diff options
Diffstat (limited to 'yt_dlp/YoutubeDL.py')
-rw-r--r-- | yt_dlp/YoutubeDL.py | 729 |
1 files changed, 421 insertions, 308 deletions
diff --git a/yt_dlp/YoutubeDL.py b/yt_dlp/YoutubeDL.py index de8a8c4d2..9125e4ea6 100644 --- a/yt_dlp/YoutubeDL.py +++ b/yt_dlp/YoutubeDL.py @@ -10,7 +10,6 @@ import json import locale import operator import os -import platform import random import re import shutil @@ -25,12 +24,12 @@ import urllib.request from string import ascii_letters from .cache import Cache -from .compat import HAS_LEGACY as compat_has_legacy from .compat import compat_os_name, compat_shlex_quote from .cookies import load_cookies from .downloader import FFmpegFD, get_suitable_downloader, shorten_protocol_name from .downloader.rtmp import rtmpdump_version from .extractor import gen_extractor_classes, get_info_extractor +from .extractor.common import UnsupportedURLIE from .extractor.openload import PhantomJSwrapper from .minicurses import format_text from .postprocessor import _PLUGIN_CLASSES as plugin_postprocessors @@ -44,6 +43,7 @@ from .postprocessor import ( FFmpegFixupTimestampPP, FFmpegMergerPP, FFmpegPostProcessor, + FFmpegVideoConvertorPP, MoveFilesAfterDownloadPP, get_postprocessor, ) @@ -51,6 +51,7 @@ from .utils import ( DEFAULT_OUTTMPL, IDENTITY, LINK_TEMPLATES, + MEDIA_EXTENSIONS, NO_DEFAULT, NUMBER_RE, OUTTMPL_TYPES, @@ -64,6 +65,7 @@ from .utils import ( EntryNotInPlaylist, ExistingVideoReached, ExtractorError, + FormatSorter, GeoRestrictedError, HEADRequest, ISO3166Utils, @@ -79,17 +81,21 @@ from .utils import ( RejectedVideoReached, SameFileError, UnavailableVideoError, + UserNotLive, YoutubeDLCookieProcessor, YoutubeDLHandler, YoutubeDLRedirectHandler, age_restricted, args_to_str, + bug_reports_message, date_from_str, + deprecation_warning, determine_ext, determine_protocol, encode_compat_str, encodeFilename, error_to_compat_str, + escapeHTML, expand_path, filter_dict, float_or_none, @@ -97,19 +103,22 @@ from .utils import ( format_decimal_suffix, format_field, formatSeconds, + get_compatible_ext, get_domain, int_or_none, iri_to_uri, + is_path_like, join_nonempty, locked_file, + make_archive_id, make_dir, make_HTTPS_handler, merge_headers, network_exceptions, number_of_digits, orderedSet, + orderedSet_from_options, parse_filesize, - platform_name, preferredencoding, prepend_extension, register_socks_protocols, @@ -125,9 +134,11 @@ from .utils import ( strftime_or_none, subtitles_filename, supports_terminal_sequences, + system_identifier, timetuple_from_msec, to_high_limit_path, traverse_obj, + try_call, try_get, url_basename, variadic, @@ -136,7 +147,7 @@ from .utils import ( write_json_file, write_string, ) -from .version import RELEASE_GIT_HEAD, __version__ +from .version import RELEASE_GIT_HEAD, VARIANT, __version__ if compat_os_name == 'nt': import ctypes @@ -228,7 +239,7 @@ class YoutubeDL: Default is 'only_download' for CLI, but False for API skip_playlist_after_errors: Number of allowed failures until the rest of the playlist is skipped - force_generic_extractor: Force downloader to use the generic extractor + allowed_extractors: List of regexes to match against extractor names that are allowed overwrites: Overwrite all video and metadata files if True, overwrite only non-video files if None and don't overwrite any file if False @@ -240,8 +251,8 @@ class YoutubeDL: matchtitle: Download only matching titles. rejecttitle: Reject downloads for matching titles. logger: Log messages to a logging.Logger instance. - logtostderr: Log messages to stderr instead of stdout. - consoletitle: Display progress in console window's titlebar. + logtostderr: Print everything to stderr instead of stdout. + consoletitle: Display progress in console window's titlebar. writedescription: Write the video description to a .description file writeinfojson: Write the video description to a .info.json file clean_infojson: Remove private fields from the infojson @@ -264,7 +275,7 @@ class YoutubeDL: subtitleslangs: List of languages of the subtitles to download (can be regex). The list may contain "all" to refer to all the available subtitles. The language can be prefixed with a "-" to - exclude it from the requested languages. Eg: ['all', '-live_chat'] + exclude it from the requested languages, e.g. ['all', '-live_chat'] keepvideo: Keep the video file after post-processing daterange: A DateRange object, download only if the upload_date is in the range. skip_download: Skip the actual download of the video file @@ -282,9 +293,8 @@ class YoutubeDL: downloaded. Videos without view count information are always downloaded. None for no limit. - download_archive: File name of a file where all downloads are recorded. - Videos already present in the file are not downloaded - again. + download_archive: A set, or the name of a file where all downloads are recorded. + Videos already present in the file are not downloaded again. break_on_existing: Stop the download process after attempting to download a file that is in the archive. break_on_reject: Stop the download process when encountering a video that @@ -293,8 +303,9 @@ class YoutubeDL: should act on each input URL as opposed to for the entire queue cookiefile: File name or text stream from where cookies should be read and dumped to cookiesfrombrowser: A tuple containing the name of the browser, the profile - name/pathfrom where cookies are loaded, and the name of the - keyring. Eg: ('chrome', ) or ('vivaldi', 'default', 'BASICTEXT') + name/path from where cookies are loaded, the name of the keyring, + and the container name, e.g. ('chrome', ) or + ('vivaldi', 'default', 'BASICTEXT') or ('firefox', 'default', None, 'Meta') legacyserverconnect: Explicitly allow HTTPS connection to servers that do not support RFC 5746 secure renegotiation nocheckcertificate: Do not verify SSL certificates @@ -303,7 +314,7 @@ class YoutubeDL: client_certificate_password: Password for client certificate private key, if encrypted. If not provided and the key is encrypted, yt-dlp will ask interactively prefer_insecure: Use HTTP instead of HTTPS to retrieve information. - At the moment, this is only supported by YouTube. + (Only supported by some extractors) http_headers: A dictionary of custom headers to be used for all requests proxy: URL of the proxy server to use geo_verification_proxy: URL of the proxy to use for IP address verification @@ -315,9 +326,14 @@ class YoutubeDL: default_search: Prepend this string if an input url is not valid. 'auto' for elaborate guessing encoding: Use this encoding instead of the system-specified. - extract_flat: Do not resolve URLs, return the immediate result. - Pass in 'in_playlist' to only show this behavior for - playlist items. + extract_flat: Whether to resolve and process url_results further + * False: Always process (default) + * True: Never process + * 'in_playlist': Do not process inside playlist/multi_video + * 'discard': Always process, but don't return the result + from inside playlist/multi_video + * 'discard_in_playlist': Same as "discard", but only for + playlists (not multi_video) wait_for_video: If given, wait for scheduled streams to become available. The value should be a tuple containing the range (min_secs, max_secs) to wait between retries @@ -361,7 +377,7 @@ class YoutubeDL: Progress hooks are guaranteed to be called at least twice (with status "started" and "finished") if the processing is successful. - merge_output_format: Extension to use when merging formats. + merge_output_format: "/" separated list of extensions to use when merging formats. final_ext: Expected final extension; used to detect when the file was already downloaded and converted fixup: Automatically correct known faults of the file. @@ -421,19 +437,23 @@ class YoutubeDL: retry_sleep_functions: Dictionary of functions that takes the number of attempts as argument and returns the time to sleep in seconds. Allowed keys are 'http', 'fragment', 'file_access' - download_ranges: A function that gets called for every video with the signature - (info_dict, *, ydl) -> Iterable[Section]. - Only the returned sections will be downloaded. Each Section contains: + download_ranges: A callback function that gets called for every video with + the signature (info_dict, ydl) -> Iterable[Section]. + Only the returned sections will be downloaded. + Each Section is a dict with the following keys: * start_time: Start time of the section in seconds * end_time: End time of the section in seconds * title: Section title (Optional) * index: Section number (Optional) + force_keyframes_at_cuts: Re-encode the video when downloading ranges to get precise cuts + noprogress: Do not print the progress bar + live_from_start: Whether to download livestreams videos from the start The following parameters are not used by YoutubeDL itself, they are used by the downloader (see yt_dlp/downloader/common.py): nopart, updatetime, buffersize, ratelimit, throttledratelimit, min_filesize, max_filesize, test, noresizebuffer, retries, file_access_retries, fragment_retries, - continuedl, noprogress, xattr_set_filesize, hls_use_mpegts, http_chunk_size, + continuedl, xattr_set_filesize, hls_use_mpegts, http_chunk_size, external_downloader_args, concurrent_fragment_downloads. The following options are used by the post processors: @@ -454,11 +474,13 @@ class YoutubeDL: discontinuities such as ad breaks (default: False) extractor_args: A dictionary of arguments to be passed to the extractors. See "EXTRACTOR ARGUMENTS" for details. - Eg: {'youtube': {'skip': ['dash', 'hls']}} + E.g. {'youtube': {'skip': ['dash', 'hls']}} mark_watched: Mark videos watched (even with --simulate). Only for YouTube The following options are deprecated and may be removed in the future: + force_generic_extractor: Force downloader to use the generic extractor + - Use allowed_extractors = ['generic', 'default'] playliststart: - Use playlist_items Playlist item to start at. playlistend: - Use playlist_items @@ -511,7 +533,8 @@ class YoutubeDL: """ _NUMERIC_FIELDS = { - 'width', 'height', 'tbr', 'abr', 'asr', 'vbr', 'fps', 'filesize', 'filesize_approx', + 'width', 'height', 'asr', 'audio_channels', 'fps', + 'tbr', 'abr', 'vbr', 'filesize', 'filesize_approx', 'timestamp', 'release_timestamp', 'duration', 'view_count', 'like_count', 'dislike_count', 'repost_count', 'average_rating', 'comment_count', 'age_limit', @@ -523,17 +546,17 @@ class YoutubeDL: _format_fields = { # NB: Keep in sync with the docstring of extractor/common.py 'url', 'manifest_url', 'manifest_stream_number', 'ext', 'format', 'format_id', 'format_note', - 'width', 'height', 'resolution', 'dynamic_range', 'tbr', 'abr', 'acodec', 'asr', - 'vbr', 'fps', 'vcodec', 'container', 'filesize', 'filesize_approx', + 'width', 'height', 'aspect_ratio', 'resolution', 'dynamic_range', 'tbr', 'abr', 'acodec', 'asr', 'audio_channels', + 'vbr', 'fps', 'vcodec', 'container', 'filesize', 'filesize_approx', 'rows', 'columns', 'player_url', 'protocol', 'fragment_base_url', 'fragments', 'is_from_start', 'preference', 'language', 'language_preference', 'quality', 'source_preference', 'http_headers', 'stretched_ratio', 'no_resume', 'has_drm', 'downloader_options', 'page_url', 'app', 'play_path', 'tc_url', 'flash_version', 'rtmp_live', 'rtmp_conn', 'rtmp_protocol', 'rtmp_real_time' } _format_selection_exts = { - 'audio': {'m4a', 'mp3', 'ogg', 'aac'}, - 'video': {'mp4', 'flv', 'webm', '3gp'}, - 'storyboards': {'mhtml'}, + 'audio': set(MEDIA_EXTENSIONS.common_audio), + 'video': set(MEDIA_EXTENSIONS.common_video + ('3gp', )), + 'storyboards': set(MEDIA_EXTENSIONS.storyboards), } def __init__(self, params=None, auto_init=True): @@ -573,10 +596,13 @@ class YoutubeDL: for type_, stream in self._out_files.items_ if type_ != 'console' }) - MIN_SUPPORTED, MIN_RECOMMENDED = (3, 6), (3, 7) + # The code is left like this to be reused for future deprecations + MIN_SUPPORTED, MIN_RECOMMENDED = (3, 7), (3, 7) current_version = sys.version_info[:2] if current_version < MIN_RECOMMENDED: - msg = 'Support for Python version %d.%d has been deprecated and will break in future versions of yt-dlp' + msg = ('Support for Python version %d.%d has been deprecated. ' + 'See https://github.com/yt-dlp/yt-dlp/issues/3764 for more details.' + '\n You will no longer receive updates on this version') if current_version < MIN_SUPPORTED: msg = 'Python version %d.%d is no longer supported' self.deprecation_warning( @@ -589,6 +615,30 @@ class YoutubeDL: ' If you experience any issues while using this option, ' f'{self._format_err("DO NOT", self.Styles.ERROR)} open a bug report') + if self.params.get('bidi_workaround', False): + try: + import pty + master, slave = pty.openpty() + width = shutil.get_terminal_size().columns + width_args = [] if width is None else ['-w', str(width)] + sp_kwargs = {'stdin': subprocess.PIPE, 'stdout': slave, 'stderr': self._out_files.error} + try: + self._output_process = Popen(['bidiv'] + width_args, **sp_kwargs) + except OSError: + self._output_process = Popen(['fribidi', '-c', 'UTF-8'] + width_args, **sp_kwargs) + self._output_channel = os.fdopen(master, 'rb') + except OSError as ose: + if ose.errno == errno.ENOENT: + self.report_warning( + 'Could not find fribidi executable, ignoring --bidi-workaround. ' + 'Make sure that fribidi is an executable file in one of the directories in your $PATH.') + else: + raise + + self.params['compat_opts'] = set(self.params.get('compat_opts', ())) + if auto_init and auto_init != 'no_verbose_header': + self.print_debug_header() + def check_deprecated(param, option, suggestion): if self.params.get(param) is not None: self.report_warning(f'{option} is deprecated. Use {suggestion} instead') @@ -606,11 +656,8 @@ class YoutubeDL: for msg in self.params.get('_warnings', []): self.report_warning(msg) for msg in self.params.get('_deprecation_warnings', []): - self.deprecation_warning(msg) + self.deprecated_feature(msg) - self.params['compat_opts'] = set(self.params.get('compat_opts', ())) - if not compat_has_legacy: - self.params['compat_opts'].add('no-compat-legacy') if 'list-formats' in self.params['compat_opts']: self.params['listformats_table'] = False @@ -624,6 +671,13 @@ class YoutubeDL: else: self.params['nooverwrites'] = not self.params['overwrites'] + if self.params.get('simulate') is None and any(( + self.params.get('list_thumbnails'), + self.params.get('listformats'), + self.params.get('listsubtitles'), + )): + self.params['simulate'] = 'list_only' + self.params.setdefault('forceprint', {}) self.params.setdefault('print_to_file', {}) @@ -631,29 +685,7 @@ class YoutubeDL: if not isinstance(params['forceprint'], dict): self.params['forceprint'] = {'video': params['forceprint']} - if self.params.get('bidi_workaround', False): - try: - import pty - master, slave = pty.openpty() - width = shutil.get_terminal_size().columns - width_args = [] if width is None else ['-w', str(width)] - sp_kwargs = {'stdin': subprocess.PIPE, 'stdout': slave, 'stderr': self._out_files.error} - try: - self._output_process = Popen(['bidiv'] + width_args, **sp_kwargs) - except OSError: - self._output_process = Popen(['fribidi', '-c', 'UTF-8'] + width_args, **sp_kwargs) - self._output_channel = os.fdopen(master, 'rb') - except OSError as ose: - if ose.errno == errno.ENOENT: - self.report_warning( - 'Could not find fribidi executable, ignoring --bidi-workaround. ' - 'Make sure that fribidi is an executable file in one of the directories in your $PATH.') - else: - raise - if auto_init: - if auto_init != 'no_verbose_header': - self.print_debug_header() self.add_default_info_extractors() if (sys.platform != 'win32' @@ -698,21 +730,23 @@ class YoutubeDL: def preload_download_archive(fn): """Preload the archive, if any is specified""" + archive = set() if fn is None: - return False + return archive + elif not is_path_like(fn): + return fn + self.write_debug(f'Loading archive file {fn!r}') try: with locked_file(fn, 'r', encoding='utf-8') as archive_file: for line in archive_file: - self.archive.add(line.strip()) + archive.add(line.strip()) except OSError as ioe: if ioe.errno != errno.ENOENT: raise - return False - return True + return archive - self.archive = set() - preload_download_archive(self.params.get('download_archive')) + self.archive = preload_download_archive(self.params.get('download_archive')) def warn_if_short_id(self, argv): # short YouTube ID starting with dash? @@ -738,13 +772,6 @@ class YoutubeDL: self._ies_instances[ie_key] = ie ie.set_downloader(self) - def _get_info_extractor_class(self, ie_key): - ie = self._ies.get(ie_key) - if ie is None: - ie = get_info_extractor(ie_key) - self.add_info_extractor(ie) - return ie - def get_info_extractor(self, ie_key): """ Get an instance of an IE with name ie_key, it will try to get one from @@ -761,8 +788,19 @@ class YoutubeDL: """ Add the InfoExtractors returned by gen_extractors to the end of the list """ - for ie in gen_extractor_classes(): - self.add_info_extractor(ie) + all_ies = {ie.IE_NAME.lower(): ie for ie in gen_extractor_classes()} + all_ies['end'] = UnsupportedURLIE() + try: + ie_names = orderedSet_from_options( + self.params.get('allowed_extractors', ['default']), { + 'all': list(all_ies), + 'default': [name for name, ie in all_ies.items() if ie._ENABLED], + }, use_regex=True) + except re.error as e: + raise ValueError(f'Wrong regex for allowed_extractors: {e.pattern}') + for name in ie_names: + self.add_info_extractor(all_ies[name]) + self.write_debug(f'Loaded {len(ie_names)} extractors') def add_post_processor(self, pp, when='post_process'): """Add a PostProcessor object to the end of the chain.""" @@ -808,12 +846,14 @@ class YoutubeDL: def to_stdout(self, message, skip_eol=False, quiet=None): """Print message to stdout""" if quiet is not None: - self.deprecation_warning('"YoutubeDL.to_stdout" no longer accepts the argument quiet. Use "YoutubeDL.to_screen" instead') + self.deprecation_warning('"YoutubeDL.to_stdout" no longer accepts the argument quiet. ' + 'Use "YoutubeDL.to_screen" instead') if skip_eol is not False: - self.deprecation_warning('"YoutubeDL.to_stdout" no longer accepts the argument skip_eol. Use "YoutubeDL.to_screen" instead') + self.deprecation_warning('"YoutubeDL.to_stdout" no longer accepts the argument skip_eol. ' + 'Use "YoutubeDL.to_screen" instead') self._write_string(f'{self._bidi_workaround(message)}\n', self._out_files.out) - def to_screen(self, message, skip_eol=False, quiet=None): + def to_screen(self, message, skip_eol=False, quiet=None, only_once=False): """Print message to screen if not in quiet mode""" if self.params.get('logger'): self.params['logger'].debug(message) @@ -822,7 +862,7 @@ class YoutubeDL: return self._write_string( '%s%s' % (self._bidi_workaround(message), ('' if skip_eol else '\n')), - self._out_files.screen) + self._out_files.screen, only_once=only_once) def to_stderr(self, message, only_once=False): """Print message to stderr""" @@ -946,11 +986,14 @@ class YoutubeDL: return self.to_stderr(f'{self._format_err("WARNING:", self.Styles.WARNING)} {message}', only_once) - def deprecation_warning(self, message): + def deprecation_warning(self, message, *, stacklevel=0): + deprecation_warning( + message, stacklevel=stacklevel + 1, printer=self.report_error, is_error=False) + + def deprecated_feature(self, message): if self.params.get('logger') is not None: - self.params['logger'].warning(f'DeprecationWarning: {message}') - else: - self.to_stderr(f'{self._format_err("DeprecationWarning:", self.Styles.ERROR)} {message}', True) + self.params['logger'].warning(f'Deprecated Feature: {message}') + self.to_stderr(f'{self._format_err("Deprecated Feature:", self.Styles.ERROR)} {message}', True) def report_error(self, message, *args, **kwargs): ''' @@ -1010,7 +1053,7 @@ class YoutubeDL: def get_output_path(self, dir_type='', filename=None): paths = self.params.get('paths', {}) - assert isinstance(paths, dict) + assert isinstance(paths, dict), '"paths" parameter must be a dictionary' path = os.path.join( expand_path(paths.get('home', '').strip()), expand_path(paths.get(dir_type, '').strip()) if dir_type else '', @@ -1028,7 +1071,7 @@ class YoutubeDL: # outtmpl should be expand_path'ed before template dict substitution # because meta fields may contain env variables we don't want to - # be expanded. For example, for outtmpl "%(title)s.%(ext)s" and + # be expanded. E.g. for outtmpl "%(title)s.%(ext)s" and # title "Hello $PATH", we don't want `$PATH` to be expanded. return expand_path(outtmpl).replace(sep, '') @@ -1044,7 +1087,7 @@ class YoutubeDL: def validate_outtmpl(cls, outtmpl): ''' @return None or Exception object ''' outtmpl = re.sub( - STR_FORMAT_RE_TMPL.format('[^)]*', '[ljqBUDS]'), + STR_FORMAT_RE_TMPL.format('[^)]*', '[ljhqBUDS]'), lambda mobj: f'{mobj.group(0)[:-1]}s', cls._outtmpl_expandpath(outtmpl)) try: @@ -1087,14 +1130,18 @@ class YoutubeDL: } TMPL_DICT = {} - EXTERNAL_FORMAT_RE = re.compile(STR_FORMAT_RE_TMPL.format('[^)]*', f'[{STR_FORMAT_TYPES}ljqBUDS]')) + EXTERNAL_FORMAT_RE = re.compile(STR_FORMAT_RE_TMPL.format('[^)]*', f'[{STR_FORMAT_TYPES}ljhqBUDS]')) MATH_FUNCTIONS = { '+': float.__add__, '-': float.__sub__, } # Field is of the form key1.key2... - # where keys (except first) can be string, int or slice - FIELD_RE = r'\w*(?:\.(?:\w+|{num}|{num}?(?::{num}?){{1,2}}))*'.format(num=r'(?:-?\d+)') + # where keys (except first) can be string, int, slice or "{field, ...}" + FIELD_INNER_RE = r'(?:\w+|%(num)s|%(num)s?(?::%(num)s?){1,2})' % {'num': r'(?:-?\d+)'} + FIELD_RE = r'\w*(?:\.(?:%(inner)s|{%(field)s(?:,%(field)s)*}))*' % { + 'inner': FIELD_INNER_RE, + 'field': rf'\w*(?:\.{FIELD_INNER_RE})*' + } MATH_FIELD_RE = rf'(?:{FIELD_RE}|-?{NUMBER_RE})' MATH_OPERATORS_RE = r'(?:%s)' % '|'.join(map(re.escape, MATH_FUNCTIONS.keys())) INTERNAL_FORMAT_RE = re.compile(rf'''(?x) @@ -1108,11 +1155,20 @@ class YoutubeDL: (?:\|(?P<default>.*?))? )$''') - def _traverse_infodict(k): - k = k.split('.') - if k[0] == '': - k.pop(0) - return traverse_obj(info_dict, k, is_user_input=True, traverse_string=True) + def _traverse_infodict(fields): + fields = [f for x in re.split(r'\.({.+?