extract(self, url):
+        playlist_id = self._match_id(url)
+        webpage = self._download_webpage(url, playlist_id)
+        episodes = re.finditer(
+            r'''(?i)<a\b[^>]+?\bhref\s*=\s*(?P<q>"|')(?P<href>/watch/(?:(?!(?P=q)).)+)(?P=q)''',
+            webpage)
+        return self.playlist_from_matches(
+            episodes, playlist_id, getter=lambda x: urljoin(url, x.group('href')), ie='TalkTV')

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: