Merge 1204472ade 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/ard.py

@@ -359,7 +359,9 @@ class ARDIE(InfoExtractor):
 
 
 class ARDBetaMediathekIE(ARDMediathekBaseIE):
-    _VALID_URL = r'https://(?:(?:beta|www)\.)?ardmediathek\.de/(?:[^/]+/)?(?:player|live|video)/(?:[^/]+/)*(?P<id>Y3JpZDovL[a-zA-Z0-9]+)'
+
+    _VALID_URL = r'https://(?:(?:beta|www)\.)?ardmediathek\.de/(((?:[^/]+/)?(?:player|live|video|serie|sendung)/(?:[^/]+/)*(?P<id>Y3JpZDovL[a-zA-Z0-9]+))|(((?P<sender>[a-zA-Z0-9\-]+)([/]))?(?P<name>[a-zA-Z0-9\-]+)))'
+
     _TESTS = [{
         'url': 'https://www.ardmediathek.de/mdr/video/die-robuste-roswita/Y3JpZDovL21kci5kZS9iZWl0cmFnL2Ntcy84MWMxN2MzZC0wMjkxLTRmMzUtODk4ZS0wYzhlOWQxODE2NGI/',
         'md5': 'a1dc75a39c61601b980648f7c9f9f71d',
@@ -395,73 +397,189 @@ class ARDBetaMediathekIE(ARDMediathekBaseIE):
     }, {
         'url': 'https://www.ardmediathek.de/ard/player/Y3JpZDovL3dkci5kZS9CZWl0cmFnLWQ2NDJjYWEzLTMwZWYtNGI4NS1iMTI2LTU1N2UxYTcxOGIzOQ/tatort-duo-koeln-leipzig-ihr-kinderlein-kommet',
         'only_matching': True,
+    }, {
+        'url': 'https://ardmediathek.de/sendung/saartalk/saartalk-gesellschaftsgift-haltung-gegen-hass/sr-fernsehen/Y3JpZDovL3NyLW9ubGluZS5kZS9TVF84MTY4MA/',
+        'only_matching': True,
+    }, {
+        'url': 'https://ardmediathek.de/serie/saartalk/saartalk-gesellschaftsgift-haltung-gegen-hass/sr-fernsehen/Y3JpZDovL3NyLW9ubGluZS5kZS9TVF84MTY4MA/',
+        'only_matching': True,
+    }, {
+        'url': 'https://www.ardmediathek.de/br/dahoam-is-dahoam',
+        'only_matching': True,
     }]
 
     def _real_extract(self, url):
         video_id = self._match_id(url)
+        video_name = self._match_id(url, group_name='name')
+        sender = self._match_id(url, group_name='sender')
+
+        if '/serie/' in url or '/sendung/' in url:
+            return self._real_extract_serie(video_id)
+        elif 'none' != video_name.lower():
+            return self._real_extract_named_serie(video_name, sender if 'none' != sender.lower() else "ard")
+        else:
+            return self._real_extract_video(video_id)
+
+    def _real_extract_video(self, video_id):
 
         player_page = self._download_json(
-            'https://api.ardmediathek.de/public-gateway',
-            video_id, data=json.dumps({
-                'query': '''{
-  playerPage(client: "ard", clipId: "%s") {
-    blockedByFsk
-    broadcastedOn
-    maturityContentRating
-    mediaCollection {
-      _duration
-      _geoblocked
-      _isLive
-      _mediaArray {
-        _mediaStreamArray {
-          _quality
-          _server
-          _stream
-        }
-      }
-      _previewImage
-      _subtitleUrl
-      _type
-    }
-    show {
-      title
-    }
-    synopsis
-    title
-    tracking {
-      atiCustomVars {
-        contentId
-      }
-    }
-  }
-}''' % video_id,
-            }).encode(), headers={
-                'Content-Type': 'application/json'
-            })['data']['playerPage']
+                f'https://api.ardmediathek.de/page-gateway/pages/ard/item/{video_id}',
+                video_id
+        )
+
         title = player_page['title']
-        content_id = str_or_none(try_get(
-            player_page, lambda x: x['tracking']['atiCustomVars']['contentId']))
-        media_collection = player_page.get('mediaCollection') or {}
+        content_id = str_or_none(
+                try_get(
+                        player_page, lambda x: x['tracking']['atiCustomVars']['contentId']
+                )
+        )
+        media_collection = player_page['widgets'][0].get('mediaCollection') or {}
         if not media_collection and content_id:
             media_collection = self._download_json(
                     'https://www.ardmediathek.de/play/media/' + content_id,
-                content_id, fatal=False) or {}
+                    content_id, fatal=False
+            ) or {}
+
         info = self._parse_media_info(
-            media_collection, content_id or video_id,
-            player_page.get('blockedByFsk'))
+            media_collection['embedded'], content_id or video_id,
+            player_page['widgets'][0].get('blockedByFsk')
+        )
+
         age_limit = None
-        description = player_page.get('synopsis')
-        maturity_content_rating = player_page.get('maturityContentRating')
+        description = player_page['widgets'][0].get('synopsis')
+        maturity_content_rating = player_page['widgets'][0].get('maturityContentRating')
         if maturity_content_rating:
             age_limit = int_or_none(maturity_content_rating.lstrip('FSK'))
         if not age_limit and description:
-            age_limit = int_or_none(self._search_regex(
-                r'\(FSK\s*(\d+)\)\s*$', description, 'age limit', default=None))
-        info.update({
+            age_limit = int_or_none(
+                self._search_regex(
+                    r'\(FSK\s*(\d+)\)\s*$', description, 'age limit', default=None
+                )
+            )
+
+        session_episode_match = re.search(r"\(S(\d+)/E(\d+)\)", title)
+        episode_match = re.search(r"\((\d+)\)", title)
+        episode_name_match = re.search(r"(Folge\s\d+)", title)
+
+        if session_episode_match:
+            season_number = session_episode_match.group(1)
+            episode_number = session_episode_match.group(2)
+
+            alt_title = re.sub(r"\(S\d+/E\d+\)", "", title)
+            alt_title = re.sub(r"(Folge \d+(\:|\s\-))", "", alt_title)
+            alt_title = alt_title.replace("|", "").replace("  ", " ").replace(" .", "").strip()
+
+            info.update(
+                {
+                    'season_number': int(season_number),
+                    'episode_number': int(episode_number),
+                    'alt_title':  alt_title
+                }
+            )
+        elif episode_name_match:
+            episode_number = episode_name_match.group(1).replace("Folge ", "")
+
+            alt_title = re.sub(r"(Folge\s\d+)", "", title)
+            alt_title = alt_title.replace("|", "").replace("  ", " ").replace(" .", "").strip()
+
+            info.update(
+                {
+                    'season_number': int(0),
+                    'episode_number': int(episode_number),
+                    'alt_title':  alt_title
+                    }
+            )
+
+        elif episode_match:
+            episode_number = episode_match.group(1)
+            info.update(
+                {
+                    'season_number': int(0),
+                    'episode_number': int(episode_number),
+                    'alt_title':re.sub(r"\(\d+\)", "", title).replace("  ", "").strip()
+                }
+            )
+        else:
+            info.update(
+                {
+                    'alt_title': title
+                }
+            )
+
+        info.update(
+            {
                 'age_limit': age_limit,
                 'title': title,
                 'description': description,
-            'timestamp': unified_timestamp(player_page.get('broadcastedOn')),
-            'series': try_get(player_page, lambda x: x['show']['title']),
-        })
+                'timestamp': unified_timestamp(player_page['widgets'][0].get('broadcastedOn')),
+                'series': try_get(player_page['widgets'][0], lambda x: x['show']['title']),
+                'channel': player_page['widgets'][0]['publicationService']['name'],
+                'channel_id': player_page['widgets'][0]['publicationService']['id'],
+                'channel_url': f"https://www.ardmediathek.de/{player_page['widgets'][0]['publicationService']['id']}",
+            }
+        )
         return info
+
+    def _real_extract_serie(self, video_id):
+        entries = []
+
+        page_number = 0
+        page_size = 100
+
+        while True:
+
+            widgets = self._download_json(
+                    f'https://api.ardmediathek.de/page-gateway/widgets/ard/asset/{video_id}',
+                    video_id,
+                    query={'pageSize':str(page_size), 'pageNumber':page_number}
+            )
+
+            for teaser in widgets['teasers']:
+                if 'EPISODE' == teaser['coreAssetType']:
+                    item = self._real_extract_video(teaser['id'])
+                    item['webpage_url'] = f"https://www.ardmediathek.de/video/{teaser['id']}"
+
+                    entries.append(item)
+
+            total = widgets['pagination']['totalElements']
+            if (page_number + 1) * page_size > total:
+                break
+            page_number += 1
+
+        return self.playlist_result(entries)
+
+    def _real_extract_named_serie(self, video_id, sender):
+
+        entries = []
+        page_size = 100
+
+        widgets = self._download_json(
+                f'https://api.ardmediathek.de/page-gateway/pages/{sender}/editorial/{video_id}',
+                video_id,
+                query={'pageSize': str(10), 'pageNumber': 0}
+        )['widgets']
+
+        for widget in widgets:
+            widget_id = widget['id']
+            page_number = 0
+
+            while True:
+                page_data = self._download_json(
+                        f'https://api.ardmediathek.de/page-gateway/widgets/{sender}/editorials/{widget_id}',
+                        video_id,
+                        query={'pageSize': page_size, 'pageNumber': page_number}
+                )
+
+                for teaser in page_data['teasers']:
+                    if 'EPISODE' == teaser.get('coreAssetType', None) and teaser['type'] not in ['poster'] and ':' not in teaser['id']:
+
+                        item = self._real_extract_video(teaser['id'])
+                        item['webpage_url'] = f"https://www.ardmediathek.de/video/{teaser['id']}"
+                        entries.append(item)
+
+                total = page_data['pagination']['totalElements']
+                if (page_number + 1) * page_size > total:
+                    break
+                page_number += 1
+
+        return self.playlist_result(entries)

youtube_dl/extractor/common.py

@@ -454,10 +454,10 @@ class InfoExtractor(object):
         return cls.__match_valid_url(url) is not None
 
     @classmethod
-    def _match_id(cls, url):
+    def _match_id(cls, url, group_name = 'id'):
         m = cls.__match_valid_url(url)
         assert m
-        return compat_str(m.group('id'))
+        return compat_str(m.group(group_name))
 
     @classmethod
     def working(cls):
@@ -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,
 )
 
 
@@ -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/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(
+    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)
+                _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: