Merge ab5617be9e 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/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/sbs.py

@@ -2,32 +2,62 @@
 from __future__ import unicode_literals
 
 from .common import InfoExtractor
+from ..compat import (
+    compat_kwargs,
+    compat_str,
+)
 from ..utils import (
-    smuggle_url,
-    ExtractorError,
+    HEADRequest,
+    float_or_none,
+    int_or_none,
+    merge_dicts,
+    parse_duration,
+    parse_iso8601,
+    traverse_obj,
+    update_url_query,
+    url_or_none,
 )
 
 
 class SBSIE(InfoExtractor):
     IE_DESC = 'sbs.com.au'
-    _VALID_URL = r'https?://(?:www\.)?sbs\.com\.au/(?:ondemand(?:/video/(?:single/)?|.*?\bplay=|/watch/)|news/(?:embeds/)?video/)(?P<id>[0-9]+)'
+    _VALID_URL = r'''(?x)
+        https?://(?:www\.)?sbs\.com\.au/(?:
+            ondemand(?:
+                /video/(?:single/)?|
+                /(?:movie|tv-program)/[^/]+/|
+                /(?:tv|news)-series/(?:[^/]+/){3}|
+                .*?\bplay=|/watch/
+            )|news/(?:embeds/)?video/
+        )(?P<id>[0-9]+)'''
+    _EMBED_REGEX = [r'''(?x)]
+            (?:
+                <meta\s+property="og:video"\s+content=|
+                <iframe[^>]+?src=
+            )
+            (["\'])(?P<url>https?://(?:www\.)?sbs\.com\.au/ondemand/video/.+?)\1''']
 
     _TESTS = [{
-        # Original URL is handled by the generic IE which finds the iframe:
-        # http://www.sbs.com.au/thefeed/blog/2014/08/21/dingo-conservation
+        # Exceptional unrestricted show for testing, thanks SBS,
+        # from an iframe of this page, handled by the generic IE, now 404:
+        # http://www.sbs.com.au/thefeed/blog/2014/08/21/dingo-conservation, but replaced by
+        # https://www.sbs.com.au/programs/video/320403011771/Dingo-Conservation-The-Feed
         'url': 'http://www.sbs.com.au/ondemand/video/single/320403011771/?source=drupal&vertical=thefeed',
-        'md5': '3150cf278965eeabb5b4cea1c963fe0a',
+        'md5': 'e49d0290cb4f40d893b8dfe760dce6b0',
         'info_dict': {
-            'id': '_rFBPRPO4pMR',
+            'id': '320403011771',  # formerly '_rFBPRPO4pMR', no longer found
             'ext': 'mp4',
             'title': 'Dingo Conservation (The Feed)',
             'description': 'md5:f250a9856fca50d22dec0b5b8015f8a5',
-            'thumbnail': r're:http://.*\.jpg',
+            'thumbnail': r're:https?://.*\.jpg',
             'duration': 308,
             'timestamp': 1408613220,
             'upload_date': '20140821',
             'uploader': 'SBSC',
+            'tags': None,
+            'categories': None,
         },
+        'expected_warnings': ['Unable to download JSON metadata'],
     }, {
         'url': 'http://www.sbs.com.au/ondemand/video/320403011771/Dingo-Conservation-The-Feed',
         'only_matching': True,
@@ -46,33 +76,160 @@ class SBSIE(InfoExtractor):
     }, {
         'url': 'https://www.sbs.com.au/ondemand/watch/1698704451971',
         'only_matching': True,
+    }, {
+        'url': 'https://www.sbs.com.au/ondemand/movie/coherence/1469404227931',
+        'only_matching': True,
+    }, {
+        'note': 'Live stream',
+        'url': 'https://www.sbs.com.au/ondemand/video/1726824003663/sbs-24x7-live-stream-nsw',
+        'only_matching': True,
+    }, {
+        'url': 'https://www.sbs.com.au/ondemand/news-series/dateline/dateline-2022/dateline-s2022-ep26/2072245827515',
+        'only_matching': True,
+    }, {
+        'url': 'https://www.sbs.com.au/ondemand/tv-series/the-handmaids-tale/season-5/the-handmaids-tale-s5-ep1/2065631811776',
+        'only_matching': True,
+    }, {
+        'url': 'https://www.sbs.com.au/ondemand/tv-program/autun-romes-forgotten-sister/2116212803602',
+        'only_matching': True,
     }]
 
+    def _extract_m3u8_formats(self, m3u8_url, video_id, *args, **kwargs):
+        # ext, entry_protocol, preference, m3u8_id, note, errnote, fatal,
+        # live, data, headers, query
+        entry_protocol = args[1] if len(args) > 1 else kwargs.get('entry_protocol')
+        if not entry_protocol:
+            entry_protocol = 'm3u8_native'
+            if len(args) > 1:
+                args = list(args)
+                args[1] = entry_protocol
+            else:
+                kwargs['entry_protocol'] = entry_protocol
+                kwargs = compat_kwargs(kwargs)
+
+        return super(SBSIE, self)._extract_m3u8_formats(m3u8_url, video_id, *args, **kwargs)
+
+    _GEO_COUNTRIES = ['AU']
+    # naming for exportability
+    AUS_TV_PARENTAL_GUIDELINES = {
+        'P': 0,
+        'C': 7,
+        'G': 0,
+        'PG': 0,
+        'M': 14,
+        'MA15+': 15,
+        'AV15+': 15,
+        'MAV15+': 15,
+        'R18+': 18,
+        'NC': 0,  # not classified (unofficial, used by SBS)
+    }
+    _PLAYER_API = 'https://www.sbs.com.au/api/v3'
+    _CATALOGUE_API = 'https://catalogue.pr.sbsod.com/'
+    _VOD_BASE_URL = 'https://sbs-vod-prod-01.akamaized.net/'
+
+    def _call_api(self, video_id, path, query=None, data=None, headers=None, fatal=True):
+        return self._download_json(update_url_query(
+            self._CATALOGUE_API + path, query),
+            video_id, headers=headers or {}, fatal=fatal) or {}
+
+    def _get_smil_url(self, video_id):
+        return update_url_query(
+            self._PLAYER_API + 'video_smil', {'id': video_id})
+
+    def _get_player_data(self, video_id, headers=None, fatal=False):
+        return self._download_json(update_url_query(
+            self._PLAYER_API + 'video_stream', {'id': video_id, 'context': 'tv'}),
+            video_id, headers=headers or {}, fatal=fatal) or {}
+
     def _real_extract(self, url):
         video_id = self._match_id(url)
-        player_params = self._download_json(
-            'http://www.sbs.com.au/api/video_pdkvars/id/%s?form=json' % video_id, video_id)
+        # get media links directly though later metadata may contain contentUrl
+        smil_url = self._get_smil_url(video_id)
+        formats = self._extract_smil_formats(smil_url, video_id, fatal=False) or []
 
-        error = player_params.get('error')
-        if error:
-            error_message = 'Sorry, The video you are looking for does not exist.'
-            video_data = error.get('results') or {}
-            error_code = error.get('errorCode')
-            if error_code == 'ComingSoon':
-                error_message = '%s is not yet available.' % video_data.get('title', '')
-            elif error_code in ('Forbidden', 'intranetAccessOnly'):
-                error_message = 'Sorry, This video cannot be accessed via this website'
-            elif error_code == 'Expired':
-                error_message = 'Sorry, %s is no longer available.' % video_data.get('title', '')
-            raise ExtractorError('%s said: %s' % (self.IE_NAME, error_message), expected=True)
+        if not formats:
+            urlh = self._request_webpage(
+                HEADRequest(self._VOD_BASE_URL), video_id,
+                note='Checking geo-restriction', fatal=False, expected_status=403)
+            if urlh:
+                error_reasons = urlh.headers.get_all('x-error-reason') or []
+                if 'geo-blocked' in error_reasons:
+                    self.raise_geo_restricted(countries=self._GEO_COUNTRIES)
 
-        urls = player_params['releaseUrls']
-        theplatform_url = (urls.get('progressive') or urls.get('html')
-                           or urls.get('standard') or player_params['relatedItemsURL'])
+        self._sort_formats(formats)
 
-        return {
-            '_type': 'url_transparent',
-            'ie_key': 'ThePlatform',
+        # try for metadata from the same source
+        player_data = self._get_player_data(video_id, fatal=False)
+        media = traverse_obj(player_data, 'video_object', expected_type=dict) or {}
+        # get, or add, metadata from catalogue
+        media.update(self._call_api(video_id, 'mpx-media/' + video_id, fatal=not media))
+
+        # utils candidate for use with traverse_obj()
+        def txt_or_none(s):
+            return (s.strip() or None) if isinstance(s, compat_str) else None
+
+        # expected_type fn for thumbs
+        def xlate_thumb(t):
+            u = url_or_none(t.get('contentUrl'))
+            return u and {
+                'id': t.get('name'),
+                'url': u,
+                'width': int_or_none(t.get('width')),
+                'height': int_or_none(t.get('height')),
+            }
+
+        # may be numeric or timecoded
+        def really_parse_duration(d):
+            result = float_or_none(d)
+            if result is None:
+                result = parse_duration(d)
+            return result
+
+        def traverse_media(*args, **kwargs):
+            nkwargs = None
+            if 'expected_type' not in kwargs:
+                kwargs['expected_type'] = txt_or_none
+                nkwargs = kwargs
+            if 'get_all' not in kwargs:
+                kwargs['get_all'] = False
+                nkwargs = kwargs
+            if nkwargs:
+                kwargs = compat_kwargs(nkwargs)
+            return traverse_obj(media, *args, **kwargs)
+
+        # For named episodes, use the catalogue's title to set episode, rather than generic 'Episode N'.
+        if traverse_media('partOfSeries', expected_type=dict):
+            media['epName'] = traverse_media('title')
+
+        return merge_dicts(*reversed(({
             'id': video_id,
-            'url': smuggle_url(self._proto_relative_url(theplatform_url), {'force_smil_url': True}),
-        }
+        }, dict((k, traverse_media(v)) for k, v in {
+            'title': 'name',
+            'description': 'description',
+            'channel': ('taxonomy', 'channel', 'name'),
+            'series': ((('partOfSeries', 'name'), 'seriesTitle')),
+            'series_id': ((('partOfSeries', 'uuid'), 'seriesID')),
+            'episode': 'epName',
+        }.items()), {
+            'season_number': traverse_media((('partOfSeries', None), 'seasonNumber'), expected_type=int_or_none),
+            'episode_number': traverse_media('episodeNumber', expected_type=int_or_none),
+            'timestamp': traverse_media('datePublished', ('publication', 'startDate'),
+                                        expected_type=parse_iso8601),
+            'release_year': traverse_media('releaseYear', expected_type=int_or_none),
+            'duration': traverse_media('duration', expected_type=really_parse_duration),
+            'is_live': traverse_media('liveStream', expected_type=bool),
+            'age_limit': self.AUS_TV_PARENTAL_GUIDELINES.get(traverse_media(
+                'classificationID', 'contentRating', default='').upper()),
+            'categories': traverse_media(
+                ('genres', Ellipsis), ('taxonomy', ('genre', 'subgenre'), 'name'),
+                get_all=True) or None,
+            'tags': traverse_media(
+                (('consumerAdviceTexts', ('sbsSubCertification', 'consumerAdvice')), Ellipsis),
+                get_all=True) or None,
+            'thumbnails': traverse_media(('thumbnails', Ellipsis),
+                                         expected_type=xlate_thumb, get_all=True),
+            'formats': formats,
+            # TODO: _extract_smil_formats_and_subtitles()
+            # 'subtitles': subtitles,
+            'uploader': 'SBSC',
+        })))

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: