diff options
Diffstat (limited to 'youtube_dl')
| -rwxr-xr-x | youtube_dl/YoutubeDL.py | 37 | ||||
| -rw-r--r-- | youtube_dl/__init__.py | 6 | ||||
| -rw-r--r-- | youtube_dl/cache.py | 95 | ||||
| -rw-r--r-- | youtube_dl/casefold.py | 12 | ||||
| -rw-r--r-- | youtube_dl/compat.py | 277 | ||||
| -rw-r--r-- | youtube_dl/downloader/common.py | 28 | ||||
| -rw-r--r-- | youtube_dl/extractor/bokecc.py | 2 | ||||
| -rw-r--r-- | youtube_dl/extractor/cloudy.py | 2 | ||||
| -rw-r--r-- | youtube_dl/extractor/common.py | 92 | ||||
| -rw-r--r-- | youtube_dl/extractor/itv.py | 17 | ||||
| -rw-r--r-- | youtube_dl/extractor/senateisvp.py | 2 | ||||
| -rw-r--r-- | youtube_dl/extractor/youtube.py | 944 | ||||
| -rw-r--r-- | youtube_dl/jsinterp.py | 414 | ||||
| -rw-r--r-- | youtube_dl/options.py | 15 | ||||
| -rw-r--r-- | youtube_dl/utils.py | 8 | ||||
| -rw-r--r-- | youtube_dl/version.py | 2 |
16 files changed, 1399 insertions, 554 deletions
diff --git a/youtube_dl/YoutubeDL.py b/youtube_dl/YoutubeDL.py index 9e5620eef..8367b6e53 100755 --- a/youtube_dl/YoutubeDL.py +++ b/youtube_dl/YoutubeDL.py @@ -540,10 +540,14 @@ class YoutubeDL(object): """Print message to stdout if not in quiet mode.""" return self.to_stdout(message, skip_eol, check_quiet=True) - def _write_string(self, s, out=None): + def _write_string(self, s, out=None, only_once=False, _cache=set()): + if only_once and s in _cache: + return write_string(s, out=out, encoding=self.params.get('encoding')) + if only_once: + _cache.add(s) - def to_stdout(self, message, skip_eol=False, check_quiet=False): + def to_stdout(self, message, skip_eol=False, check_quiet=False, only_once=False): """Print message to stdout if not in quiet mode.""" if self.params.get('logger'): self.params['logger'].debug(message) @@ -552,9 +556,9 @@ class YoutubeDL(object): terminator = ['\n', ''][skip_eol] output = message + terminator - self._write_string(output, self._screen_file) + self._write_string(output, self._screen_file, only_once=only_once) - def to_stderr(self, message): + def to_stderr(self, message, only_once=False): """Print message to stderr.""" assert isinstance(message, compat_str) if self.params.get('logger'): @@ -562,7 +566,7 @@ class YoutubeDL(object): else: message = self._bidi_workaround(message) output = message + '\n' - self._write_string(output, self._err_file) + self._write_string(output, self._err_file, only_once=only_once) def to_console_title(self, message): if not self.params.get('consoletitle', False): @@ -641,18 +645,11 @@ class YoutubeDL(object): raise DownloadError(message, exc_info) self._download_retcode = 1 - def report_warning(self, message, only_once=False, _cache={}): + def report_warning(self, message, only_once=False): ''' Print the message to stderr, it will be prefixed with 'WARNING:' If stderr is a tty file the 'WARNING:' will be colored ''' - if only_once: - m_hash = hash((self, message)) - m_cnt = _cache.setdefault(m_hash, 0) - _cache[m_hash] = m_cnt + 1 - if m_cnt > 0: - return - if self.params.get('logger') is not None: self.params['logger'].warning(message) else: @@ -663,7 +660,7 @@ class YoutubeDL(object): else: _msg_header = 'WARNING:' warning_message = '%s %s' % (_msg_header, message) - self.to_stderr(warning_message) + self.to_stderr(warning_message, only_once=only_once) def report_error(self, message, *args, **kwargs): ''' @@ -677,6 +674,16 @@ class YoutubeDL(object): kwargs['message'] = '%s %s' % (_msg_header, message) self.trouble(*args, **kwargs) + def write_debug(self, message, only_once=False): + '''Log debug message or Print message to stderr''' + if not self.params.get('verbose', False): + return + message = '[debug] {0}'.format(message) + if self.params.get('logger'): + self.params['logger'].debug(message) + else: + self.to_stderr(message, only_once) + def report_unscoped_cookies(self, *args, **kwargs): # message=None, tb=False, is_error=False if len(args) <= 2: @@ -2514,7 +2521,7 @@ class YoutubeDL(object): self.get_encoding())) write_string(encoding_str, encoding=None) - writeln_debug = lambda *s: self._write_string('[debug] %s\n' % (''.join(s), )) + writeln_debug = lambda *s: self.write_debug(''.join(s)) writeln_debug('youtube-dl version ', __version__) if _LAZY_LOADER: writeln_debug('Lazy loading extractors enabled') diff --git a/youtube_dl/__init__.py b/youtube_dl/__init__.py index 06bdfb689..202f2c9b9 100644 --- a/youtube_dl/__init__.py +++ b/youtube_dl/__init__.py @@ -18,7 +18,7 @@ from .compat import ( compat_getpass, compat_register_utf8, compat_shlex_split, - workaround_optparse_bug9161, + _workaround_optparse_bug9161, ) from .utils import ( _UnsafeExtensionError, @@ -50,7 +50,7 @@ def _real_main(argv=None): # Compatibility fix for Windows compat_register_utf8() - workaround_optparse_bug9161() + _workaround_optparse_bug9161() setproctitle('youtube-dl') @@ -409,6 +409,8 @@ def _real_main(argv=None): 'include_ads': opts.include_ads, 'default_search': opts.default_search, 'youtube_include_dash_manifest': opts.youtube_include_dash_manifest, + 'youtube_player_js_version': opts.youtube_player_js_version, + 'youtube_player_js_variant': opts.youtube_player_js_variant, 'encoding': opts.encoding, 'extract_flat': opts.extract_flat, 'mark_watched': opts.mark_watched, diff --git a/youtube_dl/cache.py b/youtube_dl/cache.py index 54123da0e..eb0a729c2 100644 --- a/youtube_dl/cache.py +++ b/youtube_dl/cache.py @@ -1,3 +1,4 @@ +# coding: utf-8 from __future__ import unicode_literals import errno @@ -10,12 +11,14 @@ import traceback from .compat import ( compat_getenv, compat_open as open, + compat_os_makedirs, ) from .utils import ( error_to_compat_str, + escape_rfc3986, expand_path, is_outdated_version, - try_get, + traverse_obj, write_json_file, ) from .version import __version__ @@ -30,23 +33,35 @@ class Cache(object): def __init__(self, ydl): self._ydl = ydl + def _write_debug(self, *args, **kwargs): + self._ydl.write_debug(*args, **kwargs) + + def _report_warning(self, *args, **kwargs): + self._ydl.report_warning(*args, **kwargs) + + def _to_screen(self, *args, **kwargs): + self._ydl.to_screen(*args, **kwargs) + + def _get_param(self, k, default=None): + return self._ydl.params.get(k, default) + def _get_root_dir(self): - res = self._ydl.params.get('cachedir') + res = self._get_param('cachedir') if res is None: cache_root = compat_getenv('XDG_CACHE_HOME', '~/.cache') res = os.path.join(cache_root, self._YTDL_DIR) return expand_path(res) def _get_cache_fn(self, section, key, dtype): - assert re.match(r'^[a-zA-Z0-9_.-]+$', section), \ + assert re.match(r'^[\w.-]+$', section), \ 'invalid section %r' % section - assert re.match(r'^[a-zA-Z0-9_.-]+$', key), 'invalid key %r' % key + key = escape_rfc3986(key, safe='').replace('%', ',') # encode non-ascii characters return os.path.join( self._get_root_dir(), section, '%s.%s' % (key, dtype)) @property def enabled(self): - return self._ydl.params.get('cachedir') is not False + return self._get_param('cachedir') is not False def store(self, section, key, data, dtype='json'): assert dtype in ('json',) @@ -56,61 +71,75 @@ class Cache(object): fn = self._get_cache_fn(section, key, dtype) try: - try: - os.makedirs(os.path.dirname(fn)) - except OSError as ose: - if ose.errno != errno.EEXIST: - raise + compat_os_makedirs(os.path.dirname(fn), exist_ok=True) + self._write_debug('Saving {section}.{key} to cache'.format(section=section, key=key)) write_json_file({self._VERSION_KEY: __version__, 'data': data}, fn) except Exception: tb = traceback.format_exc() - self._ydl.report_warning( - 'Writing cache to %r failed: %s' % (fn, tb)) + self._report_warning('Writing cache to {fn!r} failed: {tb}'.format(fn=fn, tb=tb)) + + def clear(self, section, key, dtype='json'): + + if not self.enabled: + return + + fn = self._get_cache_fn(section, key, dtype) + self._write_debug('Clearing {section}.{key} from cache'.format(section=section, key=key)) + try: + os.remove(fn) + except Exception as e: + if getattr(e, 'errno') == errno.ENOENT: + # file not found + return + tb = traceback.format_exc() + self._report_warning('Clearing cache from {fn!r} failed: {tb}'.format(fn=fn, tb=tb)) def _validate(self, data, min_ver): - version = try_get(data, lambda x: x[self._VERSION_KEY]) + version = traverse_obj(data, self._VERSION_KEY) if not version: # Backward compatibility data, version = {'data': data}, self._DEFAULT_VERSION if not is_outdated_version(version, min_ver or '0', assume_new=False): return data['data'] - self._ydl.to_screen( - 'Discarding old cache from version {version} (needs {min_ver})'.format(**locals())) + self._write_debug('Discarding old cache from version {version} (needs {min_ver})'.format(version=version, min_ver=min_ver)) - def load(self, section, key, dtype='json', default=None, min_ver=None): + def load(self, section, key, dtype='json', default=None, **kw_min_ver): assert dtype in ('json',) + min_ver = kw_min_ver.get('min_ver') if not self.enabled: return default cache_fn = self._get_cache_fn(section, key, dtype) try: + with open(cache_fn, encoding='utf-8') as cachef: + self._write_debug('Loading {section}.{key} from cache'.format(section=section, key=key), only_once=True) + return self._validate(json.load(cachef), min_ver) + except (ValueError, KeyError): try: - with open(cache_fn, 'r', encoding='utf-8') as cachef: - return self._validate(json.load(cachef), min_ver) - except ValueError: - try: - file_size = os.path.getsize(cache_fn) - except (OSError, IOError) as oe: - file_size = error_to_compat_str(oe) - self._ydl.report_warning( - 'Cache retrieval from %s failed (%s)' % (cache_fn, file_size)) - except IOError: - pass # No cache available + file_size = 'size: %d' % os.path.getsize(cache_fn) + except (OSError, IOError) as oe: + file_size = error_to_compat_str(oe) + self._report_warning('Cache retrieval from %s failed (%s)' % (cache_fn, file_size)) + except Exception as e: + if getattr(e, 'errno') == errno.ENOENT: + # no cache available + return + self._report_warning('Cache retrieval from %s failed' % (cache_fn,)) return default def remove(self): if not self.enabled: - self._ydl.to_screen('Cache is disabled (Did you combine --no-cache-dir and --rm-cache-dir?)') + self._to_screen('Cache is disabled (Did you combine --no-cache-dir and --rm-cache-dir?)') return cachedir = self._get_root_dir() if not any((term in cachedir) for term in ('cache', 'tmp')): - raise Exception('Not removing directory %s - this does not look like a cache dir' % cachedir) + raise Exception('Not removing directory %s - this does not look like a cache dir' % (cachedir,)) - self._ydl.to_screen( - 'Removing cache dir %s .' % cachedir, skip_eol=True) + self._to_screen( + 'Removing cache dir %s .' % (cachedir,), skip_eol=True, ), if os.path.exists(cachedir): - self._ydl.to_screen('.', skip_eol=True) + self._to_screen('.', skip_eol=True) shutil.rmtree(cachedir) - self._ydl.to_screen('.') + self._to_screen('.') diff --git a/youtube_dl/casefold.py b/youtube_dl/casefold.py index ad9c66f8e..712b2e7fa 100644 --- a/youtube_dl/casefold.py +++ b/youtube_dl/casefold.py @@ -10,9 +10,10 @@ from .compat import ( # https://github.com/unicode-org/icu/blob/main/icu4c/source/data/unidata/CaseFolding.txt # In case newly foldable Unicode characters are defined, paste the new version # of the text inside the ''' marks. -# The text is expected to have only blank lines andlines with 1st character #, +# The text is expected to have only blank lines and lines with 1st character #, # all ignored, and fold definitions like this: -# `from_hex_code; space_separated_to_hex_code_list; comment` +# `from_hex_code; status; space_separated_to_hex_code_list; comment` +# Only `status` C/F are used. _map_str = ''' # CaseFolding-15.0.0.txt @@ -1657,11 +1658,6 @@ _map = dict( del _map_str -def casefold(s): +def _casefold(s): assert isinstance(s, compat_str) return ''.join((_map.get(c, c) for c in s)) - - -__all__ = [ - 'casefold', -] diff --git a/youtube_dl/compat.py b/youtube_dl/compat.py index ed1a33cf2..ebe22bdf9 100644 --- a/youtube_dl/compat.py +++ b/youtube_dl/compat.py @@ -16,7 +16,6 @@ import os import platform import re import shlex -import shutil import socket import struct import subprocess @@ -24,11 +23,15 @@ import sys import types import xml.etree.ElementTree +_IDENTITY = lambda x: x + # naming convention # 'compat_' + Python3_name.replace('.', '_') # other aliases exist for convenience and/or legacy +# wrap disposable test values in type() to reclaim storage -# deal with critical unicode/str things first +# deal with critical unicode/str things first: +# compat_str, compat_basestring, compat_chr try: # Python 2 compat_str, compat_basestring, compat_chr = ( @@ -39,18 +42,23 @@ except NameError: str, (str, bytes), chr ) -# casefold + +# compat_casefold try: compat_str.casefold compat_casefold = lambda s: s.casefold() except AttributeError: - from .casefold import casefold as compat_casefold + from .casefold import _casefold as compat_casefold + +# compat_collections_abc try: import collections.abc as compat_collections_abc except ImportError: import collections as compat_collections_abc + +# compat_urllib_request try: import urllib.request as compat_urllib_request except ImportError: # Python 2 @@ -79,11 +87,15 @@ except TypeError: _add_init_method_arg(compat_urllib_request.Request) del _add_init_method_arg + +# compat_urllib_error try: import urllib.error as compat_urllib_error except ImportError: # Python 2 import urllib2 as compat_urllib_error + +# compat_urllib_parse try: import urllib.parse as compat_urllib_parse except ImportError: # Python 2 @@ -98,17 +110,23 @@ except ImportError: # Python 2 compat_urlparse = compat_urllib_parse compat_urllib_parse_urlparse = compat_urllib_parse.urlparse + +# compat_urllib_response try: import urllib.response as compat_urllib_response except ImportError: # Python 2 import urllib as compat_urllib_response + +# compat_urllib_response.addinfourl try: compat_urllib_response.addinfourl.status except AttributeError: # .getcode() is deprecated in Py 3. compat_urllib_response.addinfourl.status = property(lambda self: self.getcode()) + +# compat_http_cookiejar try: import http.cookiejar as compat_cookiejar except ImportError: # Python 2 @@ -127,12 +145,16 @@ else: compat_cookiejar_Cookie = compat_cookiejar.Cookie compat_http_cookiejar_Cookie = compat_cookiejar_Cookie + +# compat_http_cookies try: import http.cookies as compat_cookies except ImportError: # Python 2 import Cookie as compat_cookies compat_http_cookies = compat_cookies + +# compat_http_cookies_SimpleCookie if sys.version_info[0] == 2 or sys.version_info < (3, 3): class compat_cookies_SimpleCookie(compat_cookies.SimpleCookie): def load(self, rawdata): @@ -155,11 +177,15 @@ else: compat_cookies_SimpleCookie = compat_cookies.SimpleCookie compat_http_cookies_SimpleCookie = compat_cookies_SimpleCookie + +# compat_html_entities, probably useless now try: import html.entities as compat_html_entities except ImportError: # Python 2 import htmlentitydefs as compat_html_entities + +# compat_html_entities_html5 try: # Python >= 3.3 compat_html_entities_html5 = compat_html_entities.html5 except AttributeError: @@ -2408,18 +2434,24 @@ except AttributeError: # Py < 3.1 compat_http_client.HTTPResponse.getcode = lambda self: self.status + +# compat_urllib_HTTPError try: from urllib.error import HTTPError as compat_HTTPError except ImportError: # Python 2 from urllib2 import HTTPError as compat_HTTPError compat_urllib_HTTPError = compat_HTTPError + +# compat_urllib_request_urlretrieve try: from urllib.request import urlretrieve as compat_urlretrieve except ImportError: # Python 2 from urllib import urlretrieve as compat_urlretrieve compat_urllib_request_urlretrieve = compat_urlretrieve + +# compat_html_parser_HTMLParser, compat_html_parser_HTMLParseError try: from HTMLParser import ( HTMLParser as compat_HTMLParser, @@ -2432,22 +2464,33 @@ except ImportError: # Python 3 # HTMLParseError was deprecated in Python 3.3 and removed in # Python 3.5. Introducing dummy exception for Python >3.5 for compatible # and uniform cross-version exception handling + class compat_HTMLParseError(Exception): pass + compat_html_parser_HTMLParser = compat_HTMLParser compat_html_parser_HTMLParseError = compat_HTMLParseError + +# compat_subprocess_get_DEVNULL try: _DEVNULL = subprocess.DEVNULL compat_subprocess_get_DEVNULL = lambda: _DEVNULL except AttributeError: compat_subprocess_get_DEVNULL = lambda: open(os.path.devnull, 'w') + +# compat_http_server try: import http.server as compat_http_server except ImportError: import BaseHTTPServer as compat_http_server + +# compat_urllib_parse_unquote_to_bytes, +# compat_urllib_parse_unquote, compat_urllib_parse_unquote_plus, +# compat_urllib_parse_urlencode, +# compat_urllib_parse_parse_qs try: from urllib.parse import unquote_to_bytes as compat_urllib_parse_unquote_to_bytes from urllib.parse import unquote as compat_urllib_parse_unquote @@ -2455,8 +2498,7 @@ try: from urllib.parse import urlencode as compat_urllib_parse_urlencode from urllib.parse import parse_qs as compat_parse_qs except ImportError: # Python 2 - _asciire = (compat_urllib_parse._asciire if hasattr(compat_urllib_parse, '_asciire') - else re.compile(r'([\x00-\x7f]+)')) + _asciire = getattr(compat_urllib_parse, '_asciire', None) or re.compile(r'([\x00-\x7f]+)') # HACK: The following are the correct unquote_to_bytes, unquote and unquote_plus # implementations from cpython 3.4.3's stdlib. Python 2's version @@ -2524,24 +2566,21 @@ except ImportError: # Python 2 # Possible solutions are to either port it from python 3 with all # the friends or manually ensure input query contains only byte strings. # We will stick with latter thus recursively encoding the whole query. - def compat_urllib_parse_urlencode(query, doseq=0, encoding='utf-8'): + def compat_urllib_parse_urlencode(query, doseq=0, safe='', encoding='utf-8', errors='strict'): + def encode_elem(e): if isinstance(e, dict): e = encode_dict(e) elif isinstance(e, (list, tuple,)): - list_e = encode_list(e) - e = tuple(list_e) if isinstance(e, tuple) else list_e + e = type(e)(encode_elem(el) for el in e) elif isinstance(e, compat_str): - e = e.encode(encoding) + e = e.encode(encoding, errors) return e def encode_dict(d): - return dict((encode_elem(k), encode_elem(v)) for k, v in d.items()) + return tuple((encode_elem(k), encode_elem(v)) for k, v in d.items()) - def encode_list(l): - return [encode_elem(e) for e in l] - - return compat_urllib_parse._urlencode(encode_elem(query), doseq=doseq) + return compat_urllib_parse._urlencode(encode_elem(query), doseq=doseq).decode('ascii') # HACK: The following is the correct parse_qs implementation from cpython 3's stdlib. # Python 2's version is apparently totally broken @@ -2596,8 +2635,61 @@ except ImportError: # Python 2 ('parse_qs', compat_parse_qs)): setattr(compat_urllib_parse, name, fix) + try: + all(chr(i) in b'' for i in range(256)) + except TypeError: + # not all chr(i) are str: patch Python2 quote + + _safemaps = getattr(compat_urllib_parse, '_safemaps', {}) + _always_safe = frozenset(compat_urllib_parse.always_safe) + + def _quote(s, safe='/'): + """quote('abc def') -> 'abc%20def'""" + + if not s and s is not None: # fast path + return s + safe = frozenset(safe) + cachekey = (safe, _always_safe) + try: + safe_map = _safemaps[cachekey] + except KeyError: + safe = _always_safe | safe + safe_map = {} + for i in range(256): + c = chr(i) + safe_map[c] = ( + c if (i < 128 and c in safe) + else b'%{0:02X}'.format(i)) + _safemaps[cachekey] = safe_map + + if safe.issuperset(s): + return s + return ''.join(safe_map[c] for c in s) + + # linked code + def _quote_plus(s, safe=''): + return ( + _quote(s, safe + b' ').replace(b' ', b'+') if b' ' in s + else _quote(s, safe)) + + # linked code + def _urlcleanup(): + if compat_urllib_parse._urlopener: + compat_urllib_parse._urlopener.cleanup() + _safemaps.clear() + compat_urllib_parse.ftpcache.clear() + + for name, fix in ( + ('quote', _quote), + ('quote_plus', _quote_plus), + ('urlcleanup', _urlcleanup)): + setattr(compat_urllib_parse, '_' + name, getattr(compat_urllib_parse, name)) + setattr(compat_urllib_parse, name, fix) + compat_urllib_parse_parse_qs = compat_parse_qs + +# compat_urllib_request_DataHandler try: from urllib.request import DataHandler as compat_urllib_request_DataHandler except ImportError: # Python < 3.4 @@ -2632,16 +2724,20 @@ except ImportError: # Python < 3.4 return compat_urllib_response.addinfourl(io.BytesIO(data), headers, url) + +# compat_xml_etree_ElementTree_ParseError try: from xml.etree.ElementTree import ParseError as compat_xml_parse_error except ImportError: # Python 2.6 from xml.parsers.expat import ExpatError as compat_xml_parse_error compat_xml_etree_ElementTree_ParseError = compat_xml_parse_error -etree = xml.etree.ElementTree +# compat_xml_etree_ElementTree_Element +_etree = xml.etree.ElementTree -class _TreeBuilder(etree.TreeBuilder): + +class _TreeBuilder(_etree.TreeBuilder): def doctype(self, name, pubid, system): pass @@ -2650,7 +2746,7 @@ try: # xml.etree.ElementTree.Element is a method in Python <=2.6 and # the following will crash with: # TypeError: isinstance() arg 2 must be a class, type, or tuple of classes and types - isinstance(None, etree.Element) + isinstance(None, _etree.Element) from xml.etree.ElementTree import Element as compat_etree_Element except TypeError: # Python <=2.6 from xml.etree.ElementTree import _ElementInterface as compat_etree_Element @@ -2658,12 +2754,12 @@ compat_xml_etree_ElementTree_Element = compat_etree_Element if sys.version_info[0] >= 3: def compat_etree_fromstring(text): - return etree.XML(text, parser=etree.XMLParser(target=_TreeBuilder())) + return _etree.XML(text, parser=_etree.XMLParser(target=_TreeBuilder())) else: # python 2.x tries to encode unicode strings with ascii (see the # XMLParser._fixtext method) try: - _etree_iter = etree.Element.iter + _etree_iter = _etree.Element.iter except AttributeError: # Python <=2.6 def _etree_iter(root): for el in root.findall('*'): @@ -2675,27 +2771,29 @@ else: # 2.7 source def _XML(text, parser=None): if not parser: - parser = etree.XMLParser(target=_TreeBuilder()) + parser = _etree.XMLParser(target=_TreeBuilder()) parser.feed(text) return parser.close() def _element_factory(*args, **kwargs): - el = etree.Element(*args, **kwargs) + el = _etree.Element(*args, **kwargs) for k, v in el.items(): if isinstance(v, bytes): el.set(k, v.decode('utf-8')) return el def compat_etree_fromstring(text): - doc = _XML(text, parser=etree.XMLParser(target=_TreeBuilder(element_factory=_element_factory))) + doc = _XML(text, parser=_etree.XMLParser(target=_TreeBuilder(element_factory=_element_factory))) for el in _etree_iter(doc): if el.text is not None and isinstance(el.text, bytes): el.text = el.text.decode('utf-8') return doc -if hasattr(etree, 'register_namespace'): - compat_etree_register_namespace = etree.register_namespace -else: + +# compat_xml_etree_register_namespace +try: + compat_etree_register_namespace = _etree.register_namespace +except AttributeError: def compat_etree_register_namespace(prefix, uri): """Register a namespace prefix. The registry is global, and any existing mapping for either the @@ -2704,14 +2802,16 @@ else: attributes in this namespace will be serialized with prefix if possible. ValueError is raised if prefix is reserved or is invalid. """ - if re.match(r"ns\d+$", prefix): - raise ValueError("Prefix format reserved for internal use") - for k, v in list(etree._namespace_map.items()): + if re.match(r'ns\d+$', prefix): + raise ValueError('Prefix format reserved for internal use') + for k, v in list(_etree._namespace_map.items()): if k == uri or v == prefix: - del etree._namespace_map[k] - etree._namespace_map[uri] = prefix + del _etree._namespace_map[k] + _etree._namespace_map[uri] = prefix compat_xml_etree_register_namespace = compat_etree_register_namespace + +# compat_xpath, compat_etree_iterfind if sys.version_info < (2, 7): # Here comes the crazy part: In 2.6, if the xpath is a unicode, # .//node does not match if a node is a direct child of . ! @@ -2898,7 +2998,6 @@ if sys.version_info < (2, 7): def __init__(self, root): self.root = root - ## # Generate all matching objects. def compat_etree_iterfind(elem, path, namespaces=None): @@ -2933,13 +3032,15 @@ if sys.version_info < (2, 7): else: - compat_xpath = lambda xpath: xpath compat_etree_iterfind = lambda element, match: element.iterfind(match) + compat_xpath = _IDENTITY +# compat_os_name compat_os_name = os._name if os.name == 'java' else os.name +# compat_shlex_quote if compat_os_name == 'nt': def compat_shlex_quote(s): return s if re.match(r'^[-_\w./]+$', s) else '"%s"' % s.replace('"', '\\"') @@ -2954,6 +3055,7 @@ else: return "'" + s.replace("'", "'\"'\"'") + "'" +# compat_shlex.split try: args = shlex.split('中文') assert (isinstance(args, list) @@ -2969,6 +3071,7 @@ except (AssertionError, UnicodeEncodeError): return list(map(lambda s: s.decode('utf-8'), shlex.split(s, comments, posix))) +# compat_ord def compat_ord(c): if isinstance(c, int): return c @@ -2976,6 +3079,7 @@ def compat_ord(c): return ord(c) +# compat_getenv, compat_os_path_expanduser, compat_setenv if sys.version_info >= (3, 0): compat_getenv = os.getenv compat_expanduser = os.path.expanduser @@ -3063,6 +3167,22 @@ else: compat_os_path_expanduser = compat_expanduser +# compat_os_makedirs +try: + os.makedirs('.', exist_ok=True) + compat_os_makedirs = os.makedirs +except TypeError: # < Py3.2 + from errno import EEXIST as _errno_EEXIST + + def compat_os_makedirs(name, mode=0o777, exist_ok=False): + try: + return os.makedirs(name, mode=mode) + except OSError as ose: + if not (exist_ok and ose.errno == _errno_EEXIST): + raise + + +# compat_os_path_realpath if compat_os_name == 'nt' and sys.version_info < (3, 8): # os.path.realpath on Windows does not follow symbolic links # prior to Python 3.8 (see https://bugs.python.org/issue9949) @@ -3076,6 +3196,7 @@ else: compat_os_path_realpath = compat_realpath +# compat_print if sys.version_info < (3, 0): def compat_print(s): from .utils import preferredencoding @@ -3086,6 +3207,7 @@ else: print(s) +# compat_getpass_getpass if sys.version_info < (3, 0) and sys.platform == 'win32': def compat_getpass(prompt, *args, **kwargs): if isinstance(prompt, compat_str): @@ -3098,36 +3220,42 @@ else: compat_getpass_getpass = compat_getpass +# compat_input try: compat_input = raw_input except NameError: # Python 3 compat_input = input +# compat_kwargs # Python < 2.6.5 require kwargs to be bytes try: - def _testfunc(x): - pass - _testfunc(**{'x': 0}) + (lambda x: x)(**{'x': 0}) except TypeError: def compat_kwargs(kwargs): return dict((bytes(k), v) for k, v in kwargs.items()) else: - compat_kwargs = lambda kwargs: kwargs + compat_kwargs = _IDENTITY +# compat_numeric_types try: compat_numeric_types = (int, float, long, complex) except NameError: # Python 3 compat_numeric_types = (int, float, complex) +# compat_integer_types try: compat_integer_types = (int, long) except NameError: # Python 3 compat_integer_types = (int, ) +# compat_int +compat_int = compat_integer_types[-1] + +# compat_socket_create_connection if sys.version_info < (2, 7): def compat_socket_create_connection(address, timeout, source_address=None): host, port = address @@ -3154,6 +3282,7 @@ else: compat_socket_create_connection = socket.create_connection +# compat_contextlib_suppress try: from contextlib import suppress as compat_contextlib_suppress except ImportError: @@ -3196,12 +3325,12 @@ except AttributeError: # repeated .close() is OK, but just in case with compat_contextlib_suppress(EnvironmentError): f.close() - popen.wait() + popen.wait() # Fix https://github.com/ytdl-org/youtube-dl/issues/4223 # See http://bugs.python.org/issue9161 for what is broken -def workaround_optparse_bug9161(): +def _workaround_optparse_bug9161(): op = optparse.OptionParser() og = optparse.OptionGroup(op, 'foo') try: @@ -3220,9 +3349,10 @@ def workaround_optparse_bug9161(): optparse.OptionGroup.add_option = _compat_add_option -if hasattr(shutil, 'get_terminal_size'): # Python >= 3.3 - compat_get_terminal_size = shutil.get_terminal_size -else: +# compat_shutil_get_terminal_size +try: + from shutil import get_terminal_size as compat_get_terminal_size # Python >= 3.3 +except ImportError: _terminal_size = collections.namedtuple('terminal_size', ['columns', 'lines']) def compat_get_terminal_size(fallback=(80, 24)): @@ -3252,27 +3382,33 @@ else: columns = _columns if lines is None or lines <= 0: lines = _lines + return _terminal_size(columns, lines) +compat_shutil_get_terminal_size = compat_get_terminal_size + +# compat_itertools_count try: - itertools.count(start=0, step=1) + type(itertools.count(start=0, step=1)) compat_itertools_count = itertools.count -except TypeError: # Python 2.6 +except TypeError: # Python 2.6 lacks step def compat_itertools_count(start=0, step=1): while True: yield start start += step +# compat_tokenize_tokenize if sys.version_info >= (3, 0): from tokenize import tokenize as compat_tokenize_tokenize else: from tokenize import generate_tokens as compat_tokenize_tokenize +# compat_struct_pack, compat_struct_unpack, compat_Struct try: - struct.pack('!I', 0) + type(struct.pack('!I', 0)) except TypeError: # In Python 2.6 and 2.7.x < 2.7.7, struct requires a bytes argument # See https://bugs.python.org/issue19099 @@ -3304,8 +3440,10 @@ else: compat_Struct = struct.Struct -# compat_map/filter() returning an iterator, supposedly the -# same versioning as for zip below +# builtins returning an iterator + +# compat_map, compat_filter +# supposedly the same versioning as for zip below try: from future_builtins import map as compat_map except ImportError: @@ -3322,6 +3460,7 @@ except ImportError: except ImportError: compat_filter = filter +# compat_zip try: from future_builtins import zip as compat_zip except ImportError: # not 2.6+ or is 3.x @@ -3331,6 +3470,7 @@ except ImportError: # not 2.6+ or is 3.x compat_zip = zip +# compat_itertools_zip_longest # method renamed between Py2/3 try: from itertools import zip_longest as compat_itertools_zip_longest @@ -3338,7 +3478,8 @@ except ImportError: from itertools import izip_longest as compat_itertools_zip_longest -# new class in collections +# compat_collections_chain_map +# collections.ChainMap: new class try: from collections import ChainMap as compat_collections_chain_map # Py3.3's ChainMap is deficient @@ -3394,19 +3535,22 @@ except ImportError: def new_child(self, m=None, **kwargs): m = m or {} m.update(kwargs) - return compat_collections_chain_map(m, *self.maps) + # support inheritance ! + return type(self)(m, *self.maps) @property def parents(self): - return compat_collections_chain_map(*(self.maps[1:])) + return type(self)(*(self.maps[1:])) +# compat_re_Pattern, compat_re_Match # Pythons disagree on the type of a pattern (RegexObject, _sre.SRE_Pattern, Pattern, ...?) compat_re_Pattern = type(re.compile('')) # and on the type of a match compat_re_Match = type(re.match('a', 'a')) +# compat_base64_b64decode if sys.version_info < (3, 3): def compat_b64decode(s, *args, **kwargs): if isinstance(s, compat_str): @@ -3418,6 +3562,7 @@ else: compat_base64_b64decode = compat_b64decode +# compat_ctypes_WINFUNCTYPE if platform.python_implementation() == 'PyPy' and sys.pypy_version_info < (5, 4, 0): # PyPy2 prior to version 5.4.0 expects byte strings as Windows function # names, see the original PyPy issue [1] and the youtube-dl one [2]. @@ -3436,6 +3581,7 @@ else: return ctypes.WINFUNCTYPE(*args, **kwargs) +# compat_open if sys.version_info < (3, 0): # open(file, mode='r', buffering=- 1, encoding=None, errors=None, newline=None, closefd=True) not: opener=None def compat_open(file_, *args, **kwargs): @@ -3463,18 +3609,28 @@ except AttributeError: def compat_datetime_timedelta_total_seconds(td): return (td.microseconds + (td.seconds + td.days * 24 * 3600) * 10**6) / 10**6 + # optional decompression packages +# compat_brotli # PyPi brotli package implements 'br' Content-Encoding try: import brotli as compat_brotli except ImportError: compat_brotli = None +# compat_ncompress # PyPi ncompress package implements 'compress' Content-Encoding try: import ncompress as compat_ncompress except ImportError: compat_ncompress = None +# compat_zstandard +# PyPi zstandard package implements 'zstd' Content-Encoding (RFC 8878 7.2) +try: + import zstandard as compat_zstandard +except ImportError: + compat_zstandard = None + legacy = [ 'compat_HTMLParseError', @@ -3491,6 +3647,7 @@ legacy = [ 'compat_getpass', 'compat_parse_qs', 'compat_realpath', + 'compat_shlex_split', 'compat_urllib_parse_parse_qs', 'compat_urllib_parse_unquote', 'compat_urllib_parse_unquote_plus', @@ -3504,8 +3661,6 @@ legacy = [ __all__ = [ - 'compat_html_parser_HTMLParseError', - 'compat_html_parser_HTMLParser', 'compat_Struct', 'compat_base64_b64decode', 'compat_basestring', @@ -3514,13 +3669,9 @@ __all__ = [ 'compat_chr', 'compat_collections_abc', 'compat_collections_chain_map', - 'compat_datetime_timedelta_total_seconds', - 'compat_http_cookiejar', - 'compat_http_cookiejar_Cookie', - 'compat_http_cookies', - 'compat_http_cookies_SimpleCookie', 'compat_contextlib_suppress', 'compat_ctypes_WINFUNCTYPE', + 'compat_datetime_timedelta_total_seconds', 'compat_etree_fromstring', 'compat_etree_iterfind', 'compat_filter', @@ -3529,9 +3680,16 @@ __all__ = [ 'compat_getpass_getpass', 'compat_html_entities', 'compat_html_entities_html5', + 'compat_html_parser_HTMLParseError', + 'compat_html_parser_HTMLParser', + 'compat_http_cookiejar', + 'compat_http_cookiejar_Cookie', + 'compat_http_cookies', + 'compat_http_cookies_SimpleCookie', 'compat_http_client', 'compat_http_server', 'compat_input', + 'compat_int', 'compat_integer_types', 'compat_itertools_count', 'compat_itertools_zip_longest', @@ -3541,6 +3699,7 @@ __all__ = [ 'compat_numeric_types', 'compat_open', 'compat_ord', + 'compat_os_makedirs', 'compat_os_name', 'compat_os_path_expanduser', 'compat_os_path_realpath', @@ -3550,7 +3709,7 @@ __all__ = [ 'compat_register_utf8', 'compat_setenv', 'compat_shlex_quote', - 'compat_shlex_split', + 'compat_shutil_get_terminal_size', 'compat_socket_create_connection', 'compat_str', 'compat_struct_pack', @@ -3570,5 +3729,5 @@ __all__ = [ 'compat_xml_etree_register_namespace', 'compat_xpath', 'compat_zip', - 'workaround_optparse_bug9161', + 'compat_zstandard', ] diff --git a/youtube_dl/downloader/common.py b/youtube_dl/downloader/common.py index 91e691776..8354030a9 100644 --- a/youtube_dl/downloader/common.py +++ b/youtube_dl/downloader/common.py @@ -11,6 +11,7 @@ from ..utils import ( decodeArgument, encodeFilename, error_to_compat_str, + float_or_none, format_bytes, shell_quote, timeconvert, @@ -367,14 +368,27 @@ class FileDownloader(object): }) return True - min_sleep_interval = self.params.get('sleep_interval') - if min_sleep_interval: - max_sleep_interval = self.params.get('max_sleep_interval', min_sleep_interval) - sleep_interval = random.uniform(min_sleep_interval, max_sleep_interval) + min_sleep_interval, max_sleep_interval = ( + float_or_none(self.params.get(interval), default=0) + for interval in ('sleep_interval', 'max_sleep_interval')) + + sleep_note = '' + available_at = info_dict.get('available_at') + if available_at: + forced_sleep_interval = available_at - int(time.time()) + if forced_sleep_interval > min_sleep_interval: + sleep_note = 'as required by the site' + min_sleep_interval = forced_sleep_interval + if forced_sleep_interval > max_sleep_interval: + max_sleep_interval = forced_sleep_interval + + sleep_interval = random.uniform( + min_sleep_interval, max_sleep_interval or min_sleep_interval) + + if sleep_interval > 0: self.to_screen( - '[download] Sleeping %s seconds...' % ( - int(sleep_interval) if sleep_interval.is_integer() - else '%.2f' % sleep_interval)) + '[download] Sleeping %.2f seconds %s...' % ( + sleep_interval, sleep_note)) time.sleep(sleep_interval) return self.real_download(filename, info_dict) diff --git a/youtube_dl/extractor/bokecc.py b/youtube_dl/extractor/bokecc.py index 6017e8344..4b8bef391 100644 --- a/youtube_dl/extractor/bokecc.py +++ b/youtube_dl/extractor/bokecc.py @@ -32,7 +32,7 @@ class BokeCCBaseIE(InfoExtractor): class BokeCCIE(BokeCCBaseIE): - _IE_DESC = 'CC视频' + IE_DESC = 'CC视频' _VALID_URL = r'https?://union\.bokecc\.com/playvideo\.bo\?(?P<query>.*)' _TESTS = [{ diff --git a/youtube_dl/extractor/cloudy.py b/youtube_dl/extractor/cloudy.py index 85ca20ecc..d39a9a5c2 100644 --- a/youtube_dl/extractor/cloudy.py +++ b/youtube_dl/extractor/cloudy.py @@ -9,7 +9,7 @@ from ..utils import ( class CloudyIE(InfoExtractor): - _IE_DESC = 'cloudy.ec' + IE_DESC = 'cloudy.ec' _VALID_URL = r'https?://(?:www\.)?cloudy\.ec/(?:v/|embed\.php\?.*?\bid=)(?P<id>[A-Za-z0-9]+)' _TESTS = [{ 'url': 'https://www.cloudy.ec/v/af511e2527aac', diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py index 78704b557..a64fcfccc 100644 --- a/youtube_dl/extractor/common.py +++ b/youtube_dl/extractor/common.py @@ -422,6 +422,8 @@ class InfoExtractor(object): _GEO_COUNTRIES = None _GEO_IP_BLOCKS = None _WORKING = True + # supply this in public subclasses: used in supported sites list, etc + # IE_DESC = 'short description of IE' def __init__(self, downloader=None): """Constructor. Receives an optional downloader.""" @@ -503,7 +505,7 @@ class InfoExtractor(object): if not self._x_forwarded_for_ip: # Geo bypass mechanism is explicitly disabled by user - if not self._downloader.params.get('geo_bypass', True): + if not self.get_param('geo_bypass', True): return if not geo_bypass_context: @@ -525,7 +527,7 @@ class InfoExtractor(object): # Explicit IP block specified by user, use it right away # regardless of whether extractor is geo bypassable or not - ip_block = self._downloader.params.get('geo_bypass_ip_block', None) + ip_block = self.get_param('geo_bypass_ip_block', None) # Otherwise use random IP block from geo bypass context but only # if extractor is known as geo bypassable @@ -536,8 +538,8 @@ class InfoExtractor(object): if ip_block: self._x_forwarded_for_ip = GeoUtils.random_ipv4(ip_block) - if self._downloader.params.get('verbose', False): - self._downloader.to_screen( + if self.get_param('verbose', False): + self.to_screen( '[debug] Using fake IP %s as X-Forwarded-For.' % self._x_forwarded_for_ip) return @@ -546,7 +548,7 @@ class InfoExtractor(object): # Explicit country code specified by user, use it right away # regardless of whether extractor is geo bypassable or not - country = self._downloader.params.get('geo_bypass_country', None) + country = self.get_param('geo_bypass_country', None) # Otherwise use random country code from geo bypass context but # only if extractor is known as geo bypassable @@ -557,8 +559,8 @@ class InfoExtractor(object): if country: self._x_forwarded_for_ip = GeoUtils.random_ipv4(country) - if self._downloader.params.get('verbose', False): - self._downloader.to_screen( + if self.get_param('verbose', False): + self.to_screen( '[debug] Using fake IP %s (%s) as X-Forwarded-For.' % (self._x_forwarded_for_ip, country.upper())) @@ -584,9 +586,9 @@ class InfoExtractor(object): raise ExtractorError('An extractor error has occurred.', cause=e) def __maybe_fake_ip_and_retry(self, countries): - if (not self._downloader.params.get('geo_bypass_country', None) + if (not self.get_param('geo_bypass_country', None) and self._GEO_BYPASS - and self._downloader.params.get('geo_bypass', True) + and self.get_param('geo_bypass', True) and not self._x_forwarded_for_ip and countries): country_code = random.choice(countries) @@ -696,7 +698,7 @@ class InfoExtractor(object): if fatal: raise ExtractorError(errmsg, sys.exc_info()[2], cause=err) else: - self._downloader.report_warning(errmsg) + self.report_warning(errmsg) return False def _download_webpage_handle(self, url_or_request, video_id, note=None, errnote=None, fatal=True, encoding=None, data=None, headers={}, query={}, expected_status=None): @@ -768,11 +770,11 @@ class InfoExtractor(object): webpage_bytes = prefix + webpage_bytes if not encoding: encoding = self._guess_encoding_from_content(content_type, webpage_bytes) - if self._downloader.params.get('dump_intermediate_pages', False): + if self.get_param('dump_intermediate_pages', False): self.to_screen('Dumping request to ' + urlh.geturl()) dump = base64.b64encode(webpage_bytes).decode('ascii') - self._downloader.to_screen(dump) - if self._downloader.params.get('write_pages', False): + self.to_screen(dump) + if self.get_param('write_pages', False): basen = '%s_%s' % (video_id, urlh.geturl()) if len(basen) > 240: h = '___' + hashlib.md5(basen.encode('utf-8')).hexdigest() @@ -974,19 +976,9 @@ class InfoExtractor(object): """Print msg to screen, prefixing it with '[ie_name]'""" self._downloader.to_screen(self.__ie_msg(msg)) - def write_debug(self, msg, only_once=False, _cache=[]): + def write_debug(self, msg, only_once=False): '''Log debug message or Print message to stderr''' - if not self.get_param('verbose', False): - return - message = '[debug] ' + self.__ie_msg(msg) - logger = self.get_param('logger') - if logger: - logger.debug(message) - else: - if only_once and hash(message) in _cache: - return - self._downloader.to_stderr(message) - _cache.append(hash(message)) + self._downloader.write_debug(self.__ie_msg(msg), only_once=only_once) # name, default=None, *args, **kwargs def get_param(self, name, *args, **kwargs): @@ -1082,7 +1074,7 @@ class InfoExtractor(object): if mobj: break - if not self._downloader.params.get('no_color') and compat_os_name != 'nt' and sys.stderr.isatty(): + if not self.get_param('no_color') and compat_os_name != 'nt' and sys.stderr.isatty(): _name = '\033[0;34m%s\033[0m' % name else: _name = name @@ -1100,7 +1092,7 @@ class InfoExtractor(object): elif fatal: raise RegexNotFoundError('Unable to extract %s' % _name) else: - self._downloader.report_warning('unable to extract %s' % _name + bug_reports_message()) + self.report_warning('unable to extract %s' % _name + bug_reports_message()) return None def _search_json(self, start_pattern, string, name, video_id, **kwargs): @@ -1170,7 +1162,7 @@ class InfoExtractor(object): username = None password = None - if self._downloader.params.get('usenetrc', False): + if self.get_param('usenetrc', False): try: netrc_machine = netrc_machine or self._NETRC_MACHINE info = netrc.netrc().authenticators(netrc_machine) @@ -1181,7 +1173,7 @@ class InfoExtractor(object): raise netrc.NetrcParseError( 'No authenticators for %s' % netrc_machine) except (AttributeError, IOError, netrc.NetrcParseError) as err: - self._downloader.report_warning( + self.report_warning( 'parsing .netrc: %s' % error_to_compat_str(err)) return username, password @@ -1218,10 +1210,10 @@ class InfoExtractor(object): """ if self._downloader is None: return None - downloader_params = self._downloader.params - if downloader_params.get('twofactor') is not None: - return downloader_params['twofactor'] + twofactor = self.get_param('twofactor') + if twofactor is not None: + return twofactor return compat_getpass('Type %s and press [Return]: ' % note) @@ -1356,7 +1348,7 @@ class InfoExtractor(object): elif fatal: raise RegexNotFoundError('Unable to extract JSON-LD') else: - self._downloader.report_warning('unable to extract JSON-LD %s' % bug_reports_message()) + self.report_warning('unable to extract JSON-LD %s' % bug_reports_message()) return {} def _json_ld(self, json_ld, video_id, fatal=True, expected_type=None): @@ -1587,7 +1579,7 @@ class InfoExtractor(object): if f.get('vcodec') == 'none': # audio only preference -= 50 - if self._downloader.params.get('prefer_free_formats'): + if self.get_param('prefer_free_formats'): ORDER = ['aac', 'mp3', 'm4a', 'webm', 'ogg', 'opus'] else: ORDER = ['webm', 'opus', 'ogg', 'mp3', 'aac', 'm4a'] @@ -1599,7 +1591,7 @@ class InfoExtractor(object): else: if f.get('acodec') == 'none': # video only preference -= 40 - if self._downloader.params.get('prefer_free_formats'): + if self.get_param('prefer_free_formats'): ORDER = ['flv', 'mp4', 'webm'] else: ORDER = ['webm', 'flv', 'mp4'] @@ -1665,7 +1657,7 @@ class InfoExtractor(object): """ Either "http:" or "https:", depending on the user's preferences """ return ( 'http:' - if self._downloader.params.get('prefer_insecure', False) + if self.get_param('prefer_insecure', False) else 'https:') def _proto_relative_url(self, url, scheme=None): @@ -3197,7 +3189,7 @@ class InfoExtractor(object): if fatal: raise ExtractorError(msg) else: - self._downloader.report_warning(msg) + self.report_warning(msg) return res def _float(self, v, name, fatal=False, **kwargs): @@ -3207,7 +3199,7 @@ class InfoExtractor(object): if fatal: raise ExtractorError(msg) else: - self._downloader.report_warning(msg) + self.report_warning(msg) return res def _set_cookie(self, domain, name, value, expire_time=None, port=None, @@ -3216,12 +3208,12 @@ class InfoExtractor(object): 0, name, value, port, port is not None, domain, True, domain.startswith('.'), path, True, secure, expire_time, discard, None, None, rest) - self._downloader.cookiejar.set_cookie(cookie) + self.cookiejar.set_cookie(cookie) def _get_cookies(self, url): """ Return a compat_cookies_SimpleCookie with the cookies for the url """ req = sanitized_Request(url) - self._downloader.cookiejar.add_cookie_header(req) + self.cookiejar.add_cookie_header(req) return compat_cookies_SimpleCookie(req.get_header('Cookie')) def _apply_first_set_cookie_header(self, url_handle, cookie): @@ -3281,8 +3273,8 @@ class InfoExtractor(object): return not any_restricted def extract_subtitles(self, *args, **kwargs): - if (self._downloader.params.get('writesubtitles', False) - or self._downloader.params.get('listsubtitles')): + if (self.get_param('writesubtitles', False) + or self.get_param('listsubtitles')): return self._get_subtitles(*args, **kwargs) return {} @@ -3303,7 +3295,11 @@ class InfoExtractor(object): """ Merge subtitle dictionaries, language by language. """ # ..., * , target=None - target = kwargs.get('target') or dict(subtitle_dict1) + target = kwargs.get('target') + if target is None: + target = dict(subtitle_dict1) + else: + subtitle_dicts = (subtitle_dict1,) + subtitle_dicts for subtitle_dict in subtitle_dicts: for lang in subtitle_dict: @@ -3311,8 +3307,8 @@ class InfoExtractor(object): return target def extract_automatic_captions(self, *args, **kwargs): - if (self._downloader.params.get('writeautomaticsub', False) - or self._downloader.params.get('listsubtitles')): + if (self.get_param('writeautomaticsub', False) + or self.get_param('listsubtitles')): return self._get_automatic_captions(*args, **kwargs) return {} @@ -3320,9 +3316,9 @@ class InfoExtractor(object): raise NotImplementedError('This method must be implemented by subclasses') def mark_watched(self, *args, **kwargs): - if (self._downloader.params.get('mark_watched', False) + if (self.get_param('mark_watched', False) and (self._get_login_info()[0] is not None - or self._downloader.params.get('cookiefile') is not None)): + or self.get_param('cookiefile') is not None)): self._mark_watched(*args, **kwargs) def _mark_watched(self, *args, **kwargs): @@ -3330,7 +3326,7 @@ class InfoExtractor(object): def geo_verification_headers(self): headers = {} - geo_verification_proxy = self._downloader.params.get('geo_verification_proxy') + geo_verification_proxy = self.get_param('geo_verification_proxy') if geo_verification_proxy: headers['Ytdl-request-proxy'] = geo_verification_proxy return headers diff --git a/youtube_dl/extractor/itv.py b/youtube_dl/extractor/itv.py index c64af3be6..2510ad887 100644 --- a/youtube_dl/extractor/itv.py +++ b/youtube_dl/extractor/itv.py @@ -35,15 +35,6 @@ from ..utils import ( class ITVBaseIE(InfoExtractor): - def _search_nextjs_data(self, webpage, video_id, **kw): - transform_source = kw.pop('transform_source', None) - fatal = kw.pop('fatal', True) - return self._parse_json( - self._search_regex( - r'''<script\b[^>]+\bid=('|")__NEXT_DATA__\1[^>]*>(?P<js>[^<]+)</script>''', - webpage, 'next.js data', group='js', fatal=fatal, **kw), - video_id, transform_source=transform_source, fatal=fatal) - def __handle_request_webpage_error(self, err, video_id=None, errnote=None, fatal=True): if errnote is False: return False @@ -109,7 +100,9 @@ class ITVBaseIE(InfoExtractor): class ITVIE(ITVBaseIE): _VALID_URL = r'https?://(?:www\.)?itv\.com/(?:(?P<w>watch)|hub)/[^/]+/(?(w)[\w-]+/)(?P<id>\w+)' - _IE_DESC = 'ITVX' + IE_DESC = 'ITVX' + _WORKING = False + _TESTS = [{ 'note': 'Hub URLs redirect to ITVX', 'url': 'https://www.itv.com/hub/liar/2a4547a0012', @@ -270,7 +263,7 @@ class ITVIE(ITVBaseIE): 'ext': determine_ext(href, 'vtt'), }) - next_data = self._search_nextjs_data(webpage, video_id, fatal=False, default='{}') + next_data = self._search_nextjs_data(webpage, video_id, fatal=False, default={}) video_data.update(traverse_obj(next_data, ('props', 'pageProps', ('title', 'episode')), expected_type=dict)[0] or {}) title = traverse_obj(video_data, 'headerTitle', 'episodeTitle') info = self._og_extract(webpage, require_title=not title) @@ -323,7 +316,7 @@ class ITVIE(ITVBaseIE): class ITVBTCCIE(ITVBaseIE): _VALID_URL = r'https?://(?:www\.)?itv\.com/(?!(?:watch|hub)/)(?:[^/]+/)+(?P<id>[^/?#&]+)' - _IE_DESC = 'ITV articles: News, British Touring Car Championship' + IE_DESC = 'ITV articles: News, British Touring Car Championship' _TESTS = [{ 'note': 'British Touring Car Championship', 'url': 'https://www.itv.com/btcc/articles/btcc-2018-all-the-action-from-brands-hatch', diff --git a/youtube_dl/extractor/senateisvp.py b/youtube_dl/extractor/senateisvp.py index db5ef8b57..b8ac58713 100644 --- a/youtube_dl/extractor/senateisvp.py +++ b/youtube_dl/extractor/senateisvp.py @@ -47,7 +47,7 @@ class SenateISVPIE(InfoExtractor): ['vetaff', '76462', 'http://vetaff-f.akamaihd.net'], ['arch', '', 'http://ussenate-f.akamaihd.net/'] ] - _IE_NAME = 'senate.gov' + IE_NAME = 'senate.gov' _VALID_URL = r'https?://(?:www\.)?senate\.gov/isvp/?\?(?P<qs>.+)' _TESTS = [{ 'url': 'http://www.senate.gov/isvp/?comm=judiciary&type=live&stt=&filename=judiciary031715&auto_play=false&wmode=transparent&poster=http%3A%2F%2Fwww.judiciary.senate.gov%2Fthemes%2Fjudiciary%2Fimages%2Fvideo-poster-flash-fit.png', diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 56957a661..c045bc8bc 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -1,5 +1,4 @@ # coding: utf-8 - from __future__ import unicode_literals import collections @@ -27,11 +26,14 @@ from ..compat import ( ) from ..jsinterp import JSInterpreter from ..utils import ( + bug_reports_message, clean_html, dict_get, error_to_compat_str, ExtractorError, + filter_dict, float_or_none, + get_first, extract_attributes, get_element_by_attribute, int_or_none, @@ -46,6 +48,7 @@ from ..utils import ( parse_duration, parse_qs, qualities, + remove_end, remove_start, smuggle_url, str_or_none, @@ -63,6 +66,7 @@ from ..utils import ( url_or_none, urlencode_postdata, urljoin, + variadic, ) @@ -82,9 +86,79 @@ class YoutubeBaseInfoExtractor(InfoExtractor): _PLAYLIST_ID_RE = r'(?:(?:PL|LL|EC|UU|FL|RD|UL|TL|PU|OLAK5uy_)[0-9A-Za-z-_]{10,}|RDMM)' + _INNERTUBE_CLIENTS = { + 'ios': { + 'INNERTUBE_CONTEXT': { + 'client': { + 'clientName': 'IOS', + 'clientVersion': '20.10.4', + 'deviceMake': 'Apple', + 'deviceModel': 'iPhone16,2', + 'userAgent': 'com.google.ios.youtube/20.10.4 (iPhone16,2; U; CPU iOS 18_3_2 like Mac OS X;)', + 'osName': 'iPhone', + 'osVersion': '18.3.2.22D82', + }, + }, + 'INNERTUBE_CONTEXT_CLIENT_NAME': 5, + 'REQUIRE_PO_TOKEN': False, + 'REQUIRE_JS_PLAYER': False, + }, + # mweb has 'ultralow' formats + # See: https://github.com/yt-dlp/yt-dlp/pull/557 + 'mweb': { + 'INNERTUBE_CONTEXT': { + 'client': { + 'clientName': 'MWEB', + 'clientVersion': '2.2.20250925.01.00', + # mweb previously did not require PO Token with this UA + 'userAgent': 'Mozilla/5.0 (iPad; CPU OS 16_7_10 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.6 Mobile/15E148 Safari/604.1,gzip(gfe)', + }, + }, + 'INNERTUBE_CONTEXT_CLIENT_NAME': 2, + 'REQUIRE_PO_TOKEN': True, + 'SUPPORTS_COOKIES': True, + }, + 'tv': { + 'INNERTUBE_CONTEXT': { + 'client': { + 'clientName': 'TVHTML5', + 'clientVersion': '7.20250312.16.00', + # See: https://github.com/youtube/cobalt/blob/main/cobalt/browser/user_agent/user_agent_platform_info.cc#L506 + 'userAgent': 'Mozilla/5.0 (ChromiumStylePlatform) Cobalt/25.lts.30.1034943-gold (unlike Gecko), Unknown_TV_Unknown_0/Unknown (Unknown, Unknown)', + }, + }, + 'INNERTUBE_CONTEXT_CLIENT_NAME': 7, + 'SUPPORTS_COOKIES': True, + }, + + 'web': { + 'INNERTUBE_CONTEXT': { + 'client': { + 'clientName': 'WEB', + 'clientVersion': '2.20250925.01.00', + 'userAgent': 'Mozilla/5.0', + }, + }, + 'INNERTUBE_CONTEXT_CLIENT_NAME': 1, + 'REQUIRE_PO_TOKEN': True, + 'SUPPORTS_COOKIES': True, + }, + # Safari UA returns pre-merged video+audio 144p/240p/360p/720p/1080p HLS formats + 'web_safari': { + 'INNERTUBE_CONTEXT': { + 'client': { + 'clientName': 'WEB', + 'clientVersion': '2.20250925.01.00', + 'userAgent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/15.5 Safari/605.1.15,gzip(gfe)', + }, + }, + }, + } + def _login(self): """ Attempt to log in to YouTube. + True is returned if successful or skipped. False is returned if login failed. @@ -281,14 +355,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor): if not self._login(): return - _DEFAULT_API_DATA = { - 'context': { - 'client': { - 'clientName': 'WEB', - 'clientVersion': '2.20201021.03.00', - }, - }, - } + _DEFAULT_API_DATA = {'context': _INNERTUBE_CLIENTS['web']['INNERTUBE_CONTEXT']} _YT_INITIAL_DATA_RE = r'(?:window\s*\[\s*["\']ytInitialData["\']\s*\]|ytInitialData)\s*=\s*({.+?})\s*;' _YT_INITIAL_PLAYER_RESPONSE_RE = r'ytInitialPlayerResponse\s*=\s*({.+?})\s*;' @@ -321,19 +388,24 @@ class YoutubeBaseInfoExtractor(InfoExtractor): '{0} {1} {2}'.format(time_now, self._SAPISID, origin).encode('utf-8')).hexdigest() return 'SAPISIDHASH {0}_{1}'.format(time_now, sapisidhash) - def _call_api(self, ep, query, video_id, fatal=True, headers=None): + def _call_api(self, ep, query, video_id, fatal=True, headers=None, + note='Downloading API JSON'): data = self._DEFAULT_API_DATA.copy() data.update(query) real_headers = {'content-type': 'application/json'} if headers: real_headers.update(headers) + # was: 'AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8' + api_key = self.get_param('youtube_innertube_key') return self._download_json( 'https://www.youtube.com/youtubei/v1/%s' % ep, video_id=video_id, - note='Downloading API JSON', errnote='Unable to download API page', + note=note, errnote='Unable to download API page', data=json.dumps(data).encode('utf8'), fatal=fatal, - headers=real_headers, - query={'key': 'AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8'}) + headers=real_headers, query=filter_dict({ + 'key': api_key, + 'prettyPrint': 'false', + })) def _extract_yt_initial_data(self, video_id, webpage): return self._parse_json( @@ -342,11 +414,32 @@ class YoutubeBaseInfoExtractor(InfoExtractor): self._YT_INITIAL_DATA_RE), webpage, 'yt initial data'), video_id) + def _extract_visitor_data(self, *args): + """ + Extract visitorData from an API response or ytcfg + + Appears to be used to track session state + """ + visitor_data = self.get_param('youtube_visitor_data') + if visitor_data: + return visitor_data + + return get_first( + args, (('VISITOR_DATA', + ('INNERTUBE_CONTEXT', 'client', 'visitorData'), + ('responseContext', 'visitorData')), + T(compat_str))) + def _extract_ytcfg(self, video_id, webpage): - return self._parse_json( - self._search_regex( - r'ytcfg\.set\s*\(\s*({.+?})\s*\)\s*;', webpage, 'ytcfg', - default='{}'), video_id, fatal=False) or {} + ytcfg = self._search_json( + r'ytcfg\.set\s*\(', webpage, 'ytcfg', video_id, + end_pattern=r'\)\s*;', default={}) + + traverse_obj(ytcfg, ( + 'INNERTUBE_CONTEXT', 'client', 'configInfo', + T(lambda x: x.pop('appInstallData', None)))) + + return ytcfg def _extract_video(self, renderer): video_id = renderer['videoId'] @@ -381,6 +474,27 @@ class YoutubeBaseInfoExtractor(InfoExtractor): 'uploader': uploader, } + @staticmethod + def _extract_thumbnails(data, *path_list, **kw_final_key): + """ + Extract thumbnails from thumbnails dict + + @param path_list: path list to level that contains 'thumbnails' key + """ + final_key = kw_final_key.get('final_key', 'thumbnails') + + return traverse_obj(data, (( + tuple(variadic(path) + (final_key, Ellipsis) + for path in path_list or [()])), { + 'url': ('url', T(url_or_none), + # Sometimes youtube gives a wrong thumbnail URL. See: + # https://github.com/yt-dlp/yt-dlp/issues/233 + # https://github.com/ytdl-org/youtube-dl/issues/28023 + T(lambda u: update_url(u, query=None) if u and 'maxresdefault' in u else u)), + 'height': ('height', T(int_or_none)), + 'width': ('width', T(int_or_none)), + }, T(lambda t: t if t.get('url') else None))) + def _search_results(self, query, params): data = { 'context': { @@ -395,42 +509,38 @@ class YoutubeBaseInfoExtractor(InfoExtractor): data['params'] = params for page_num in itertools.count(1): search = self._download_json( - 'https://www.youtube.com/youtubei/v1/search?key=AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8', + 'https://www.youtube.com/youtubei/v1/search', video_id='query "%s"' % query, note='Downloading page %s' % page_num, errnote='Unable to download API page', fatal=False, data=json.dumps(data).encode('utf8'), + query={ + # 'key': 'AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8', + 'prettyPrint': 'false', + }, headers={'content-type': 'application/json'}) if not search: break - slr_contents = try_get( + slr_contents = traverse_obj( search, - (lambda x: x['contents']['twoColumnSearchResultsRenderer']['primaryContents']['sectionListRenderer']['contents'], - lambda x: x['onResponseReceivedCommands'][0]['appendContinuationItemsAction']['continuationItems']), - list) + ('contents', 'twoColumnSearchResultsRenderer', 'primaryContents', + 'sectionListRenderer', 'contents'), + ('onResponseReceivedCommands', 0, 'appendContinuationItemsAction', + 'continuationItems'), + expected_type=list) if not slr_contents: break - for slr_content in slr_contents: - isr_contents = try_get( - slr_content, - lambda x: x['itemSectionRenderer']['contents'], - list) - if not isr_contents: - continue - for content in isr_contents: - if not isinstance(content, dict): - continue - video = content.get('videoRenderer') - if not isinstance(video, dict): - continue - video_id = video.get('videoId') - if not video_id: - continue - yield self._extract_video(video) - token = try_get( + for video in traverse_obj( + slr_contents, + (Ellipsis, 'itemSectionRenderer', 'contents', + Ellipsis, 'videoRenderer', + T(lambda v: v if v.get('videoId') else None))): + yield self._extract_video(video) + + token = traverse_obj( slr_contents, - lambda x: x[-1]['continuationItemRenderer']['continuationEndpoint']['continuationCommand']['token'], - compat_str) + (-1, 'continuationItemRenderer', 'continuationEndpoint', + 'continuationCommand', 'token', T(compat_str))) if not token: break data['continuation'] = token @@ -590,11 +700,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'invidious': '|'.join(_INVIDIOUS_SITES), } _PLAYER_INFO_RE = ( - r'/s/player/(?P<id>[a-zA-Z0-9_-]{8,})/player', - r'/(?P<id>[a-zA-Z0-9_-]{8,})/player(?:_ias\.vflset(?:/[a-zA-Z]{2,3}_[a-zA-Z]{2,3})?|-plasma-ias-(?:phone|tablet)-[a-z]{2}_[A-Z]{2}\.vflset)/base\.js$', - r'\b(?P<id>vfl[a-zA-Z0-9_-]+)\b.*?\.js$', + r'/s/player/(?P<id>[a-zA-Z0-9_-]{8,})/(?:tv-)?player', + r'/(?P<id>[a-zA-Z0-9_-]{8,})/player(?:_ias(?:_tce)?\.vflset(?:/[a-zA-Z]{2,3}_[a-zA-Z]{2,3})?|-plasma-ias-(?:phone|tablet)-[a-z]{2}_[A-Z]{2}\.vflset)/base\.js$', + r'\b(?P<id>vfl[a-zA-Z0-9_-]{6,})\b.*?\.js$', ) - _SUBTITLE_FORMATS = ('json3', 'srv1', 'srv2', 'srv3', 'ttml', 'vtt') + _SUBTITLE_FORMATS = ('json3', 'srv1', 'srv2', 'srv3', 'ttml', 'srt', 'vtt') _GEO_BYPASS = False @@ -1485,6 +1595,18 @@ class YoutubeIE(YoutubeBaseInfoExtractor): '397': {'acodec': 'none', 'vcodec': 'av01.0.05M.08'}, } + _PLAYER_JS_VARIANT_MAP = ( + ('main', 'player_ias.vflset/en_US/base.js'), + ('tcc', 'player_ias_tcc.vflset/en_US/base.js'), + ('tce', 'player_ias_tce.vflset/en_US/base.js'), + ('es5', 'player_es5.vflset/en_US/base.js'), + ('es6', 'player_es6.vflset/en_US/base.js'), + ('tv', 'tv-player-ias.vflset/tv-player-ias.js'), + ('tv_es6', 'tv-player-es6.vflset/tv-player-es6.js'), + ('phone', 'player-plasma-ias-phone-en_US.vflset/base.js'), + ('tablet', 'player-plasma-ias-tablet-en_US.vflset/base.js'), + ) + @classmethod def suitable(cls, url): if parse_qs(url).get('list', [None])[0]: @@ -1496,6 +1618,19 @@ class YoutubeIE(YoutubeBaseInfoExtractor): self._code_cache = {} self._player_cache = {} + def _get_player_js_version(self): + player_js_version = self.get_param('youtube_player_js_version') or '20348@0004de42' + sts_hash = self._search_regex( + ('^actual$(^)?(^)?', r'^([0-9]{5,})@([0-9a-f]{8,})$'), + player_js_version, 'player_js_version', group=(1, 2), default=None) + if sts_hash: + return sts_hash + self.report_warning( + 'Invalid player JS version "{0}" specified. ' + 'It should be "{1}" or in the format of {2}'.format( + player_js_version, 'actual', 'SignatureTimeStamp@Hash'), only_once=True) + return None, None + # *ytcfgs, webpage=None def _extract_player_url(self, *ytcfgs, **kw_webpage): if ytcfgs and not isinstance(ytcfgs[0], dict): @@ -1506,64 +1641,139 @@ class YoutubeIE(YoutubeBaseInfoExtractor): webpage or '', 'player URL', fatal=False) if player_url: ytcfgs = ytcfgs + ({'PLAYER_JS_URL': player_url},) - return traverse_obj( + player_url = traverse_obj( ytcfgs, (Ellipsis, 'PLAYER_JS_URL'), (Ellipsis, 'WEB_PLAYER_CONTEXT_CONFIGS', Ellipsis, 'jsUrl'), get_all=False, expected_type=lambda u: urljoin('https://www.youtube.com', u)) + player_id_override = self._get_player_js_version()[1] + + requested_js_variant = self.get_param('youtube_player_js_variant') or 'main' + variant_js = next( + (v for k, v in self._PLAYER_JS_VARIANT_MAP if k == requested_js_variant), + None) + if variant_js: + player_id = player_id_override or self._extract_player_info(player_url) + original_url = player_url + player_url = '/s/player/{0}/{1}'.format(player_id, variant_js) + if original_url != player_url: + self.write_debug( + 'Forcing "{0}" player JS variant for player {1}\n' + ' original url = {2}'.format( + requested_js_variant, player_id, original_url), + only_once=True) + elif requested_js_variant != 'actual': + self.report_warning( + 'Invalid player JS variant name "{0}" requested. ' + 'Valid choices are: {1}'.format( + requested_js_variant, ','.join(k for k, _ in self._PLAYER_JS_VARIANT_MAP)), + only_once=True) + + return urljoin('https://www.youtube.com', player_url) + def _download_player_url(self, video_id, fatal=False): res = self._download_webpage( 'https://www.youtube.com/iframe_api', note='Downloading iframe API JS', video_id=video_id, fatal=fatal) player_version = self._search_regex( r'player\\?/([0-9a-fA-F]{8})\\?/', res or '', 'player version', fatal=fatal, - default=NO_DEFAULT if res else None) - if player_version: - return 'https://www.youtube.com/s/player/{0}/player_ias.vflset/en_US/base.js'.format(player_version) + default=NO_DEFAULT if res else None) or None + return player_version and 'https://www.youtube.com/s/player/{0}/player_ias.vflset/en_US/base.js'.format(player_version) def _signature_cache_id(self, example_sig): """ Return a string representation of a signature """ return '.'.join(compat_str(len(part)) for part in example_sig.split('.')) - @classmethod - def _extract_player_info(cls, player_url): - for player_re in cls._PLAYER_INFO_RE: - id_m = re.search(player_re, player_url) - if id_m: - break - else: - raise ExtractorError('Cannot identify player %r' % player_url) - return id_m.group('id') + def _extract_player_info(self, player_url): + try: + return self._search_regex( + self._PLAYER_INFO_RE, player_url, 'player info', group='id') + except ExtractorError as e: + raise ExtractorError( + 'Cannot identify player %r' % (player_url,), cause=e) - def _load_player(self, video_id, player_url, fatal=True, player_id=None): - if not player_id: + def _player_js_cache_key(self, player_url, extra_id=None, _cache={}): + if player_url not in _cache: player_id = self._extract_player_info(player_url) - if player_id not in self._code_cache: + player_path = remove_start( + compat_urllib_parse.urlparse(player_url).path, + '/s/player/{0}/'.format(player_id)) + variant = next((k for k, v in self._PLAYER_JS_VARIANT_MAP + if v == player_path), None) + if not variant: + variant = next( + (k for k, v in self._PLAYER_JS_VARIANT_MAP + if re.match(re.escape(v).replace('en_US', r'\w+') + '$', player_path)), + None) + if not variant: + self.write_debug( + 'Unable to determine player JS variant\n' + ' player = {0}'.format(player_url), only_once=True) + variant = re.sub(r'[^a-zA-Z0-9]', '_', remove_end(player_path, '.js')) + _cache[player_url] = join_nonempty(player_id, variant) + + if extra_id: + extra_id = '-'.join((_cache[player_url], extra_id)) + assert os.path.basename(extra_id) == extra_id + return extra_id + return _cache[player_url] + + def _load_player(self, video_id, player_url, fatal=True): + player_js_key = self._player_js_cache_key(player_url) + if player_js_key not in self._code_cache: code = self._download_webpage( player_url, video_id, fatal=fatal, - note='Downloading player ' + player_id, - errnote='Download of %s failed' % player_url) + note='Downloading player {0}'.format(player_js_key), + errnote='Download of {0} failed'.format(player_url)) if code: - self._code_cache[player_id] = code - return self._code_cache[player_id] if fatal else self._code_cache.get(player_id) + self._code_cache[player_js_key] = code + return self._code_cache.get(player_js_key) + + def _load_player_data_from_cache(self, name, player_url, extra_id=None): + cache_id = ('youtube-{0}'.format(name), self._player_js_cache_key(player_url, extra_id)) + data = self._player_cache.get(cache_id) + if data: + return data + + data = self.cache.load(*cache_id, min_ver='2025.04.07') + if data: + self._player_cache[cache_id] = data + return data + + def _store_player_data_to_cache(self, name, player_url, data, extra_id=None): + cache_id = ('youtube-{0}'.format(name), self._player_js_cache_key(player_url, extra_id)) + + if cache_id not in self._player_cache: + self.cache.store(cache_id[0], cache_id[1], data) + self._player_cache[cache_id] = data + + def _remove_player_data_from_cache(self, name, player_url, extra_id=None): + cache_id = ('youtube-{0}'.format(name), self._player_js_cache_key(player_url, extra_id)) + + if cache_id in self._player_cache: + self.cache.clear(*cache_id) + self._player_cache.pop(cache_id, None) def _extract_signature_function(self, video_id, player_url, example_sig): - player_id = self._extract_player_info(player_url) + # player_id = self._extract_player_info(player_url) # Read from filesystem cache - func_id = 'js_{0}_{1}'.format( - player_id, self._signature_cache_id(example_sig)) - assert os.path.basename(func_id) == func_id - - self.write_debug('Extracting signature function {0}'.format(func_id)) - cache_spec, code = self.cache.load('youtube-sigfuncs', func_id), None + extra_id = self._signature_cache_id(example_sig) + self.write_debug('Extracting signature function {0}-{1}'.format(player_url, extra_id)) + cache_spec, code = self._load_player_data_from_cache( + 'sigfuncs', player_url, extra_id=extra_id), None if not cache_spec: - code = self._load_player(video_id, player_url, player_id) - if code: - res = self._parse_sig_js(code) - test_string = ''.join(map(compat_chr, range(len(example_sig)))) - cache_spec = [ord(c) for c in res(test_string)] - self.cache.store('youtube-sigfuncs', func_id, cache_spec) + code = self._load_player(video_id, player_url) + if code: + res = self._parse_sig_js(code) + test_string = ''.join(map(compat_chr, range(len(example_sig)))) + cache_spec = [ord(c) for c in res(test_string)] + self._store_player_data_to_cache( + 'sigfuncs', player_url, cache_spec, extra_id=extra_id) + else: + self.report_warning( + 'Failed to compute signature function {0}-{1}'.format( + player_url, extra_id)) return lambda s: ''.join(s[i] for i in cache_spec) @@ -1609,6 +1819,23 @@ class YoutubeIE(YoutubeBaseInfoExtractor): ' return %s\n') % (signature_id_tuple, expr_code) self.to_screen('Extracted signature function:\n' + code) + def _extract_sig_fn(self, jsi, funcname): + var_ay = self._search_regex( + r'''(?x) + (?:\*/|\{|\n|^)\s*(?:'[^']+'\s*;\s*) + (var\s*[\w$]+\s*=\s*(?: + ('|")(?:\\\2|(?!\2).)+\2\s*\.\s*split\(\s*('|")\W+\3\s*\)| + \[\s*(?:('|")(?:\\\4|(?!\4).)*\4\s*(?:(?=\])|,\s*))+\] + ))(?=\s*[,;]) + ''', jsi.code, 'useful values', default='') + + sig_fn = jsi.extract_function_code(funcname) + + if var_ay: + sig_fn = (sig_fn[0], ';\n'.join((var_ay, sig_fn[1]))) + + return sig_fn + def _parse_sig_js(self, jscode): # Examples where `sig` is funcname: # sig=function(a){a=a.split(""); ... ;return a.join("")}; @@ -1634,8 +1861,12 @@ class YoutubeIE(YoutubeBaseInfoExtractor): jscode, 'Initial JS player signature function name', group='sig') jsi = JSInterpreter(jscode) - initial_function = jsi.extract_function(funcname) - return lambda s: initial_function([s]) + + initial_function = self._extract_sig_fn(jsi, funcname) + + func = jsi.extract_function_from_code(*initial_function) + + return lambda s: func([s]) def _cached(self, func, *cache_id): def inner(*args, **kwargs): @@ -1695,6 +1926,24 @@ class YoutubeIE(YoutubeBaseInfoExtractor): return ret def _extract_n_function_name(self, jscode): + func_name, idx = None, None + + def generic_n_function_search(func_name=None): + return self._search_regex( + r'''(?xs) + (?:(?<=[^\w$])|^) # instead of \b, which ignores $ + (?P<name>%s)\s*=\s*function\((?!\d)[a-zA-Z\d_$]+\) + \s*\{(?:(?!};).)+?(?: + ["']enhanced_except_ | + return\s*(?P<q>"|')[a-zA-Z\d-]+_w8_(?P=q)\s*\+\s*[\w$]+ + ) + ''' % (func_name or r'(?!\d)[a-zA-Z\d_$]+',), jscode, + 'Initial JS player n function name', group='name', + default=None if func_name else NO_DEFAULT) + + # these special cases are redundant and probably obsolete (2025-04): + # they make the tests run ~10% faster without fallback warnings + r""" func_name, idx = self._search_regex( # (y=NuD(),Mw(k),q=k.Z[y]||null)&&(q=narray[idx](q),k.set(y,q),k.V||NuD(''))}}; # (R="nn"[+J.Z],mW(J),N=J.K[R]||null)&&(N=narray[idx](N),J.set(R,N))}}; @@ -1721,41 +1970,59 @@ class YoutubeIE(YoutubeBaseInfoExtractor): \(\s*[\w$]+\s*\) ''', jscode, 'Initial JS player n function name', group=('nfunc', 'idx'), default=(None, None)) + """ + + if not func_name: + # nfunc=function(x){...}|function nfunc(x); ... + # ... var y=[nfunc]|y[idx]=nfunc); + # obvious REs hang, so use a two-stage tactic + for m in re.finditer(r'''(?x) + [\n;]var\s(?:(?:(?!,).)+,|\s)*?(?!\d)[\w$]+(?:\[(?P<idx>\d+)\])?\s*=\s* + (?(idx)|\[\s*)(?P<nfunc>(?!\d)[\w$]+)(?(idx)|\s*\]) + \s*?[;\n] + ''', jscode): + fn = self._search_regex( + r'[;,]\s*(function\s+)?({0})(?(1)|\s*=\s*function)\s*\((?!\d)[\w$]+\)\s*\{1}(?!\s*return\s)'.format( + re.escape(m.group('nfunc')), '{'), + jscode, 'Initial JS player n function name (2)', group=2, default=None) + if fn: + func_name = fn + idx = m.group('idx') + if generic_n_function_search(func_name): + # don't look any further + break + # thx bashonly: yt-dlp/yt-dlp/pull/10611 if not func_name: - self.report_warning('Falling back to generic n function search') - return self._search_regex( - r'''(?xs) - (?:(?<=[^\w$])|^) # instead of \b, which ignores $ - (?P<name>(?!\d)[a-zA-Z\d_$]+)\s*=\s*function\((?!\d)[a-zA-Z\d_$]+\) - \s*\{(?:(?!};).)+?(?: - ["']enhanced_except_ | - return\s*(?P<q>"|')[a-zA-Z\d-]+_w8_(?P=q)\s*\+\s*[\w$]+ - ) - ''', jscode, 'Initial JS player n function name', group='name') + self.report_warning('Falling back to generic n function search', only_once=True) + return generic_n_function_search() + if not idx: return func_name return self._search_json( - r'var\s+{0}\s*='.format(re.escape(func_name)), jscode, + r'(?<![\w-])var\s(?:(?:(?!,).)+,|\s)*?{0}\s*='.format(re.escape(func_name)), jscode, 'Initial JS player n function list ({0}.{1})'.format(func_name, idx), - func_name, contains_pattern=r'\[[\s\S]+\]', end_pattern='[,;]', + func_name, contains_pattern=r'\[.+\]', end_pattern='[,;]', transform_source=js_to_json)[int(idx)] def _extract_n_function_code(self, video_id, player_url): player_id = self._extract_player_info(player_url) - func_code = self.cache.load('youtube-nsig', player_id) + func_code = self._load_player_data_from_cache('nsig', player_url) jscode = func_code or self._load_player(video_id, player_url) jsi = JSInterpreter(jscode) if func_code: return jsi, player_id, func_code - func_name = self._extract_n_function_name(jscode) + return self._extract_n_function_code_jsi(video_id, jsi, player_id, player_url) - func_code = jsi.extract_function_code(func_name) + def _extract_n_function_code_jsi(self, video_id, jsi, player_id=None, player_url=None): + func_name = self._extract_n_function_name(jsi.code) - self.cache.store('youtube-nsig', player_id, func_code) + func_code = self._extract_sig_fn(jsi, func_name) + if player_url: + self._store_player_data_to_cache('nsig', player_url, func_code) return jsi, player_id, func_code def _extract_n_function_from_code(self, jsi, func_code): @@ -1788,7 +2055,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor): n_param = n_param[-1] n_response = decrypt_nsig(n_param)(n_param, video_id, player_url) if n_response is None: - # give up if descrambling failed + # give up and forget cached data if descrambling failed + self._remove_player_data_from_cache('nsig', player_url) break fmt['url'] = update_url_query(fmt['url'], {'n': n_response}) @@ -1796,21 +2064,37 @@ class YoutubeIE(YoutubeBaseInfoExtractor): def _extract_signature_timestamp(self, video_id, player_url, ytcfg=None, fatal=False): """ Extract signatureTimestamp (sts) + Required to tell API what sig/player version is in use. """ - sts = traverse_obj(ytcfg, 'STS', expected_type=int) - if not sts: - # Attempt to extract from player - if player_url is None: - error_msg = 'Cannot extract signature timestamp without player_url.' - if fatal: - raise ExtractorError(error_msg) - self.report_warning(error_msg) - return - code = self._load_player(video_id, player_url, fatal=fatal) - sts = int_or_none(self._search_regex( - r'(?:signatureTimestamp|sts)\s*:\s*(?P<sts>[0-9]{5})', code or '', - 'JS player signature timestamp', group='sts', fatal=fatal)) + sts = traverse_obj( + (self._get_player_js_version(), ytcfg), + (0, 0), + (1, 'STS'), + expected_type=int_or_none) + + if sts: + return sts + + if not player_url: + error_msg = 'Cannot extract signature timestamp without player url' + if fatal: + raise ExtractorError(error_msg) + self.report_warning(error_msg) + return None + + sts = self._load_player_data_from_cache('sts', player_url) + if sts: + return sts + + # Attempt to extract from player + code = self._load_player(video_id, player_url, fatal=fatal) + sts = int_or_none(self._search_regex( + r'(?:signatureTimestamp|sts)\s*:\s*(?P<sts>[0-9]{5})', code or '', + 'JS player signature timestamp', group='sts', fatal=fatal)) + if sts: + self._store_player_data_to_cache('sts', player_url, sts) + return sts def _mark_watched(self, video_id, player_response): @@ -1885,7 +2169,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor): raise ExtractorError('Invalid URL: %s' % url) return mobj.group(2) - def _extract_chapters_from_json(self, data, video_id, duration): + @staticmethod + def _extract_chapters_from_json(data, video_id, duration): chapters_list = try_get( data, lambda x: x['playerOverlays'] @@ -1935,8 +2220,12 @@ class YoutubeIE(YoutubeBaseInfoExtractor): video_id = self._match_id(url) base_url = self.http_scheme() + '//www.youtube.com/' webpage_url = base_url + 'watch?v=' + video_id + ua = traverse_obj(self._INNERTUBE_CLIENTS, ( + 'web', 'INNERTUBE_CONTEXT', 'client', 'userAgent')) + headers = {'User-Agent': ua} if ua else None webpage = self._download_webpage( - webpage_url + '&bpctr=9999999999&has_verified=1', video_id, fatal=False) + webpage_url + '&bpctr=9999999999&has_verified=1', video_id, + headers=headers, fatal=False) player_response = None player_url = None @@ -1944,12 +2233,16 @@ class YoutubeIE(YoutubeBaseInfoExtractor): player_response = self._extract_yt_initial_variable( webpage, self._YT_INITIAL_PLAYER_RESPONSE_RE, video_id, 'initial player response') + is_live = traverse_obj(player_response, ('videoDetails', 'isLive')) + + fetched_timestamp = None if False and not player_response: player_response = self._call_api( 'player', {'videoId': video_id}, video_id) if True or not player_response: origin = 'https://www.youtube.com' pb_context = {'html5Preference': 'HTML5_PREF_WANTS'} + fetched_timestamp = int(time.time()) player_url = self._extract_player_url(webpage) ytcfg = self._extract_ytcfg(video_id, webpage or '') @@ -1957,46 +2250,85 @@ class YoutubeIE(YoutubeBaseInfoExtractor): if sts: pb_context['signatureTimestamp'] = sts - query = { - 'playbackContext': { - 'contentPlaybackContext': pb_context, - 'contentCheckOk': True, - 'racyCheckOk': True, - }, - 'context': { - 'client': { - 'clientName': 'MWEB', - 'clientVersion': '2.20241202.07.00', - 'hl': 'en', - 'userAgent': 'Mozilla/5.0 (iPad; CPU OS 16_7_10 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/16.6 Mobile/15E148 Safari/604.1,gzip(gfe)', - 'timeZone': 'UTC', - 'utcOffsetMinutes': 0, - }, - }, - 'videoId': video_id, - } + client_names = traverse_obj(self._INNERTUBE_CLIENTS, ( + T(dict.items), lambda _, k_v: not k_v[1].get('REQUIRE_PO_TOKEN'), + 0))[:1] + if 'web' not in client_names: + # webpage links won't download: ignore links and playability + player_response = filter_dict( + player_response or {}, + lambda k, _: k not in ('streamingData', 'playabilityStatus')) + + if is_live and 'ios' not in client_names: + client_names.append('ios') + headers = { - 'X-YouTube-Client-Name': '2', - 'X-YouTube-Client-Version': '2.20241202.07.00', - 'Origin': origin, 'Sec-Fetch-Mode': 'navigate', - 'User-Agent': query['context']['client']['userAgent'], + 'Origin': origin, + 'X-Goog-Visitor-Id': self._extract_visitor_data(ytcfg) or '', } auth = self._generate_sapisidhash_header(origin) if auth is not None: headers['Authorization'] = auth headers['X-Origin'] = origin - player_response = self._call_api('player', query, video_id, fatal=False, headers=headers) + for client in traverse_obj(self._INNERTUBE_CLIENTS, (client_names, T(dict))): + + query = { + 'playbackContext': { + 'contentPlaybackContext': pb_context, + }, + 'contentCheckOk': True, + 'racyCheckOk': True, + 'context': { + 'client': merge_dicts( + traverse_obj(client, ('INNERTUBE_CONTEXT', 'client')), { + 'hl': 'en', + 'timeZone': 'UTC', + 'utcOffsetMinutes': 0, + }), + }, + 'videoId': video_id, + } + + api_headers = merge_dicts(headers, traverse_obj(client, { + 'X-YouTube-Client-Name': 'INNERTUBE_CONTEXT_CLIENT_NAME', + 'X-YouTube-Client-Version': ( + 'INNERTUBE_CONTEXT', 'client', 'clientVersion'), + 'User-Agent': ( + 'INNERTUBE_CONTEXT', 'client', 'userAgent'), + })) + + api_player_response = self._call_api( + 'player', query, video_id, fatal=False, headers=api_headers, + note=join_nonempty( + 'Downloading', traverse_obj(query, ( + 'context', 'client', 'clientName')), + 'API JSON', delim=' ')) + + hls = traverse_obj( + (player_response, api_player_response), + (Ellipsis, 'streamingData', 'hlsManifestUrl', T(url_or_none))) + fetched_timestamp = int(time.time()) + if len(hls) == 2 and not hls[0] and hls[1]: + player_response['streamingData']['hlsManifestUrl'] = hls[1] + else: + video_details = merge_dicts(*traverse_obj( + (player_response, api_player_response), + (Ellipsis, 'videoDetails', T(dict)))) + player_response.update(filter_dict( + api_player_response or {}, cndn=lambda k, _: k != 'captions')) + player_response['videoDetails'] = video_details def is_agegated(playability): - if not isinstance(playability, dict): - return + # playability: dict + if not playability: + return False if playability.get('desktopLegacyAgeGateReason'): return True - reasons = filter(None, (playability.get(r) for r in ('status', 'reason'))) + reasons = traverse_obj(playability, (('status', 'reason'),)) AGE_GATE_REASONS = ( 'confirm your age', 'age-restricted', 'inappropriate', # reason 'age_verification_required', 'age_check_required', # status @@ -2054,15 +2386,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor): trailer_video_id, self.ie_key(), trailer_video_id) def get_text(x): - if not x: - return - text = x.get('simpleText') - if text and isinstance(text, compat_str): - return text - runs = x.get('runs') - if not isinstance(runs, list): - return - return ''.join([r['text'] for r in runs if isinstance(r.get('text'), compat_str)]) + return ''.join(traverse_obj( + x, (('simpleText',),), ('runs', Ellipsis, 'text'), + expected_type=compat_str)) search_meta = ( (lambda x: self._html_search_meta(x, webpage, default=None)) @@ -2130,6 +2456,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): itag_qualities = {} q = qualities(['tiny', 'small', 'medium', 'large', 'hd720', 'hd1080', 'hd1440', 'hd2160', 'hd2880', 'highres']) CHUNK_SIZE = 10 << 20 + is_live = video_details.get('isLive') streaming_data = player_response.get('streamingData') or {} streaming_formats = streaming_data.get('formats') or [] @@ -2139,11 +2466,19 @@ class YoutubeIE(YoutubeBaseInfoExtractor): return LazyList({ 'url': update_url_query(f['url'], { 'range': '{0}-{1}'.format(range_start, min(range_start + CHUNK_SIZE - 1, f['filesize'])), - }) + }), } for range_start in range(0, f['filesize'], CHUNK_SIZE)) lower = lambda s: s.lower() + if is_live: + fetched_timestamp = None + elif fetched_timestamp is not None: + # Handle preroll waiting period + preroll_sleep = self.get_param('youtube_preroll_sleep') + preroll_sleep = int_or_none(preroll_sleep, default=6) + fetched_timestamp += preroll_sleep + for fmt in streaming_formats: if fmt.get('targetDurationSec'): continue @@ -2240,6 +2575,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'downloader_options': {'http_chunk_size': CHUNK_SIZE}, # No longer useful? }) + if fetched_timestamp: + dct['available_at'] = fetched_timestamp + formats.append(dct) def process_manifest_format(f, proto, client_name, itag, all_formats=False): @@ -2257,6 +2595,10 @@ class YoutubeIE(YoutubeBaseInfoExtractor): if f.get('source_preference') is None: f['source_preference'] = -1 + # Deprioritize since its pre-merged m3u8 formats may have lower quality audio streams + if client_name == 'web_safari' and proto == 'hls' and not is_live: + f['source_preference'] -= 1 + if itag in ('616', '235'): f['format_note'] = join_nonempty(f.get('format_note'), 'Premium', delim=' ') f['source_preference'] += 100 @@ -2273,14 +2615,15 @@ class YoutubeIE(YoutubeBaseInfoExtractor): hls_manifest_url = streaming_data.get('hlsManifestUrl') if hls_manifest_url: - for f in self._extract_m3u8_formats( - hls_manifest_url, video_id, 'mp4', fatal=False): + formats.extend( + f for f in self._extract_m3u8_formats( + hls_manifest_url, video_id, 'mp4', + entry_protocol='m3u8_native', live=is_live, fatal=False) if process_manifest_format( - f, 'hls', None, self._search_regex( - r'/itag/(\d+)', f['url'], 'itag', default=None)): - formats.append(f) + f, 'hls', None, self._search_regex( + r'/itag/(\d+)', f['url'], 'itag', default=None))) - if self._downloader.params.get('youtube_include_dash_manifest', True): + if self.get_param('youtube_include_dash_manifest', True): dash_manifest_url = streaming_data.get('dashManifestUrl') if dash_manifest_url: for f in self._extract_mpd_formats( @@ -2307,7 +2650,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): playability_status, lambda x: x['errorScreen']['playerErrorMessageRenderer'], dict) or {} - reason = get_text(pemr.get('reason')) or playability_status.get('reason') + reason = get_text(pemr.get('reason')) or playability_status.get('reason') or '' subreason = pemr.get('subreason') if subreason: subreason = clean_html(get_text(subreason)) @@ -2319,7 +2662,12 @@ class YoutubeIE(YoutubeBaseInfoExtractor): self.raise_geo_restricted( subreason, countries) reason += '\n' + subreason + if reason: + if 'sign in' in reason.lower(): + self.raise_login_required(remove_end(reason, 'This helps protect our community. Learn more')) + elif traverse_obj(playability_status, ('errorScreen', 'playerCaptchaViewModel', T(dict))): + reason += '. YouTube is requiring a captcha challenge before playback' raise ExtractorError(reason, expected=True) self._sort_formats(formats) @@ -2380,8 +2728,6 @@ class YoutubeIE(YoutubeBaseInfoExtractor): # Strictly de-prioritize damaged formats f['preference'] = -10 - is_live = video_details.get('isLive') - owner_profile_url = self._yt_urljoin(self._extract_author_var( webpage, 'url', videodetails=video_details, metadata=microformat)) @@ -2416,14 +2762,17 @@ class YoutubeIE(YoutubeBaseInfoExtractor): } pctr = traverse_obj( - player_response, - ('captions', 'playerCaptionsTracklistRenderer', T(dict))) + (player_response, api_player_response), + (Ellipsis, 'captions', 'playerCaptionsTracklistRenderer', T(dict))) if pctr: def process_language(container, base_url, lang_code, query): lang_subs = [] for fmt in self._SUBTITLE_FORMATS: query.update({ 'fmt': fmt, + # xosf=1 causes undesirable text position data for vtt, json3 & srv* subtitles + # See: https://github.com/yt-dlp/yt-dlp/issues/13654 + 'xosf': [], }) lang_subs.append({ 'ext': fmt, @@ -2434,19 +2783,21 @@ class YoutubeIE(YoutubeBaseInfoExtractor): def process_subtitles(): subtitles = {} for caption_track in traverse_obj(pctr, ( - 'captionTracks', lambda _, v: v.get('baseUrl'))): + Ellipsis, 'captionTracks', lambda _, v: ( + v.get('baseUrl') and v.get('languageCode')))): + base_url = self._yt_urljoin(caption_track['baseUrl']) if not base_url: continue + lang_code = caption_track['languageCode'] if caption_track.get('kind') != 'asr': - lang_code = caption_track.get('languageCode') - if not lang_code: - continue process_language( subtitles, base_url, lang_code, {}) continue automatic_captions = {} + process_language( + automatic_captions, base_url, lang_code, {}) for translation_language in traverse_obj(pctr, ( - 'translationLanguages', lambda _, v: v.get('languageCode'))): + Ellipsis, 'translationLanguages', lambda _, v: v.get('languageCode'))): translation_language_code = translation_language['languageCode'] process_language( automatic_captions, base_url, translation_language_code, @@ -2463,7 +2814,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): for d_k, s_ks in [('start', ('start', 't')), ('end', ('end',))]: d_k += '_time' if d_k not in info and k in s_ks: - info[d_k] = parse_duration(query[k][0]) + info[d_k] = parse_duration(v[0]) if video_description: # Youtube Music Auto-generated description @@ -2492,6 +2843,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): initial_data = self._call_api( 'next', {'videoId': video_id}, video_id, fatal=False) + initial_sdcr = None if initial_data: chapters = self._extract_chapters_from_json( initial_data, video_id, duration) @@ -2511,9 +2863,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor): for next_num, content in enumerate(contents, start=1): mmlir = content.get('macroMarkersListItemRenderer') or {} start_time = chapter_time(mmlir) - end_time = chapter_time(try_get( - contents, lambda x: x[next_num]['macroMarkersListItemRenderer'])) \ - if next_num < len(contents) else duration + end_time = (traverse_obj( + contents, (next_num, 'macroMarkersListItemRenderer', T(chapter_time))) + if next_num < len(contents) else duration) if start_time is None or end_time is None: continue chapters.append({ @@ -2619,12 +2971,13 @@ class YoutubeIE(YoutubeBaseInfoExtractor): info['track'] = mrr_contents_text # this is not extraction but spelunking! - carousel_lockups = traverse_obj( - initial_data, - ('engagementPanels', Ellipsis, 'engagementPanelSectionListRenderer', - 'content', 'structuredDescriptionContentRenderer', 'items', Ellipsis, - 'videoDescriptionMusicSectionRenderer', 'carouselLockups', Ellipsis), - expected_type=dict) or [] + initial_sdcr = traverse_obj(initial_data, ( + 'engagementPanels', Ellipsis, 'engagementPanelSectionListRenderer', + 'content', 'structuredDescriptionContentRenderer', T(dict)), + get_all=False) + carousel_lockups = traverse_obj(initial_sdcr, ( + 'items', Ellipsis, 'videoDescriptionMusicSectionRenderer', + 'carouselLockups', Ellipsis, T(dict))) or [] # try to reproduce logic from metadataRowContainerRenderer above (if it still is) fields = (('ALBUM', 'album'), ('ARTIST', 'artist'), ('SONG', 'track'), ('LICENSES', 'license')) # multiple_songs ? @@ -2649,6 +3002,23 @@ class YoutubeIE(YoutubeBaseInfoExtractor): self.mark_watched(video_id, player_response) + # Fallbacks for missing metadata + if initial_sdcr: + if info.get('description') is None: + info['description'] = traverse_obj(initial_sdcr, ( + 'items', Ellipsis, 'expandableVideoDescriptionBodyRenderer', + 'attributedDescriptionBodyText', 'content', T(compat_str)), + get_all=False) + # videoDescriptionHeaderRenderer also has publishDate/channel/handle/ucid, but not needed + if info.get('title') is None: + info['title'] = traverse_obj( + (initial_sdcr, initial_data), + (0, 'items', Ellipsis, 'videoDescriptionHeaderRenderer', T(dict)), + (1, 'playerOverlays', 'playerOverlayRenderer', 'videoDetails', + 'playerOverlayVideoDetailsRenderer', T(dict)), + expected_type=lambda x: self._get_text(x, 'title'), + get_all=False) + return merge_dicts( info, { 'uploader_id': self._extract_uploader_id(owner_profile_url), @@ -3050,13 +3420,9 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): @staticmethod def _extract_grid_item_renderer(item): - assert isinstance(item, dict) - for key, renderer in item.items(): - if not key.startswith('grid') or not key.endswith('Renderer'): - continue - if not isinstance(renderer, dict): - continue - return renderer + return traverse_obj(item, ( + T(dict.items), lambda _, k_v: k_v[0].startswith('grid') and k_v[0].endswith('Renderer'), + 1, T(dict)), get_all=False) @staticmethod def _get_text(r, k): @@ -3065,8 +3431,12 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): expected_type=txt_or_none) def _grid_entries(self, grid_renderer): - for item in grid_renderer['items']: - if not isinstance(item, dict): + for item in traverse_obj(grid_renderer, ('items', Ellipsis, T(dict))): + lockup_view_model = traverse_obj(item, ('lockupViewModel', T(dict))) + if lockup_view_model: + entry = self._extract_lockup_view_model(lockup_view_model) + if entry: + yield entry continue renderer = self._extract_grid_item_renderer(item) if not isinstance(renderer, dict): @@ -3135,8 +3505,8 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): shelf_renderer, lambda x: x['title']['runs'][0]['text'], compat_str) yield self.url_result(shelf_url, video_title=title) # Shelf may not contain shelf URL, fallback to extraction from content - for entry in self._shelf_entries_from_content(shelf_renderer): - yield entry + for from_ in self._shelf_entries_from_content(shelf_renderer): + yield from_ def _playlist_entries(self, video_list_renderer): for content in video_list_renderer['contents']: @@ -3150,10 +3520,51 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): continue yield self._extract_video(renderer) + def _extract_lockup_view_model(self, view_model): + content_id = view_model.get('contentId') + if not content_id: + return + content_type = view_model.get('contentType') + if content_type == 'LOCKUP_CONTENT_TYPE_VIDEO': + ie = YoutubeIE + url = update_url_query( + 'https://www.youtube.com/watch', {'v': content_id}) + thumb_keys = (None,) + elif content_type in ('LOCKUP_CONTENT_TYPE_PLAYLIST', 'LOCKUP_CONTENT_TYPE_PODCAST'): + ie = YoutubeTabIE + url = update_url_query( + 'https://www.youtube.com/playlist', {'list': content_id}) + thumb_keys = ('collectionThumbnailViewModel', 'primaryThumbnail') + else: + self.report_warning( + 'Unsupported lockup view model content type "{0}"{1}'.format(content_type, bug_reports_message()), + only_once=True) + return + thumb_keys = ('contentImage',) + thumb_keys + ('thumbnailViewModel', 'image') + return merge_dicts(self.url_result( + url, ie=ie.ie_key(), video_id=content_id), { + 'title': traverse_obj(view_model, ( + 'metadata', 'lockupMetadataViewModel', 'title', + 'content', T(compat_str))), + 'thumbnails': self._extract_thumbnails( + view_model, thumb_keys, final_key='sources'), + }) + + def _extract_shorts_lockup_view_model(self, view_model): + content_id = traverse_obj(view_model, ( + 'onTap', 'innertubeCommand', 'reelWatchEndpoint', 'videoId', + T(lambda v: v if YoutubeIE.suitable(v) else None))) + return merge_dicts(self.url_result( + content_id, ie=YoutubeIE.ie_key(), video_id=content_id), { + 'title': traverse_obj(view_model, ( + 'overlayMetadata', 'primaryText', 'content', T(compat_str))), + 'thumbnails': self._extract_thumbnails( + view_model, 'thumbnail', final_key='sources'), + }) if content_id else None + def _video_entry(self, video_renderer): video_id = video_renderer.get('videoId') - if video_id: - return self._extract_video(video_renderer) + return self._extract_video(video_renderer) if video_id else None def _post_thread_entries(self, post_thread_renderer): post_renderer = try_get( @@ -3185,21 +3596,15 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): yield self.url_result(ep_url, ie=YoutubeIE.ie_key(), video_id=video_id) def _post_thread_continuation_entries(self, post_thread_continuation): - contents = post_thread_continuation.get('contents') - if not isinstance(contents, list): - return - for content in contents: - renderer = content.get('backstagePostThreadRenderer') - if not isinstance(renderer, dict): - continue - for entry in self._post_thread_entries(renderer): - yield entry + for renderer in traverse_obj(post_thread_continuation, ( + 'contents', Ellipsis, 'backstagePostThreadRenderer', T(dict))): + for from_ in self._post_thread_entries(renderer): + yield from_ def _rich_grid_entries(self, contents): - for content in contents: - content = traverse_obj( - content, ('richItemRenderer', 'content'), - expected_type=dict) or {} + for content in traverse_obj( + contents, (Ellipsis, 'richItemRenderer', 'content'), + expected_type=dict): video_renderer = traverse_obj( content, 'videoRenderer', 'reelItemRenderer', expected_type=dict) @@ -3207,6 +3612,12 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): entry = self._video_entry(video_renderer) if entry: yield entry + # shorts item + shorts_lockup_view_model = content.get('shortsLockupViewModel') + if shorts_lockup_view_model: + entry = self._extract_shorts_lockup_view_model(shorts_lockup_view_model) + if entry: + yield entry # playlist renderer = traverse_obj( content, 'playlistRenderer', expected_type=dict) or {} @@ -3245,23 +3656,15 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): next_continuation = cls._extract_next_continuation_data(renderer) if next_continuation: return next_continuation - contents = [] - for key in ('contents', 'items'): - contents.extend(try_get(renderer, lambda x: x[key], list) or []) - for content in contents: - if not isinstance(content, dict): - continue - continuation_ep = try_get( - content, lambda x: x['continuationItemRenderer']['continuationEndpoint'], - dict) - if not continuation_ep: - continue - continuation = try_get( - continuation_ep, lambda x: x['continuationCommand']['token'], compat_str) + for command in traverse_obj(renderer, ( + ('contents', 'items', 'rows'), Ellipsis, 'continuationItemRenderer', + ('continuationEndpoint', ('button', 'buttonRenderer', 'command')), + (('commandExecutorCommand', 'commands', Ellipsis), None), T(dict))): + continuation = traverse_obj(command, ('continuationCommand', 'token', T(compat_str))) if not continuation: continue - ctp = continuation_ep.get('clickTrackingParams') - return YoutubeTabIE._build_continuation_query(continuation, ctp) + ctp = command.get('clickTrackingParams') + return cls._build_continuation_query(continuation, ctp) def _entries(self, tab, item_id, webpage): tab_content = try_get(tab, lambda x: x['content'], dict) @@ -3271,17 +3674,10 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): if slr_renderer: is_channels_tab = tab.get('title') == 'Channels' continuation = None - slr_contents = try_get(slr_renderer, lambda x: x['contents'], list) or [] - for slr_content in slr_contents: - if not isinstance(slr_content, dict): - continue - is_renderer = try_get(slr_content, lambda x: x['itemSectionRenderer'], dict) - if not is_renderer: - continue - isr_contents = try_get(is_renderer, lambda x: x['contents'], list) or [] - for isr_content in isr_contents: - if not isinstance(isr_content, dict): - continue + for is_renderer in traverse_obj(slr_renderer, ( + 'contents', Ellipsis, 'itemSectionRenderer', T(dict))): + for isr_content in traverse_obj(slr_renderer, ( + 'contents', Ellipsis, T(dict))): renderer = isr_content.get('playlistVideoListRenderer') if renderer: for entry in self._playlist_entries(renderer): @@ -3310,6 +3706,13 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): entry = self._video_entry(renderer) if entry: yield entry + renderer = isr_content.get('richGridRenderer') + if renderer: + for from_ in self._rich_grid_entries( + traverse_obj(renderer, ('contents', Ellipsis, T(dict)))): + yield from_ + continuation = self._extract_continuation(renderer) + continue if not continuation: continuation = self._extract_continuation(is_renderer) @@ -3319,8 +3722,9 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): rich_grid_renderer = tab_content.get('richGridRenderer') if not rich_grid_renderer: return - for entry in self._rich_grid_entries(rich_grid_renderer.get('contents') or []): - yield entry + for from_ in self._rich_grid_entries( + traverse_obj(rich_grid_renderer, ('contents', Ellipsis, T(dict)))): + yield from_ continuation = self._extract_continuation(rich_grid_renderer) @@ -3354,7 +3758,7 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): if not continuation: break if visitor_data: - headers['x-goog-visitor-id'] = visitor_data + headers['X-Goog-Visitor-Id'] = visitor_data data['continuation'] = continuation['continuation'] data['clickTracking'] = { 'clickTrackingParams': continuation['itct'], @@ -3366,8 +3770,12 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): # Downloading page may result in intermittent 5xx HTTP error # that is usually worked around with a retry response = self._download_json( - 'https://www.youtube.com/youtubei/v1/browse?key=AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8', + 'https://www.youtube.com/youtubei/v1/browse', None, 'Downloading page %d%s' % (page_num, ' (retry #%d)' % count if count else ''), + query={ + # 'key': 'AIzaSyAO_FJ2SlqU8Q4STEHLGCilw_Y9_11qcW8', + 'prettyPrint': 'false', + }, headers=headers, data=json.dumps(data).encode('utf8')) break except ExtractorError as e: @@ -3464,18 +3872,34 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): uploader['channel'] = uploader['uploader'] return uploader - @classmethod - def _extract_alert(cls, data): - alerts = [] - for alert in traverse_obj(data, ('alerts', Ellipsis), expected_type=dict): - alert_text = traverse_obj( - alert, (None, lambda x: x['alertRenderer']['text']), get_all=False) - if not alert_text: - continue - text = cls._get_text(alert_text, 'text') - if text: - alerts.append(text) - return '\n'.join(alerts) + def _extract_and_report_alerts(self, data, expected=True, fatal=True, only_once=False): + + def alerts(): + for alert in traverse_obj(data, ('alerts', Ellipsis), expected_type=dict): + alert_dict = traverse_obj( + alert, 'alertRenderer', None, expected_type=dict, get_all=False) + alert_type = traverse_obj(alert_dict, 'type') + if not alert_type: + continue + message = self._get_text(alert_dict, 'text') + if message: + yield alert_type, message + + errors, warnings = [], [] + _IGNORED_WARNINGS = T('Unavailable videos will be hidden during playback') + for alert_type, alert_message in alerts(): + if alert_type.lower() == 'error' and fatal: + errors.append([alert_type, alert_message]) + elif alert_message not in _IGNORED_WARNINGS: + warnings.append([alert_type, alert_message]) + + for alert_type, alert_message in itertools.chain(warnings, errors[:-1]): + self.report_warning( + 'YouTube said: %s - %s' % (alert_type, alert_message), + only_once=only_once) + if errors: + raise ExtractorError( + 'YouTube said: %s' % (errors[-1][1],), expected=expected) def _extract_from_tabs(self, item_id, webpage, data, tabs): selected_tab = self._extract_selected_tab(tabs) @@ -3536,10 +3960,23 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): def _real_extract(self, url): item_id = self._match_id(url) url = update_url(url, netloc='www.youtube.com') - # Handle both video/playlist URLs qs = parse_qs(url) - video_id = qs.get('v', [None])[0] - playlist_id = qs.get('list', [None])[0] + + def qs_get(key, default=None): + return qs.get(key, [default])[-1] + + # Go around for /feeds/videos.xml?playlist_id={pl_id} + if item_id == 'feeds' and '/feeds/videos.xml?' in url: + playlist_id = qs_get('playlist_id') + if playlist_id: + return self.url_result( + update_url_query('https://www.youtube.com/playlist', { + 'list': playlist_id, + }), ie=self.ie_key(), video_id=playlist_id) + + # Handle both video/playlist URLs + video_id = qs_get('v') + playlist_id = qs_get('list') if video_id and playlist_id: if self._downloader.params.get('noplaylist'): self.to_screen('Downloading just video %s because of --no-playlist' % video_id) @@ -3562,10 +3999,10 @@ class YoutubeTabIE(YoutubeBaseInfoExtractor): compat_str) or video_id if video_id: return self.url_result(video_id, ie=YoutubeIE.ie_key(), video_id=video_id) + # Capture and output alerts - alert = self._extract_alert(data) - if alert: - raise ExtractorError(alert, expected=True) + self._extract_and_report_alerts(data) + # Failed to recognize raise ExtractorError('Unable to recognize tab page') @@ -3719,7 +4156,7 @@ class YoutubeFavouritesIE(YoutubeBaseInfoExtractor): 'only_matching': True, }] - def _real_extract(self, url): + def _real_extract(self, _): return self.url_result( 'https://www.youtube.com/playlist?list=LL', ie=YoutubeTabIE.ie_key()) @@ -3791,6 +4228,7 @@ class YoutubeFeedsInfoExtractor(YoutubeTabIE): Subclasses must define the _FEED_NAME property. """ + _LOGIN_REQUIRED = True @property @@ -3800,7 +4238,7 @@ class YoutubeFeedsInfoExtractor(YoutubeTabIE): def _real_initialize(self): self._login() - def _real_extract(self, url): + def _real_extract(self, _): return self.url_result( 'https://www.youtube.com/feed/%s' % self._FEED_NAME, ie=YoutubeTabIE.ie_key()) @@ -3815,7 +4253,7 @@ class YoutubeWatchLaterIE(InfoExtractor): 'only_matching': True, }] - def _real_extract(self, url): + def _real_extract(self, _): return self.url_result( 'https://www.youtube.com/playlist?list=WL', ie=YoutubeTabIE.ie_key()) @@ -3895,7 +4333,7 @@ class YoutubeTruncatedURLIE(InfoExtractor): 'only_matching': True, }] - def _real_extract(self, url): + def _real_extract(self, _): raise ExtractorError( 'Did you forget to quote the URL? Remember that & is a meta ' 'character in most shells, so you want to put the URL in quotes, ' diff --git a/youtube_dl/jsinterp.py b/youtube_dl/jsinterp.py index 7835187f5..7630e2099 100644 --- a/youtube_dl/jsinterp.py +++ b/youtube_dl/jsinterp.py @@ -1,10 +1,12 @@ # coding: utf-8 from __future__ import unicode_literals +import calendar import itertools import json import operator import re +import time from functools import update_wrapper, wraps @@ -12,8 +14,10 @@ from .utils import ( error_to_compat_str, ExtractorError, float_or_none, + int_or_none, js_to_json, remove_quotes, + str_or_none, unified_timestamp, variadic, write_string, @@ -24,6 +28,8 @@ from .compat import ( compat_collections_chain_map as ChainMap, compat_contextlib_suppress, compat_filter as filter, + compat_int, + compat_integer_types, compat_itertools_zip_longest as zip_longest, compat_map as map, compat_numeric_types, @@ -70,14 +76,27 @@ class JS_Undefined(object): pass -def _js_bit_op(op): +def _js_bit_op(op, is_shift=False): - def zeroise(x): - return 0 if x in (None, JS_Undefined, _NaN, _Infinity) else x + def zeroise(x, is_shift_arg=False): + if isinstance(x, compat_integer_types): + return (x % 32) if is_shift_arg else (x & 0xffffffff) + try: + x = float(x) + if is_shift_arg: + x = int(x % 32) + elif x < 0: + x = -compat_int(-x % 0xffffffff) + else: + x = compat_int(x % 0xffffffff) + except (ValueError, TypeError): + # also here for int(NaN), including float('inf') % 32 + x = 0 + return x @wraps_op(op) def wrapped(a, b): - return op(zeroise(a), zeroise(b)) & 0xffffffff + return op(zeroise(a), zeroise(b, is_shift)) & 0xffffffff return wrapped @@ -135,6 +154,7 @@ def _js_to_primitive(v): ) +# more exact: yt-dlp/yt-dlp#12110 def _js_toString(v): return ( 'undefined' if v is JS_Undefined @@ -143,7 +163,7 @@ def _js_toString(v): else 'null' if v is None # bool <= int: do this first else ('false', 'true')[v] if isinstance(v, bool) - else '{0:.7f}'.format(v).rstrip('.0') if isinstance(v, compat_numeric_types) + else re.sub(r'(?<=\d)\.?0*$', '', '{0:.7f}'.format(v)) if isinstance(v, compat_numeric_types) else _js_to_primitive(v)) @@ -220,7 +240,7 @@ def _js_ternary(cndn, if_true=True, if_false=False): def _js_unary_op(op): @wraps_op(op) - def wrapped(_, a): + def wrapped(a, _): return op(a) return wrapped @@ -253,8 +273,8 @@ def _js_typeof(expr): # avoid dict to maintain order # definition None => Defined in JSInterpreter._operator _OPERATORS = ( - ('>>', _js_bit_op(operator.rshift)), - ('<<', _js_bit_op(operator.lshift)), + ('>>', _js_bit_op(operator.rshift, True)), + ('<<', _js_bit_op(operator.lshift, True)), ('+', _js_add), ('-', _js_arith_op(operator.sub)), ('*', _js_arith_op(operator.mul)), @@ -263,17 +283,6 @@ _OPERATORS = ( ('**', _js_exp), ) -_COMP_OPERATORS = ( - ('===', _js_id_op(operator.is_)), - ('!==', _js_id_op(operator.is_not)), - ('==', _js_eq), - ('!=', _js_neq), - ('<=', _js_comp_op(operator.le)), - ('>=', _js_comp_op(operator.ge)), - ('<', _js_comp_op(operator.lt)), - ('>', _js_comp_op(operator.gt)), -) - _LOG_OPERATORS = ( ('|', _js_bit_op(operator.or_)), ('^', _js_bit_op(operator.xor)), @@ -290,13 +299,27 @@ _SC_OPERATORS = ( _UNARY_OPERATORS_X = ( ('void', _js_unary_op(lambda _: JS_Undefined)), ('typeof', _js_unary_op(_js_typeof)), + # avoid functools.partial here since Py2 update_wrapper(partial) -> no __module__ + ('!', _js_unary_op(lambda x: _js_ternary(x, if_true=False, if_false=True))), ) -_OPERATOR_RE = '|'.join(map(lambda x: re.escape(x[0]), _OPERATORS + _LOG_OPERATORS)) +_COMP_OPERATORS = ( + ('===', _js_id_op(operator.is_)), + ('!==', _js_id_op(operator.is_not)), + ('==', _js_eq), + ('!=', _js_neq), + ('<=', _js_comp_op(operator.le)), + ('>=', _js_comp_op(operator.ge)), + ('<', _js_comp_op(operator.lt)), + ('>', _js_comp_op(operator.gt)), +) + +_OPERATOR_RE = '|'.join(map(lambda x: re.escape(x[0]), _OPERATORS + _LOG_OPERATORS + _SC_OPERATORS)) _NAME_RE = r'[a-zA-Z_$][\w$]*' _MATCHING_PARENS = dict(zip(*zip('()', '{}', '[]'))) _QUOTES = '\'"/' +_NESTED_BRACKETS = r'[^[\]]+(?:\[[^[\]]+(?:\[[^\]]+\])?\])?' class JS_Break(ExtractorError): @@ -333,7 +356,7 @@ class LocalNameSpace(ChainMap): raise NotImplementedError('Deleting is not supported') def __repr__(self): - return 'LocalNameSpace%s' % (self.maps, ) + return 'LocalNameSpace({0!r})'.format(self.maps) class Debugger(object): @@ -354,6 +377,9 @@ class Debugger(object): @classmethod def wrap_interpreter(cls, f): + if not cls.ENABLED: + return f + @wraps(f) def interpret_statement(self, stmt, local_vars, allow_recursion, *args, **kwargs): if cls.ENABLED and stmt.strip(): @@ -389,11 +415,22 @@ class JSInterpreter(object): class Exception(ExtractorError): def __init__(self, msg, *args, **kwargs): expr = kwargs.pop('expr', None) + msg = str_or_none(msg, default='"None"') if expr is not None: msg = '{0} in: {1!r:.100}'.format(msg.rstrip(), expr) super(JSInterpreter.Exception, self).__init__(msg, *args, **kwargs) - class JS_RegExp(object): + class JS_Object(object): + def __getitem__(self, key): + if hasattr(self, key): + return getattr(self, key) + raise KeyError(key) + + def dump(self): + """Serialise the instance""" + raise NotImplementedError + + class JS_RegExp(JS_Object): RE_FLAGS = { # special knowledge: Python's re flags are bitmask values, current max 128 # invent new bitmask values well above that for literal parsing @@ -414,15 +451,24 @@ class JSInterpreter(object): def __init__(self, pattern_txt, flags=0): if isinstance(flags, compat_str): flags, _ = self.regex_flags(flags) - # First, avoid https://github.com/python/cpython/issues/74534 self.__self = None - self.__pattern_txt = pattern_txt.replace('[[', r'[\[') + pattern_txt = str_or_none(pattern_txt) or '(?:)' + # escape unintended embedded flags + pattern_txt = re.sub( + r'(\(\?)([aiLmsux]*)(-[imsx]+:|(?<!\?)\))', + lambda m: ''.join( + (re.escape(m.group(1)), m.group(2), re.escape(m.group(3))) + if m.group(3) == ')' + else ('(?:', m.group(2), m.group(3))), + pattern_txt) + # Avoid https://github.com/python/cpython/issues/74534 + self.source = pattern_txt.replace('[[', r'[\[') self.__flags = flags def __instantiate(self): if self.__self: return - self.__self = re.compile(self.__pattern_txt, self.__flags) + self.__self = re.compile(self.source, self.__flags) # Thx: https://stackoverflow.com/questions/44773522/setattr-on-python2-sre-sre-pattern for name in dir(self.__self): # Only these? Obviously __class__, __init__. @@ -430,16 +476,15 @@ class JSInterpreter(object): # that can't be setattr'd but also can't need to be copied. if name in ('__class__', '__init__', '__weakref__'): continue - setattr(self, name, getattr(self.__self, name)) + if name == 'flags': + setattr(self, name, getattr(self.__self, name, self.__flags)) + else: + setattr(self, name, getattr(self.__self, name)) def __getattr__(self, name): self.__instantiate() - # make Py 2.6 conform to its lying documentation - if name == 'flags': - self.flags = self.__flags - return self.flags - elif name == 'pattern': - self.pattern = self.__pattern_txt + if name == 'pattern': + self.pattern = self.source return self.pattern elif hasattr(self.__self, name): v = getattr(self.__self, name) @@ -447,6 +492,26 @@ class JSInterpreter(object): return v elif name in ('groupindex', 'groups'): return 0 if name == 'groupindex' else {} + else: + flag_attrs = ( # order by 2nd elt + ('hasIndices', 'd'), + ('global', 'g'), + ('ignoreCase', 'i'), + ('multiline', 'm'), + ('dotAll', 's'), + ('unicode', 'u'), + ('unicodeSets', 'v'), + ('sticky', 'y'), + ) + for k, c in flag_attrs: + if name == k: + return bool(self.RE_FLAGS[c] & self.__flags) + else: + if name == 'flags': + return ''.join( + (c if self.RE_FLAGS[c] & self.__flags else '') + for _, c in flag_attrs) + raise AttributeError('{0} has no attribute named {1}'.format(self, name)) @classmethod @@ -460,6 +525,85 @@ class JSInterpreter(object): flags |= cls.RE_FLAGS[ch] return flags, expr[idx + 1:] + def dump(self): + return '(/{0}/{1})'.format( + re.sub(r'(?<!\\)/', r'\/', self.source), + self.flags) + + @staticmethod + def escape(string_): + return re.escape(string_) + + class JS_Date(JS_Object): + _t = None + + @staticmethod + def __ymd_etc(*args, **kw_is_utc): + # args: year, monthIndex, day, hours, minutes, seconds, milliseconds + is_utc = kw_is_utc.get('is_utc', False) + + args = list(args[:7]) + args += [0] * (9 - len(args)) + args[1] += 1 # month 0..11 -> 1..12 + ms = args[6] + for i in range(6, 9): + args[i] = -1 # don't know + if is_utc: + args[-1] = 1 + # TODO: [MDN] When a segment overflows or underflows its expected + # range, it usually "carries over to" or "borrows from" the higher segment. + try: + mktime = calendar.timegm if is_utc else time.mktime + return mktime(time.struct_time(args)) * 1000 + ms + except (OverflowError, ValueError): + return None + + @classmethod + def UTC(cls, *args): + t = cls.__ymd_etc(*args, is_utc=True) + return _NaN if t is None else t + + @staticmethod + def parse(date_str, **kw_is_raw): + is_raw = kw_is_raw.get('is_raw', False) + + t = unified_timestamp(str_or_none(date_str), False) + return int(t * 1000) if t is not None else t if is_raw else _NaN + + @staticmethod + def now(**kw_is_raw): + is_raw = kw_is_raw.get('is_raw', False) + + t = time.time() + return int(t * 1000) if t is not None else t if is_raw else _NaN + + def __init__(self, *args): + if not args: + args = [self.now(is_raw=True)] + if len(args) == 1: + if isinstance(args[0], JSInterpreter.JS_Date): + self._t = int_or_none(args[0].valueOf(), default=None) + else: + arg_type = _js_typeof(args[0]) + if arg_type == 'string': + self._t = self.parse(args[0], is_raw=True) + elif arg_type == 'number': + self._t = int(args[0]) + else: + self._t = self.