})\.?', fields) + for f in ([x] if x.startswith('{') else x.split('.'))] + for i in (0, -1): + if fields and not fields[i]: + fields.pop(i) + + for i, f in enumerate(fields): + if not f.startswith('{'): + continue + assert f.endswith('}'), f'No closing brace for {f} in {fields}' + fields[i] = {k: k.split('.') for k in f[1:-1].split(',')} + + return traverse_obj(info_dict, fields, is_user_input=True, traverse_string=True) def get_value(mdict): # Object traversal @@ -1148,6 +1204,9 @@ class YoutubeDL: if mdict['strf_format']: value = strftime_or_none(value, mdict['strf_format'].replace('\\,', ',')) + # XXX: Workaround for https://github.com/yt-dlp/yt-dlp/issues/4485 + if sanitize and value == '': + value = None return value na = self.params.get('outtmpl_na_placeholder', 'NA') @@ -1195,7 +1254,11 @@ class YoutubeDL: delim = '\n' if '#' in flags else ', ' value, fmt = delim.join(map(str, variadic(value, allowed_types=(str, bytes)))), str_fmt elif fmt[-1] == 'j': # json - value, fmt = json.dumps(value, default=_dumpjson_default, indent=4 if '#' in flags else None), str_fmt + value, fmt = json.dumps( + value, default=_dumpjson_default, + indent=4 if '#' in flags else None, ensure_ascii='+' not in flags), str_fmt + elif fmt[-1] == 'h': # html + value, fmt = escapeHTML(str(value)), str_fmt elif fmt[-1] == 'q': # quoted value = map(str, variadic(value) if '#' in flags else [value]) value, fmt = ' '.join(map(compat_shlex_quote, value)), str_fmt @@ -1293,11 +1356,19 @@ class YoutubeDL: return self.get_output_path(dir_type, filename) def _match_entry(self, info_dict, incomplete=False, silent=False): - """ Returns None if the file should be downloaded """ + """Returns None if the file should be downloaded""" + _type = info_dict.get('_type', 'video') + assert incomplete or _type == 'video', 'Only video result can be considered complete' - video_title = info_dict.get('title', info_dict.get('id', 'video')) + video_title = info_dict.get('title', info_dict.get('id', 'entry')) def check_filter(): + if _type in ('playlist', 'multi_video'): + return + elif _type in ('url', 'url_transparent') and not try_call( + lambda: self.get_info_extractor(info_dict['ie_key']).is_single_video(info_dict['url'])): + return + if 'title' in info_dict: # This can happen when we're just evaluating the playlist title = info_dict['title'] @@ -1309,6 +1380,7 @@ class YoutubeDL: if rejecttitle: if re.search(rejecttitle, title, re.IGNORECASE): return '"' + title + '" title matched reject pattern "' + rejecttitle + '"' + date = info_dict.get('upload_date') if date is not None: dateRange = self.params.get('daterange', DateRange()) @@ -1367,18 +1439,19 @@ class YoutubeDL: def extract_info(self, url, download=True, ie_key=None, extra_info=None, process=True, force_generic_extractor=False): """ - Return a list with a dictionary for each video extracted. + Extract and return the information dictionary of the URL Arguments: - url -- URL to extract + @param url URL to extract Keyword arguments: - download -- whether to download videos during extraction - ie_key -- extractor key hint - extra_info -- dictionary containing the extra values to add to each result - process -- whether to resolve all unresolved references (URLs, playlist items), - must be True for download to work. - force_generic_extractor -- force using the generic extractor + @param download Whether to download videos + @param process Whether to resolve all unresolved references (URLs, playlist items). + Must be True for download to work + @param ie_key Use only the extractor with this key + + @param extra_info Dictionary containing the extra values to add to the info (For internal use only) + @force_generic_extractor Force using the generic extractor (Deprecated; use ie_key='Generic') """ if extra_info is None: @@ -1388,11 +1461,11 @@ class YoutubeDL: ie_key = 'Generic' if ie_key: - ies = {ie_key: self._get_info_extractor_class(ie_key)} + ies = {ie_key: self._ies[ie_key]} if ie_key in self._ies else {} else: ies = self._ies - for ie_key, ie in ies.items(): + for key, ie in ies.items(): if not ie.suitable(url): continue @@ -1401,14 +1474,16 @@ class YoutubeDL: 'and will probably not work.') temp_id = ie.get_temp_id(url) - if temp_id is not None and self.in_download_archive({'id': temp_id, 'ie_key': ie_key}): - self.to_screen(f'[{ie_key}] {temp_id}: has already been recorded in the archive') + if temp_id is not None and self.in_download_archive({'id': temp_id, 'ie_key': key}): + self.to_screen(f'[{key}] {temp_id}: has already been recorded in the archive') if self.params.get('break_on_existing', False): raise ExistingVideoReached() break - return self.__extract_info(url, self.get_info_extractor(ie_key), download, extra_info, process) + return self.__extract_info(url, self.get_info_extractor(key), download, extra_info, process) else: - self.report_error('no suitable InfoExtractor for URL %s' % url) + extractors_restricted = self.params.get('allowed_extractors') not in (None, ['default']) + self.report_error(f'No suitable extractor{format_field(ie_key, None, " (%s)")} found for URL {url}', + tb=False if extractors_restricted else None) def _handle_extraction_exceptions(func): @functools.wraps(func) @@ -1442,7 +1517,7 @@ class YoutubeDL: break return wrapper - def _wait_for_video(self, ie_result): + def _wait_for_video(self, ie_result={}): if (not self.params.get('wait_for_video') or ie_result.get('_type', 'video') != 'video' or ie_result.get('formats') or ie_result.get('url')): @@ -1453,7 +1528,12 @@ class YoutubeDL: def progress(msg): nonlocal last_msg - self.to_screen(msg + ' ' * (len(last_msg) - len(msg)) + '\r', skip_eol=True) + full_msg = f'{msg}\n' + if not self.params.get('noprogress'): + full_msg = msg + ' ' * (len(last_msg) - len(msg)) + '\r' + elif last_msg: + return + self.to_screen(full_msg, skip_eol=True) last_msg = msg min_wait, max_wait = self.params.get('wait_for_video') @@ -1461,7 +1541,7 @@ class YoutubeDL: if diff is None and ie_result.get('live_status') == 'is_upcoming': diff = round(random.uniform(min_wait, max_wait) if (max_wait and min_wait) else (max_wait or min_wait), 0) self.report_warning('Release time of video is not known') - elif (diff or 0) <= 0: + elif ie_result and (diff or 0) <= 0: self.report_warning('Video should already be available according to extracted info') diff = min(max(diff or 0, min_wait or 0), max_wait or float('inf')) self.to_screen(f'[wait] Waiting for {format_dur(diff)} - Press Ctrl+C to try now') @@ -1485,8 +1565,16 @@ class YoutubeDL: @_handle_extraction_exceptions def __extract_info(self, url, ie, download, extra_info, process): - ie_result = ie.extract(url) + try: + ie_result = ie.extract(url) + except UserNotLive as e: + if process: + if self.params.get('wait_for_video'): + self.report_warning(e) + self._wait_for_video() + raise if ie_result is None: # Finished already (backwards compatibility; listformats and friends should be moved here) + self.report_warning(f'Extractor {ie.IE_NAME} returned nothing{bug_reports_message()}') return if isinstance(ie_result, list): # Backwards compatibility: old IE result format @@ -1534,7 +1622,8 @@ class YoutubeDL: result_type = ie_result.get('_type', 'video') if result_type in ('url', 'url_transparent'): - ie_result['url'] = sanitize_url(ie_result['url']) + ie_result['url'] = sanitize_url( + ie_result['url'], scheme='http' if self.params.get('prefer_insecure') else 'https') if ie_result.get('original_url'): extra_info.setdefault('original_url', ie_result['original_url']) @@ -1548,6 +1637,7 @@ class YoutubeDL: self.add_default_extra_info(info_copy, ie, ie_result['url']) self.add_extra_info(info_copy, extra_info) info_copy, _ = self.pre_process(info_copy) + self._fill_common_fields(info_copy, False) self.__forced_printings(info_copy, self.prepare_filename(info_copy), incomplete=True) self._raise_pending_errors(info_copy) if self.params.get('force_write_download_archive', False): @@ -1614,8 +1704,8 @@ class YoutubeDL: elif result_type in ('playlist', 'multi_video'): # Protect from infinite recursion due to recursively nested playlists # (see https://github.com/ytdl-org/youtube-dl/issues/27833) - webpage_url = ie_result['webpage_url'] - if webpage_url in self._playlist_urls: + webpage_url = ie_result.get('webpage_url') # Playlists maynot have webpage_url + if webpage_url and webpage_url in self._playlist_urls: self.to_screen( '[download] Skipping already downloaded playlist: %s' % ie_result.get('title') or ie_result.get('id')) @@ -1657,22 +1747,41 @@ class YoutubeDL: return make_dir(path, self.report_error) @staticmethod - def _playlist_infodict(ie_result, **kwargs): - return { - **ie_result, + def _playlist_infodict(ie_result, strict=False, **kwargs): + info = { + 'playlist_count': ie_result.get('playlist_count'), 'playlist': ie_result.get('title') or ie_result.get('id'), 'playlist_id': ie_result.get('id'), 'playlist_title': ie_result.get('title'), 'playlist_uploader': ie_result.get('uploader'), 'playlist_uploader_id': ie_result.get('uploader_id'), - 'playlist_index': 0, **kwargs, } + if strict: + return info + if ie_result.get('webpage_url'): + info.update({ + 'webpage_url': ie_result['webpage_url'], + 'webpage_url_basename': url_basename(ie_result['webpage_url']), + 'webpage_url_domain': get_domain(ie_result['webpage_url']), + }) + return { + **info, + 'playlist_index': 0, + '__last_playlist_index': max(ie_result['requested_entries'] or (0, 0)), + 'extractor': ie_result['extractor'], + 'extractor_key': ie_result['extractor_key'], + } def __process_playlist(self, ie_result, download): """Process each entry in the playlist""" - title = ie_result.get('title') or ie_result.get('id') or '<Untitled>' - self.to_screen(f'[download] Downloading playlist: {title}') + assert ie_result['_type'] in ('playlist', 'multi_video') + + common_info = self._playlist_infodict(ie_result, strict=True) + title = common_info.get('playlist') or '<Untitled>' + if self._match_entry(common_info, incomplete=True) is not None: + return + self.to_screen(f'[download] Downloading {ie_result["_type"]}: {title}') all_entries = PlaylistEntries(self, ie_result) entries = orderedSet(all_entries.get_requested_items(), lazy=True) @@ -1689,12 +1798,14 @@ class YoutubeDL: # Better to do this after potentially exhausting entries ie_result['playlist_count'] = all_entries.get_full_count() + extra = self._playlist_infodict(ie_result, n_entries=int_or_none(n_entries)) + ie_copy = collections.ChainMap(ie_result, extra) + _infojson_written = False write_playlist_files = self.params.get('allow_playlist_files', True) if write_playlist_files and self.params.get('list_thumbnails'): self.list_thumbnails(ie_result) if write_playlist_files and not self.params.get('simulate'): - ie_copy = self._playlist_infodict(ie_result, n_entries=int_or_none(n_entries)) _infojson_written = self._write_info_json( 'playlist', ie_result, self.prepare_filename(ie_copy, 'pl_infojson')) if _infojson_written is None: @@ -1703,7 +1814,7 @@ class YoutubeDL: self.prepare_filename(ie_copy, 'pl_description')) is None: return # TODO: This should be passed to ThumbnailsConvertor if necessary - self._write_thumbnails('playlist', ie_copy, self.prepare_filename(ie_copy, 'pl_thumbnail')) + self._write_thumbnails('playlist', ie_result, self.prepare_filename(ie_copy, 'pl_thumbnail')) if lazy: if self.params.get('playlistreverse') or self.params.get('playlistrandom'): @@ -1713,53 +1824,62 @@ class YoutubeDL: elif self.params.get('playlistrandom'): random.shuffle(entries) - self.to_screen(f'[{ie_result["extractor"]}] Playlist {title}: Downloading {n_entries} videos' + self.to_screen(f'[{ie_result["extractor"]}] Playlist {title}: Downloading {n_entries} items' f'{format_field(ie_result, "playlist_count", " of %s")}') + keep_resolved_entries = self.params.get('extract_flat') != 'discard' + if self.params.get('extract_flat') == 'discard_in_playlist': + keep_resolved_entries = ie_result['_type'] != 'playlist' + if keep_resolved_entries: + self.write_debug('The information of all playlist entries will be held in memory') + failures = 0 max_failures = self.params.get('skip_playlist_after_errors') or float('inf') for i, (playlist_index, entry) in enumerate(entries): if lazy: resolved_entries.append((playlist_index, entry)) - - # TODO: Add auto-generated fields - if not entry or self._match_entry(entry, incomplete=True) is not None: + if not entry: continue - self.to_screen('[download] Downloading video %s of %s' % ( - self._format_screen(i + 1, self.Styles.ID), self._format_screen(n_entries, self.Styles.EMPHASIS))) - entry['__x_forwarded_for_ip'] = ie_result.get('__x_forwarded_for_ip') if not lazy and 'playlist-index' in self.params.get('compat_opts', []): playlist_index = ie_result['requested_entries'][i] - entry_result = self.__process_iterable_entry(entry, download, { + entry_copy = collections.ChainMap(entry, { + **common_info, 'n_entries': int_or_none(n_entries), - '__last_playlist_index': max(ie_result['requested_entries'] or (0, 0)), - 'playlist_count': ie_result.get('playlist_count'), 'playlist_index': playlist_index, 'playlist_autonumber': i + 1, - 'playlist': title, - 'playlist_id': ie_result.get('id'), - 'playlist_title': ie_result.get('title'), - 'playlist_uploader': ie_result.get('uploader'), - 'playlist_uploader_id': ie_result.get('uploader_id'), - 'extractor': ie_result['extractor'], - 'webpage_url': ie_result['webpage_url'], - 'webpage_url_basename': url_basename(ie_result['webpage_url']), - 'webpage_url_domain': get_domain(ie_result['webpage_url']), - 'extractor_key': ie_result['extractor_key'], }) + + if self._match_entry(entry_copy, incomplete=True) is not None: + # For compatabilty with youtube-dl. See https://github.com/yt-dlp/yt-dlp/issues/4369 + resolved_entries[i] = (playlist_index, NO_DEFAULT) + continue + + self.to_screen('[download] Downloading item %s of %s' % ( + self._format_screen(i + 1, self.Styles.ID), self._format_screen(n_entries, self.Styles.EMPHASIS))) + + extra.update({ + 'playlist_index': playlist_index, + 'playlist_autonumber': i + 1, + }) + entry_result = self.__process_iterable_entry(entry, download, extra) if not entry_result: failures += 1 if failures >= max_failures: self.report_error( f'Skipping the remaining entries in playlist "{title}" since {failures} items failed extraction') break - resolved_entries[i] = (playlist_index, entry_result) + if keep_resolved_entries: + resolved_entries[i] = (playlist_index, entry_result) # Update with processed data - ie_result['requested_entries'], ie_result['entries'] = tuple(zip(*resolved_entries)) or ([], []) + ie_result['entries'] = [e for _, e in resolved_entries if e is not NO_DEFAULT] + ie_result['requested_entries'] = [i for i, e in resolved_entries if e is not NO_DEFAULT] + if ie_result['requested_entries'] == try_call(lambda: list(range(1, ie_result['playlist_count'] + 1))): + # Do not set for full playlist + ie_result.pop('requested_entries') # Write the updated info to json if _infojson_written is True and self._write_info_json( @@ -1916,8 +2036,8 @@ class YoutubeDL: filter_parts.append(string) def _remove_unused_ops(tokens): - # Remove operators that we don't use and join them with the surrounding strings - # for example: 'mp4' '-' 'baseline' '-' '16x9' is converted to 'mp4-baseline-16x9' + # Remove operators that we don't use and join them with the surrounding strings. + # E.g. 'mp4' '-' 'baseline' '-' '16x9' is converted to 'mp4-baseline-16x9' ALLOWED_OPS = ('/', '+', ',', '(', ')') last_string, last_start, last_end, last_line = None, None, None, None for type, string, start, end, line in tokens: @@ -2033,14 +2153,13 @@ class YoutubeDL: the_only_video = video_fmts[0] if len(video_fmts) == 1 else None the_only_audio = audio_fmts[0] if len(audio_fmts) == 1 else None - output_ext = self.params.get('merge_output_format') - if not output_ext: - if the_only_video: - output_ext = the_only_video['ext'] - elif the_only_audio and not video_fmts: - output_ext = the_only_audio['ext'] - else: - output_ext = 'mkv' + output_ext = get_compatible_ext( + vcodecs=[f.get('vcodec') for f in video_fmts], + acodecs=[f.get('acodec') for f in audio_fmts], + vexts=[f['ext'] for f in video_fmts], + aexts=[f['ext'] for f in audio_fmts], + preferences=(try_call(lambda: self.params['merge_output_format'].split('/')) + or self.params.get('prefer_free_formats') and ('webm', 'mkv'))) filtered = lambda *keys: filter(None, (traverse_obj(fmt, *keys) for fmt in formats_info)) @@ -2066,6 +2185,7 @@ class YoutubeDL: 'vcodec': the_only_video.get('vcodec'), 'vbr': the_only_video.get('vbr'), 'stretched_ratio': the_only_video.get('stretched_ratio'), + 'aspect_ratio': the_only_video.get('aspect_ratio'), }) if the_only_audio: @@ -2073,6 +2193,7 @@ class YoutubeDL: 'acodec': the_only_audio.get('acodec'), 'abr': the_only_audio.get('abr'), 'asr': the_only_audio.get('asr'), + 'audio_channels': the_only_audio.get('audio_channels') }) return new_dict @@ -2279,10 +2400,9 @@ class YoutubeDL: else: info_dict['thumbnails'] = thumbnails - def _fill_common_fields(self, info_dict, is_video=True): + def _fill_common_fields(self, info_dict, final=True): # TODO: move sanitization here - if is_video: - # playlists are allowed to lack "title" + if final: title = info_dict.get('title', NO_DEFAULT) if title is NO_DEFAULT: raise ExtractorError('Missing "title" field in extractor result', @@ -2326,11 +2446,13 @@ class YoutubeDL: for key in live_keys: if info_dict.get(key) is None: info_dict[key] = (live_status == key) + if live_status == 'post_live': + info_dict['was_live'] = True # Auto generate title fields corresponding to the *_number fields when missing # in order to always have clean titles. This is very common for TV series. for field in ('chapter', 'season', 'episode'): - if info_dict.get('%s_number' % field) is not None and not info_dict.get(field): + if final and info_dict.get('%s_number' % field) is not None and not info_dict.get(field): info_dict[field] = '%s %d' % (field.capitalize(), info_dict['%s_number' % field]) def _raise_pending_errors(self, info): @@ -2338,6 +2460,18 @@ class YoutubeDL: if err: self.report_error(err, tb=False) + def sort_formats(self, info_dict): + formats = self._get_formats(info_dict) + if not formats: + return + # Backward compatibility with InfoExtractor._sort_formats + field_preference = formats[0].pop('__sort_fields', None) + if field_preference: + info_dict['_format_sort_fields'] = field_preference + + formats.sort(key=FormatSorter( + self, info_dict.get('_format_sort_fields', [])).calculate_preference) + def process_video_result(self, info_dict, download=True): assert info_dict.get('_type', 'video') == 'video' self._num_videos += 1 @@ -2375,13 +2509,18 @@ class YoutubeDL: self.report_warning('"duration" field is negative, there is an error in extractor') chapters = info_dict.get('chapters') or [] + if chapters and chapters[0].get('start_time'): + chapters.insert(0, {'start_time': 0}) + dummy_chapter = {'end_time': 0, 'start_time': info_dict.get('duration')} - for prev, current, next_ in zip( - (dummy_chapter, *chapters), chapters, (*chapters[1:], dummy_chapter)): + for idx, (prev, current, next_) in enumerate(zip( + (dummy_chapter, *chapters), chapters, (*chapters[1:], dummy_chapter)), 1): if current.get('start_time') is None: current['start_time'] = prev.get('end_time') if not current.get('end_time'): current['end_time'] = next_.get('start_time') + if not current.get('title'): + current['title'] = f'<Untitled Chapter {idx}>' if 'playlist' not in info_dict: # It isn't part of a playlist @@ -2418,21 +2557,18 @@ class YoutubeDL: info_dict['requested_subtitles'] = self.process_subtitles( info_dict['id'], subtitles, automatic_captions) - if info_dict.get('formats') is None: - # There's only one format available - formats = [info_dict] - else: - formats = info_dict['formats'] + self.sort_formats(info_dict) + formats = self._get_formats(info_dict) # or None ensures --clean-infojson removes it info_dict['_has_drm'] = any(f.get('has_drm') for f in formats) or None if not self.params.get('allow_unplayable_formats'): formats = [f for f in formats if not f.get('has_drm')] - if info_dict['_has_drm'] and all( - f.get('acodec') == f.get('vcodec') == 'none' for f in formats): - self.report_warning( - 'This video is DRM protected and only images are available for download. ' - 'Use --list-formats to see them') + + if formats and all(f.get('acodec') == f.get('vcodec') == 'none' for f in formats): + self.report_warning( + f'{"This video is DRM protected and " if info_dict["_has_drm"] else ""}' + 'only images are available for download. Use --list-formats to see them'.capitalize()) get_from_start = not info_dict.get('is_live') or bool(self.params.get('live_from_start')) if not get_from_start: @@ -2444,9 +2580,6 @@ class YoutubeDL: '--live-from-start is passed, but there are no formats that can be downloaded from the start. ' 'If you want to download from the current time, use --no-live-from-start')) - if not formats: - self.raise_no_formats(info_dict) - def is_wellformed(f): url = f.get('url') if not url: @@ -2459,7 +2592,10 @@ class YoutubeDL: return True # Filter out malformed formats for better extraction robustness - formats = list(filter(is_wellformed, formats)) + formats = list(filter(is_wellformed, formats or [])) + + if not formats: + self.raise_no_formats(info_dict) formats_dict = {} @@ -2505,6 +2641,8 @@ class YoutubeDL: format['resolution'] = self.format_resolution(format, default=None) if format.get('dynamic_range') is None and format.get('vcodec') != 'none': format['dynamic_range'] = 'SDR' + if format.get('aspect_ratio') is None: + format['aspect_ratio'] = try_call(lambda: round(format['width'] / format['height'], 2)) if (info_dict.get('duration') and format.get('tbr') and not format.get('filesize') and not format.get('filesize_approx')): format['filesize_approx'] = int(info_dict['duration'] * format['tbr'] * (1024 / 8)) @@ -2537,10 +2675,9 @@ class YoutubeDL: info_dict, _ = self.pre_process(info_dict, 'after_filter') # The pre-processors may have modified the formats - formats = info_dict.get('formats', [info_dict]) + formats = self._get_formats(info_dict) - list_only = self.params.get('simulate') is None and ( - self.params.get('list_thumbnails') or self.params.get('listformats') or self.params.get('listsubtitles')) + list_only = self.params.get('simulate') == 'list_only' interactive_format_selection = not list_only and self.format_selector == '-' if self.params.get('list_thumbnails'): self.list_thumbnails(info_dict) @@ -2595,31 +2732,30 @@ class YoutubeDL: # Process what we can, even without any available formats. formats_to_download = [{}] - requested_ranges = self.params.get('download_ranges') - if requested_ranges: - requested_ranges = tuple(requested_ranges(info_dict, self)) - + requested_ranges = tuple(self.params.get('download_ranges', lambda *_: [{}])(info_dict, self)) best_format, downloaded_formats = formats_to_download[-1], [] if download: - if best_format: + if best_format and requested_ranges: def to_screen(*msg): self.to_screen(f'[info] {info_dict["id"]}: {" ".join(", ".join(variadic(m)) for m in msg)}') to_screen(f'Downloading {len(formats_to_download)} format(s):', (f['format_id'] for f in formats_to_download)) - if requested_ranges: + if requested_ranges != ({}, ): to_screen(f'Downloading {len(requested_ranges)} time ranges:', - (f'{int(c["start_time"])}-{int(c["end_time"])}' for c in requested_ranges)) + (f'{c["start_time"]:.1f}-{c["end_time"]:.1f}' for c in requested_ranges)) max_downloads_reached = False - for fmt, chapter in itertools.product(formats_to_download, requested_ranges or [{}]): + for fmt, chapter in itertools.product(formats_to_download, requested_ranges): new_info = self._copy_infodict(info_dict) new_info.update(fmt) offset, duration = info_dict.get('section_start') or 0, info_dict.get('duration') or float('inf') + end_time = offset + min(chapter.get('end_time', duration), duration) if chapter or offset: new_info.update({ 'section_start': offset + chapter.get('start_time', 0), - 'section_end': offset + min(chapter.get('end_time', duration), duration), + # duration may not be accurate. So allow deviations <1sec + 'section_end': end_time if end_time <= offset + duration + 1 else None, 'section_title': chapter.get('title'), 'section_number': chapter.get('index'), }) @@ -2661,42 +2797,26 @@ class YoutubeDL: if lang not in available_subs: available_subs[lang] = cap_info - if (not self.params.get('writesubtitles') and not - self.params.get('writeautomaticsub') or not - available_subs): + if not available_subs or ( + not self.params.get('writesubtitles') + and not self.params.get('writeautomaticsub')): return None all_sub_langs = tuple(available_subs.keys()) if self.params.get('allsubtitles', False): requested_langs = all_sub_langs elif self.params.get('subtitleslangs', False): - # A list is used so that the order of languages will be the same as - # given in subtitleslangs. See https://github.com/yt-dlp/yt-dlp/issues/1041 - requested_langs = [] - for lang_re in self.params.get('subtitleslangs'): - discard = lang_re[0] == '-' - if discard: - lang_re = lang_re[1:] - if lang_re == 'all': - if discard: - requested_langs = [] - else: - requested_langs.extend(all_sub_langs) - continue - current_langs = filter(re.compile(lang_re + '$').match, all_sub_langs) - if discard: - for lang in current_langs: - while lang in requested_langs: - requested_langs.remove(lang) - else: - requested_langs.extend(current_langs) - requested_langs = orderedSet(requested_langs) + try: + requested_langs = orderedSet_from_options( + self.params.get('subtitleslangs'), {'all': all_sub_langs}, use_regex=True) + except re.error as e: + raise ValueError(f'Wrong regex for subtitlelangs: {e.pattern}') elif normal_sub_langs: requested_langs = ['en'] if 'en' in normal_sub_langs else normal_sub_langs[:1] else: requested_langs = ['en'] if 'en' in all_sub_langs else all_sub_langs[:1] if requested_langs: - self.write_debug('Downloading subtitles: %s' % ', '.join(requested_langs)) + self.to_screen(f'[info] {video_id}: Downloading subtitles: {", ".join(requested_langs)}') formats_query = self.params.get('subtitlesformat', 'best') formats_preference = formats_query.split('/') if formats_query else [] @@ -2732,12 +2852,16 @@ class YoutubeDL: info_copy['automatic_captions_table'] = self.render_subtitles_table(info_dict.get('id'), info_dict.get('automatic_captions')) def format_tmpl(tmpl): - mobj = re.match(r'\w+(=?)$', tmpl) - if mobj and mobj.group(1): - return f'{tmpl[:-1]} = %({tmpl[:-1]})r' - elif mobj: - return f'%({tmpl})s' - return tmpl + mobj = re.fullmatch(r'([\w.:,]|-\d|(?P<dict>{([\w.:,]|-\d)+}))+=?', tmpl) + if not mobj: + return tmpl + + fmt = '%({})s' + if tmpl.startswith('{'): + tmpl = f'.{tmpl}' + if tmpl.endswith('='): + tmpl, fmt = tmpl[:-1], '{0} = %({0})#j' + return '\n'.join(map(fmt.format, [tmpl] if mobj.group('dict') else tmpl.split(','))) for tmpl in self.params['forceprint'].get(key, []): self.to_stdout(self.evaluate_outtmpl(format_tmpl(tmpl), info_copy)) @@ -2845,8 +2969,6 @@ class YoutubeDL: if 'format' not in info_dict and 'ext' in info_dict: info_dict['format'] = info_dict['ext'] - # This is mostly just for backward compatibility of process_info - # As a side-effect, this allows for format-specific filters if self._match_entry(info_dict) is not None: info_dict['__write_download_archive'] = 'ignore' return @@ -3007,33 +3129,9 @@ class YoutubeDL: return if info_dict.get('requested_formats') is not None: - - def compatible_formats(formats): - # TODO: some formats actually allow this (mkv, webm, ogg, mp4), but not all of them. - video_formats = [format for format in formats if format.get('vcodec') != 'none'] - audio_formats = [format for format in formats if format.get('acodec') != 'none'] - if len(video_formats) > 2 or len(audio_formats) > 2: - return False - - # Check extension - exts = {format.get('ext') for format in formats} - COMPATIBLE_EXTS = ( - {'mp3', 'mp4', 'm4a', 'm4p', 'm4b', 'm4r', 'm4v', 'ismv', 'isma'}, - {'webm'}, - ) - for ext_sets in COMPATIBLE_EXTS: - if ext_sets.issuperset(exts): - return True - # TODO: Check acodec/vcodec - return False - requested_formats = info_dict['requested_formats'] old_ext = info_dict['ext'] if self.params.get('merge_output_format') is None: - if not compatible_formats(requested_formats): - info_dict['ext'] = 'mkv' - self.report_warning( - 'Requested formats are incompatible for merge and will be merged into mkv') if (info_dict['ext'] == 'webm' and info_dict.get('thumbnails') # check with type instead of pp_key, __name__, or isinstance @@ -3171,22 +3269,23 @@ class YoutubeDL: self.report_warning(f'{vid}: {msg}. Install ffmpeg to fix this automatically') stretched_ratio = info_dict.get('stretched_ratio') - ffmpeg_fixup( - stretched_ratio not in (1, None), - f'Non-uniform pixel ratio {stretched_ratio}', - FFmpegFixupStretchedPP) - - ffmpeg_fixup( - (info_dict.get('requested_formats') is None - and info_dict.get('container') == 'm4a_dash' - and info_dict.get('ext') == 'm4a'), - 'writing DASH m4a. Only some players support this container', - FFmpegFixupM4aPP) + ffmpeg_fixup(stretched_ratio not in (1, None), + f'Non-uniform pixel ratio {stretched_ratio}', + FFmpegFixupStretchedPP) downloader = get_suitable_downloader(info_dict, self.params) if 'protocol' in info_dict else None downloader = downloader.FD_NAME if downloader else None - if info_dict.get('requested_formats') is None: # Not necessary if doing merger + ext = info_dict.get('ext') + postprocessed_by_ffmpeg = info_dict.get('requested_formats') or any(( + isinstance(pp, FFmpegVideoConvertorPP) + and resolve_recode_mapping(ext, pp.mapping)[0] not in (ext, None) + ) for pp in self._pps['post_process']) + + if not postprocessed_by_ffmpeg: + ffmpeg_fixup(ext == 'm4a' and info_dict.get('container') == 'm4a_dash', + 'writing DASH m4a. Only some players support this container', + FFmpegFixupM4aPP) ffmpeg_fixup(downloader == 'hlsnative' and not self.params.get('hls_use_mpegts') or info_dict.get('is_live') and self.params.get('hls_use_mpegts') is None, 'Possible MPEG-TS in MP4 container or malformed AAC timestamps', @@ -3227,6 +3326,7 @@ class YoutubeDL: self.to_screen(f'[info] {e}') if not self.params.get('break_per_url'): raise + self._num_downloads = 0 else: if self.params.get('dump_single_json', False): self.post_extract(res) @@ -3275,6 +3375,12 @@ class YoutubeDL: return info_dict info_dict.setdefault('epoch', int(time.time())) info_dict.setdefault('_type', 'video') + info_dict.setdefault('_version', { + 'version': __version__, + 'current_git_head': current_git_head(), + 'release_git_head': RELEASE_GIT_HEAD, + 'repository': REPOSITORY, + }) if remove_private_keys: reject = lambda k, v: v is None or k.startswith('__') or k in { @@ -3392,18 +3498,15 @@ class YoutubeDL: break else: return - return f'{extractor.lower()} {video_id}' + return make_archive_id(extractor, video_id) def in_download_archive(self, info_dict): - fn = self.params.get('download_archive') - if fn is None: + if not self.archive: return False - vid_id = self._make_archive_id(info_dict) - if not vid_id: - return False # Incomplete video information - - return vid_id in self.archive + vid_ids = [self._make_archive_id(info_dict)] + vid_ids.extend(info_dict.get('_old_archive_ids') or []) + return any(id_ in self.archive for id_ in vid_ids) def record_download_archive(self, info_dict): fn = self.params.get('download_archive') @@ -3411,9 +3514,11 @@ class YoutubeDL: return vid_id = self._make_archive_id(info_dict) assert vid_id + self.write_debug(f'Adding to archive: {vid_id}') - with locked_file(fn, 'a', encoding='utf-8') as archive_file: - archive_file.write(vid_id + '\n') + if is_path_like(fn): + with locked_file(fn, 'a', encoding='utf-8') as archive_file: + archive_file.write(vid_id + '\n') self.archive.add(vid_id) @staticmethod @@ -3495,11 +3600,17 @@ class YoutubeDL: res += '~' + format_bytes(fdict['filesize_approx']) return res - def render_formats_table(self, info_dict): - if not info_dict.get('formats') and not info_dict.get('url'): - return None + def _get_formats(self, info_dict): + if info_dict.get('formats') is None: + if info_dict.get('url') and info_dict.get('_type', 'video') == 'video': + return [info_dict] + return [] + return info_dict['formats'] - formats = info_dict.get('formats', [info_dict]) + def render_formats_table(self, info_dict): + formats = self._get_formats(info_dict) + if not formats: + return if not self.params.get('listformats_table', True) is not False: table = [ [ @@ -3507,31 +3618,43 @@ class YoutubeDL: format_field(f, 'ext'), self.format_resolution(f), self._format_note(f) - ] for f in formats if f.get('preference') is None or f['preference'] >= -1000] + ] for f in formats if (f.get('preference') or 0) >= -1000] return render_table(['format code', 'extension', 'resolution', 'note'], table, extra_gap=1) + def simplified_codec(f, field): + assert field in ('acodec', 'vcodec') + codec = f.get(field, 'unknown') + if not codec: + return 'unknown' + elif codec != 'none': + return '.'.join(codec.split('.')[:4]) + + if field == 'vcodec' and f.get('acodec') == 'none': + return 'images' + elif field == 'acodec' and f.get('vcodec') == 'none': + return '' + return self._format_out('audio only' if field == 'vcodec' else 'video only', + self.Styles.SUPPRESS) + delim = self._format_out('\u2502', self.Styles.DELIM, '|', test_encoding=True) table = [ [ self._format_out(format_field(f, 'format_id'), self.Styles.ID), format_field(f, 'ext'), format_field(f, func=self.format_resolution, ignore=('audio only', 'images')), - format_field(f, 'fps', '\t%d'), + format_field(f, 'fps', '\t%d', func=round), format_field(f, 'dynamic_range', '%s', ignore=(None, 'SDR')).replace('HDR', ''), + format_field(f, 'audio_channels', '\t%s'), delim, format_field(f, 'filesize', ' \t%s', func=format_bytes) + format_field(f, 'filesize_approx', '~\t%s', func=format_bytes), - format_field(f, 'tbr', '\t%dk'), + format_field(f, 'tbr', '\t%dk', func=round), shorten_protocol_name(f.get('protocol', '')), delim, - format_field(f, 'vcodec', default='unknown').replace( - 'none', 'images' if f.get('acodec') == 'none' - else self._format_out('audio only', self.Styles.SUPPRESS)), - format_field(f, 'vbr', '\t%dk'), - format_field(f, 'acodec', default='unknown').replace( - 'none', '' if f.get('vcodec') == 'none' - else self._format_out('video only', self.Styles.SUPPRESS)), - format_field(f, 'abr', '\t%dk'), - format_field(f, 'asr', '\t%dHz'), + simplified_codec(f, 'vcodec'), + format_field(f, 'vbr', '\t%dk', func=round), + simplified_codec(f, 'acodec'), + format_field(f, 'abr', '\t%dk', func=round), + format_field(f, 'asr', '\t%s', func=format_decimal_suffix), join_nonempty( self._format_out('UNSUPPORTED', 'light red') if f.get('ext') in ('f4f', 'f4m') else None, format_field(f, 'language', '[%s]'), @@ -3541,7 +3664,7 @@ class YoutubeDL: delim=' '), ] for f in formats if f.get('preference') is None or f['preference'] >= -1000] header_line = self._list_format_headers( - 'ID', 'EXT', 'RESOLUTION', '\tFPS', 'HDR', delim, '\tFILESIZE', '\tTBR', 'PROTO', + 'ID', 'EXT', 'RESOLUTION', '\tFPS', 'HDR', 'CH', delim, '\tFILESIZE', '\tTBR', 'PROTO', delim, 'VCODEC', '\tVBR', 'ACODEC', '\tABR', '\tASR', 'MORE INFO') return render_table( @@ -3554,7 +3677,7 @@ class YoutubeDL: return None return render_table( self._list_format_headers('ID', 'Width', 'Height', 'URL'), - [[t.get('id'), t.get('width', 'unknown'), t.get('height', 'unknown'), t['url']] for t in thumbnails]) + [[t.get('id'), t.get('width') or 'unknown', t.get('height') or 'unknown', t['url']] for t in thumbnails]) def render_subtitles_table(self, video_id, subtitles): def _row(lang, formats): @@ -3597,6 +3720,8 @@ class YoutubeDL: if not self.params.get('verbose'): return + from . import _IN_CLI # Must be delayed import + # These imports can be slow. So import them only as needed from .extractor.extractors import _LAZY_LOADER from .extractor.extractors import _PLUGIN_CLASSES as plugin_extractors @@ -3626,11 +3751,19 @@ class YoutubeDL: write_debug = lambda msg: self._write_string(f'[debug] {msg}\n') source = detect_variant() + if VARIANT not in (None, 'pip'): + source += '*' write_debug(join_nonempty( - 'yt-dlp version', __version__, + f'{"yt-dlp" if REPOSITORY == "yt-dlp/yt-dlp" else REPOSITORY} version', + __version__, f'[{RELEASE_GIT_HEAD}]' if RELEASE_GIT_HEAD else '', '' if source == 'unknown' else f'({source})', + '' if _IN_CLI else 'API', delim=' ')) + + if not _IN_CLI: + write_debug(f'params: {self.params}') + if not _LAZY_LOADER: if os.environ.get('YTDLP_NO_LAZY_EXTRACTORS'): write_debug('Lazy loading extractors is forcibly disabled') @@ -3643,29 +3776,9 @@ class YoutubeDL: if self.params['compat_opts']: write_debug('Compatibility options: %s' % ', '.join(self.params['compat_opts'])) - if source == 'source': - try: - stdout, _, _ = Popen.run( - ['git', 'rev-parse', '--short', 'HEAD'], - text=True, cwd=os.path.dirname(os.path.abspath(__file__)), - stdout=subprocess.PIPE, stderr=subprocess.PIPE) - if re.fullmatch('[0-9a-f]+', stdout.strip()): - write_debug(f'Git HEAD: {stdout.strip()}') - except Exception: - with contextlib.suppress(Exception): - sys.exc_clear() - - def python_implementation(): - impl_name = platform.python_implementation() - if impl_name == 'PyPy' and hasattr(sys, 'pypy_version_info'): - return impl_name + ' version %d.%d.%d' % sys.pypy_version_info[:3] - return impl_name - - write_debug('Python version %s (%s %s) - %s' % ( - platform.python_version(), - python_implementation(), - platform.architecture()[0], - platform_name())) + if current_git_head(): + write_debug(f'Git HEAD: {current_git_head()}') + write_debug(system_identifier()) exe_versions, ffmpeg_features = FFmpegPostProcessor.get_versions_and_features(self) ffmpeg_features = {key for key, val in ffmpeg_features.items() if val} |