__ymd_etc(*args) + + def toString(self): + try: + return time.strftime('%a %b %0d %Y %H:%M:%S %Z%z', self._t).rstrip() + except TypeError: + return "Invalid Date" + + def valueOf(self): + return _NaN if self._t is None else self._t + + def dump(self): + return '(new Date({0}))'.format(self.toString()) + @classmethod def __op_chars(cls): op_chars = set(';,[') @@ -563,6 +707,68 @@ class JSInterpreter(object): _SC_OPERATORS, _LOG_OPERATORS, _COMP_OPERATORS, _OPERATORS, _UNARY_OPERATORS_X)) return _cached + def _separate_at_op(self, expr, max_split=None): + + for op, _ in self._all_operators(): + # hackety: </> have higher priority than <</>>, but don't confuse them + skip_delim = (op + op) if op in '<>*?' else None + if op == '?': + skip_delim = (skip_delim, '?.') + separated = list(self._separate(expr, op, skip_delims=skip_delim)) + if len(separated) < 2: + continue + + 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 + separated = [s.strip() for s in separated] + while len(separated) > 1 and not separated[-1]: + undone += 1 + separated.pop() + if op == '-' and undone % 2 != 0: + right_expr = op + right_expr + elif op == '+': + while len(separated) > 1 and set(separated[-1]) <= self.OP_CHARS: + right_expr = separated.pop() + right_expr + if separated[-1][-1:] in self.OP_CHARS: + right_expr = separated.pop() + right_expr + # hanging op at end of left => unary + (strip) or - (push right) + separated.append(right_expr) + dm_ops = ('*', '%', '/', '**') + dm_chars = set(''.join(dm_ops)) + + def yield_terms(s): + skip = False + for i, term in enumerate(s[:-1]): + if skip: + skip = False + continue + if not (dm_chars & set(term)): + yield term + continue + for dm_op in dm_ops: + bodmas = list(self._separate(term, dm_op, skip_delims=skip_delim)) + if len(bodmas) > 1 and not bodmas[-1].strip(): + bodmas[-1] = (op if op == '-' else '') + s[i + 1] + yield dm_op.join(bodmas) + skip = True + break + else: + if term: + yield term + + if not skip and s[-1]: + yield s[-1] + + separated = list(yield_terms(separated)) + right_expr = separated.pop() if len(separated) > 1 else None + expr = op.join(separated) + if right_expr is None: + continue + return op, separated, right_expr + def _operator(self, op, left_val, right_expr, expr, local_vars, allow_recursion): if op in ('||', '&&'): if (op == '&&') ^ _js_ternary(left_val): @@ -573,7 +779,7 @@ class JSInterpreter(object): elif op == '?': right_expr = _js_ternary(left_val, *self._separate(right_expr, ':', 1)) - right_val = self.interpret_expression(right_expr, local_vars, allow_recursion) + right_val = self.interpret_expression(right_expr, local_vars, allow_recursion) if right_expr else left_val opfunc = op and next((v for k, v in self._all_operators() if k == op), None) if not opfunc: return right_val @@ -584,18 +790,21 @@ class JSInterpreter(object): except Exception as e: raise self.Exception('Failed to evaluate {left_val!r:.50} {op} {right_val!r:.50}'.format(**locals()), expr, cause=e) - def _index(self, obj, idx, allow_undefined=True): + def _index(self, obj, idx, allow_undefined=None): if idx == 'length' and isinstance(obj, list): return len(obj) try: return obj[int(idx)] if isinstance(obj, list) else obj[compat_str(idx)] - except (TypeError, KeyError, IndexError) as e: - if allow_undefined: - # when is not allowed? + except (TypeError, KeyError, IndexError, ValueError) as e: + # allow_undefined is None gives correct behaviour + if allow_undefined or ( + allow_undefined is None and not isinstance(e, TypeError)): return JS_Undefined raise self.Exception('Cannot get index {idx!r:.100}'.format(**locals()), expr=repr(obj), cause=e) def _dump(self, obj, namespace): + if obj is JS_Undefined: + return 'undefined' try: return json.dumps(obj) except TypeError: @@ -615,51 +824,9 @@ class JSInterpreter(object): _FINALLY_RE = re.compile(r'finally\s*\{') _SWITCH_RE = re.compile(r'switch\s*\(') - def handle_operators(self, expr, local_vars, allow_recursion): - - for op, _ in self._all_operators(): - # hackety: </> have higher priority than <</>>, but don't confuse them - skip_delim = (op + op) if op in '<>*?' else None - if op == '?': - skip_delim = (skip_delim, '?.') - separated = list(self._separate(expr, op, skip_delims=skip_delim)) - if len(separated) < 2: - continue - - 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 - separated = [s.strip() for s in separated] - while len(separated) > 1 and not separated[-1]: - undone += 1 - separated.pop() - if op == '-' and undone % 2 != 0: - right_expr = op + right_expr - elif op == '+': - while len(separated) > 1 and set(separated[-1]) <= self.OP_CHARS: - right_expr = separated.pop() + right_expr - if separated[-1][-1:] 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] if separated else '' - 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: - continue - - left_val = self.interpret_expression(op.join(separated), local_vars, allow_recursion) - return self._operator(op, left_val, right_expr, expr, local_vars, allow_recursion), True + def _eval_operator(self, op, left_expr, right_expr, expr, local_vars, allow_recursion): + left_val = self.interpret_expression(left_expr, local_vars, allow_recursion) + return self._operator(op, left_val, right_expr, expr, local_vars, allow_recursion) @Debugger.wrap_interpreter def interpret_statement(self, stmt, local_vars, allow_recursion=100): @@ -700,7 +867,7 @@ class JSInterpreter(object): new_kw, _, obj = expr.partition('new ') if not new_kw: - for klass, konstr in (('Date', lambda x: int(unified_timestamp(x, False) * 1000)), + for klass, konstr in (('Date', lambda *x: self.JS_Date(*x).valueOf()), ('RegExp', self.JS_RegExp), ('Error', self.Exception)): if not obj.startswith(klass + '('): @@ -715,15 +882,19 @@ class JSInterpreter(object): else: raise self.Exception('Unsupported object {obj:.100}'.format(**locals()), expr=expr) + # apply unary operators (see new above) for op, _ in _UNARY_OPERATORS_X: if not expr.startswith(op): continue operand = expr[len(op):] - if not operand or operand[0] != ' ': + if not operand or (op.isalpha() and operand[0] != ' '): continue - op_result = self.handle_operators(expr, local_vars, allow_recursion) - if op_result: - return op_result[0], should_return + separated = self._separate_at_op(operand, max_split=1) + if separated: + next_op, separated, right_expr = separated + separated.append(right_expr) + operand = next_op.join(separated) + return self._eval_operator(op, operand, '', expr, local_vars, allow_recursion), should_return if expr.startswith('{'): inner, outer = self._separate_at_paren(expr) @@ -918,15 +1089,18 @@ class JSInterpreter(object): m = re.match(r'''(?x) (?P<assign> - (?P<out>{_NAME_RE})(?:\[(?P<out_idx>(?:.+?\]\s*\[)*.+?)\])?\s* + (?P<out>{_NAME_RE})(?P<out_idx>(?:\[{_NESTED_BRACKETS}\])+)?\s* (?P<op>{_OPERATOR_RE})? =(?!=)(?P<expr>.*)$ )|(?P<return> (?!if|return|true|false|null|undefined|NaN|Infinity)(?P<name>{_NAME_RE})$ - )|(?P<indexing> - (?P<in>{_NAME_RE})\[(?P<in_idx>(?:.+?\]\s*\[)*.+?)\]$ )|(?P<attribute> - (?P<var>{_NAME_RE})(?:(?P<nullish>\?)?\.(?P<member>[^(]+)|\[(?P<member2>[^\]]+)\])\s* + (?P<var>{_NAME_RE})(?: + (?P<nullish>\?)?\.(?P<member>[^(]+)| + \[(?P<member2>{_NESTED_BRACKETS})\] + )\s* + )|(?P<indexing> + (?P<in>{_NAME_RE})(?P<in_idx>\[.+\])$ )|(?P<function> (?P<fname>{_NAME_RE})\((?P<args>.*)\)$ )'''.format(**globals()), expr) @@ -941,13 +1115,18 @@ class JSInterpreter(object): elif left_val in (None, JS_Undefined): raise self.Exception('Cannot index undefined variable ' + m.group('out'), expr=expr) - indexes = re.split(r'\]\s*\[', m.group('out_idx')) - for i, idx in enumerate(indexes, 1): + indexes = md['out_idx'] + while indexes: + idx, indexes = self._separate_at_paren(indexes) idx = self.interpret_expression(idx, local_vars, allow_recursion) - if i < len(indexes): + if indexes: left_val = self._index(left_val, idx) if isinstance(idx, float): idx = int(idx) + if isinstance(left_val, list) and len(left_val) <= int_or_none(idx, default=-1): + # JS Array is a sparsely assignable list + # TODO: handle extreme sparsity without memory bloat, eg using auxiliary dict + left_val.extend((idx - len(left_val) + 1) * [JS_Undefined]) left_val[idx] = self._operator( m.group('op'), self._index(left_val, idx) if m.group('op') else None, m.group('expr'), expr, local_vars, allow_recursion) @@ -985,14 +1164,17 @@ class JSInterpreter(object): if md.get('indexing'): val = local_vars[m.group('in')] - for idx in re.split(r'\]\s*\[', m.group('in_idx')): + indexes = m.group('in_idx') + while indexes: + idx, indexes = self._separate_at_paren(indexes) idx = self.interpret_expression(idx, local_vars, allow_recursion) val = self._index(val, idx) return val, should_return - op_result = self.handle_operators(expr, local_vars, allow_recursion) - if op_result: - return op_result[0], should_return + separated = self._separate_at_op(expr) + if separated: + op, separated, right_expr = separated + return self._eval_operator(op, op.join(separated), right_expr, expr, local_vars, allow_recursion), should_return if md.get('attribute'): variable, member, nullish = m.group('var', 'member', 'nullish') @@ -1013,12 +1195,15 @@ class JSInterpreter(object): def eval_method(variable, member): if (variable, member) == ('console', 'debug'): if Debugger.ENABLED: - Debugger.write(self.interpret_expression('[{}]'.format(arg_str), local_vars, allow_recursion)) + Debugger.write(self.interpret_expression('[{0}]'.format(arg_str), local_vars, allow_recursion)) return types = { 'String': compat_str, 'Math': float, 'Array': list, + 'Date': self.JS_Date, + 'RegExp': self.JS_RegExp, + # 'Error': self.Exception, # has no std static methods } obj = local_vars.get(variable) if obj in (JS_Undefined, None): @@ -1026,7 +1211,7 @@ class JSInterpreter(object): if obj is JS_Undefined: try: if variable not in self._objects: - self._objects[variable] = self.extract_object(variable) + self._objects[variable] = self.extract_object(variable, local_vars) obj = self._objects[variable] except self.Exception: if not nullish: @@ -1037,7 +1222,7 @@ class JSInterpreter(object): # Member access if arg_str is None: - return self._index(obj, member) + return self._index(obj, member, nullish) # Function call argvals = [ @@ -1071,6 +1256,8 @@ class JSInterpreter(object): assertion(len(argvals) == 2, 'takes two arguments') return argvals[0] ** argvals[1] raise self.Exception('Unsupported Math method ' + member, expr=expr) + elif obj is self.JS_Date: + return getattr(obj, member)(*argvals) if member == 'split': assertion(len(argvals) <= 2, 'takes at most two arguments') @@ -1111,9 +1298,10 @@ class JSInterpreter(object): elif member == 'join': assertion(isinstance(obj, list), 'must be applied on a list') assertion(len(argvals) <= 1, 'takes at most one argument') - return (',' if len(argvals) == 0 else argvals[0]).join( - ('' if x in (None, JS_Undefined) else _js_toString(x)) - for x in obj) + return (',' if len(argvals) == 0 or argvals[0] in (None, JS_Undefined) + else argvals[0]).join( + ('' if x in (None, JS_Undefined) else _js_toString(x)) + for x in obj) elif member == 'reverse': assertion(not argvals, 'does not take any arguments') obj.reverse() @@ -1177,7 +1365,8 @@ class JSInterpreter(object): assertion(len(argvals) == 2, 'takes exactly two arguments') # TODO: argvals[1] callable, other Py vs JS edge cases if isinstance(argvals[0], self.JS_RegExp): - count = 0 if argvals[0].flags & self.JS_RegExp.RE_FLAGS['g'] else 1 + # access JS member with Py reserved name + count = 0 if self._index(argvals[0], 'global') else 1 assertion(member != 'replaceAll' or count == 0, 'replaceAll must be called with a global RegExp') return argvals[0].sub(argvals[1], obj, count=count) @@ -1218,7 +1407,7 @@ class JSInterpreter(object): for v in self._separate(list_txt): yield self.interpret_expression(v, local_vars, allow_recursion) - def extract_object(self, objname): + def extract_object(self, objname, *global_stack): _FUNC_NAME_RE = r'''(?:{n}|"{n}"|'{n}')'''.format(n=_NAME_RE) obj = {} fields = next(filter(None, ( @@ -1239,7 +1428,8 @@ class JSInterpreter(object): fields): argnames = self.build_arglist(f.group('args')) name = remove_quotes(f.group('key')) - obj[name] = function_with_repr(self.build_function(argnames, f.group('code')), 'F<{0}>'.format(name)) + obj[name] = function_with_repr( + self.build_function(argnames, f.group('code'), *global_stack), 'F<{0}>'.format(name)) return obj @@ -1271,19 +1461,21 @@ class JSInterpreter(object): code, _ = self._separate_at_paren(func_m.group('code')) # refine the match return self.build_arglist(func_m.group('args')), code - def extract_function(self, funcname): + def extract_function(self, funcname, *global_stack): return function_with_repr( - self.extract_function_from_code(*self.extract_function_code(funcname)), + self.extract_function_from_code(*itertools.chain( + self.extract_function_code(funcname), global_stack)), 'F<%s>' % (funcname,)) def extract_function_from_code(self, argnames, code, *global_stack): local_vars = {} + start = None while True: - mobj = re.search(r'function\((?P<args>[^)]*)\)\s*{', code) + mobj = re.search(r'function\((?P<args>[^)]*)\)\s*{', code[start:]) if mobj is None: break - start, body_start = mobj.span() + start, body_start = ((start or 0) + x for x in mobj.span()) body, remaining = self._separate_at_paren(code[body_start - 1:]) name = self._named_object(local_vars, self.extract_function_from_code( [x.strip() for x in mobj.group('args').split(',')], diff --git a/youtube_dl/options.py b/youtube_dl/options.py index 61705d1f0..ce3633c41 100644 --- a/youtube_dl/options.py +++ b/youtube_dl/options.py @@ -405,6 +405,10 @@ def parseOpts(overrideArguments=None): action='store_true', dest='listformats', help='List all available formats of requested videos') video_format.add_option( + '--no-list-formats', + action='store_false', dest='listformats', + help='Do not list available formats of requested videos (default)') + video_format.add_option( '--youtube-include-dash-manifest', action='store_true', dest='youtube_include_dash_manifest', default=True, help=optparse.SUPPRESS_HELP) @@ -413,6 +417,17 @@ def parseOpts(overrideArguments=None): action='store_false', dest='youtube_include_dash_manifest', help='Do not download the DASH manifests and related data on YouTube videos') video_format.add_option( + '--youtube-player-js-variant', + action='store', dest='youtube_player_js_variant', + help='For YouTube, the player javascript variant to use for n/sig deciphering; `actual` to follow the site; default `%default`.', + choices=('actual', 'main', 'tcc', 'tce', 'es5', 'es6', 'tv', 'tv_es6', 'phone', 'tablet'), + default='main', metavar='VARIANT') + video_format.add_option( + '--youtube-player-js-version', + action='store', dest='youtube_player_js_version', + help='For YouTube, the player javascript version to use for n/sig deciphering, specified as `signature_timestamp@hash`, or `actual` to follow the site; default `%default`', + default='20348@0004de42', metavar='STS@HASH') + video_format.add_option( '--merge-output-format', action='store', dest='merge_output_format', metavar='FORMAT', default=None, help=( diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py index ac1e78002..c4262936e 100644 --- a/youtube_dl/utils.py +++ b/youtube_dl/utils.py @@ -4204,12 +4204,16 @@ def lowercase_escape(s): s) -def escape_rfc3986(s): +def escape_rfc3986(s, safe=None): """Escape non-ASCII characters as suggested by RFC 3986""" if sys.version_info < (3, 0): s = _encode_compat_str(s, 'utf-8') + if safe is not None: + safe = _encode_compat_str(safe, 'utf-8') + if safe is None: + safe = b"%/;:@&=+$,!~*'()?#[]" # ensure unicode: after quoting, it can always be converted - return compat_str(compat_urllib_parse.quote(s, b"%/;:@&=+$,!~*'()?#[]")) + return compat_str(compat_urllib_parse.quote(s, safe)) def escape_url(url): diff --git a/youtube_dl/version.py b/youtube_dl/version.py index b82fbc702..c70d9d2af 100644 --- a/youtube_dl/version.py +++ b/youtube_dl/version.py @@ -1,3 +1,3 @@ from __future__ import unicode_literals -__version__ = '2021.12.17' +__version__ = '2025.04.07' |