You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
yt-dlp/yt_dlp/YoutubeDL.py

3983 lines
187 KiB

import collections
import contextlib
import datetime
import errno
import fileinput
import functools
import io
import itertools
import json
import locale
import operator
import os
import random
import re
import shutil
import subprocess
import sys
import tempfile
import time
import tokenize
import traceback
import unicodedata
import urllib.request
from string import ascii_letters
from .cache import Cache
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
from .postprocessor import (
EmbedThumbnailPP,
FFmpegFixupDuplicateMoovPP,
FFmpegFixupDurationPP,
FFmpegFixupM3u8PP,
FFmpegFixupM4aPP,
FFmpegFixupStretchedPP,
FFmpegFixupTimestampPP,
FFmpegMergerPP,
FFmpegPostProcessor,
FFmpegVideoConvertorPP,
MoveFilesAfterDownloadPP,
get_postprocessor,
)
from .postprocessor.ffmpeg import resolve_mapping as resolve_recode_mapping
from .update import REPOSITORY, current_git_head, detect_variant
from .utils import (
DEFAULT_OUTTMPL,
IDENTITY,
LINK_TEMPLATES,
MEDIA_EXTENSIONS,
NO_DEFAULT,
NUMBER_RE,
OUTTMPL_TYPES,
POSTPROCESS_WHEN,
STR_FORMAT_RE_TMPL,
STR_FORMAT_TYPES,
ContentTooShortError,
DateRange,
DownloadCancelled,
DownloadError,
EntryNotInPlaylist,
ExistingVideoReached,
ExtractorError,
GeoRestrictedError,
HEADRequest,
ISO3166Utils,
LazyList,
MaxDownloadsReached,
Namespace,
PagedList,
PerRequestProxyHandler,
PlaylistEntries,
Popen,
PostProcessingError,
ReExtractInfo,
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,
format_bytes,
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,
preferredencoding,
prepend_extension,
register_socks_protocols,
remove_terminal_sequences,
render_table,
replace_extension,
sanitize_filename,
sanitize_path,
sanitize_url,
sanitized_Request,
std_headers,
str_or_none,
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,
version_tuple,
windows_enable_vt_mode,
write_json_file,
write_string,
)
from .version import RELEASE_GIT_HEAD, VARIANT, __version__
if compat_os_name == 'nt':
import ctypes
class YoutubeDL:
"""YoutubeDL class.
YoutubeDL objects are the ones responsible of downloading the
actual video file and writing it to disk if the user has requested
it, among some other tasks. In most cases there should be one per
program. As, given a video URL, the downloader doesn't know how to
extract all the needed information, task that InfoExtractors do, it
has to pass the URL to one of them.
For this, YoutubeDL objects have a method that allows
InfoExtractors to be registered in a given order. When it is passed
a URL, the YoutubeDL object handles it to the first InfoExtractor it
finds that reports being able to handle it. The InfoExtractor extracts
all the information about the video or videos the URL refers to, and
YoutubeDL process the extracted information, possibly using a File
Downloader to download the video.
YoutubeDL objects accept a lot of parameters. In order not to saturate
the object constructor with arguments, it receives a dictionary of
options instead. These options are available through the params
attribute for the InfoExtractors to use. The YoutubeDL also
registers itself as the downloader in charge for the InfoExtractors
that are added to it, so this is a "mutual registration".
Available options:
username: Username for authentication purposes.
password: Password for authentication purposes.
videopassword: Password for accessing a video.
ap_mso: Adobe Pass multiple-system operator identifier.
ap_username: Multiple-system operator account username.
ap_password: Multiple-system operator account password.
usenetrc: Use netrc for authentication instead.
verbose: Print additional info to stdout.
quiet: Do not print messages to stdout.
no_warnings: Do not print out anything for warnings.
forceprint: A dict with keys WHEN mapped to a list of templates to
print to stdout. The allowed keys are video or any of the
items in utils.POSTPROCESS_WHEN.
For compatibility, a single list is also accepted
print_to_file: A dict with keys WHEN (same as forceprint) mapped to
a list of tuples with (template, filename)
forcejson: Force printing info_dict as JSON.
dump_single_json: Force printing the info_dict of the whole playlist
(or video) as a single JSON line.
force_write_download_archive: Force writing download archive regardless
of 'skip_download' or 'simulate'.
simulate: Do not download the video files. If unset (or None),
simulate only if listsubtitles, listformats or list_thumbnails is used
format: Video format code. see "FORMAT SELECTION" for more details.
You can also pass a function. The function takes 'ctx' as
argument and returns the formats to download.
See "build_format_selector" for an implementation
allow_unplayable_formats: Allow unplayable formats to be extracted and downloaded.
ignore_no_formats_error: Ignore "No video formats" error. Usefull for
extracting metadata even if the video is not actually
available for download (experimental)
format_sort: A list of fields by which to sort the video formats.
See "Sorting Formats" for more details.
format_sort_force: Force the given format_sort. see "Sorting Formats"
for more details.
prefer_free_formats: Whether to prefer video formats with free containers
over non-free ones of same quality.
allow_multiple_video_streams: Allow multiple video streams to be merged
into a single file
allow_multiple_audio_streams: Allow multiple audio streams to be merged
into a single file
check_formats Whether to test if the formats are downloadable.
Can be True (check all), False (check none),
'selected' (check selected formats),
or None (check only if requested by extractor)
paths: Dictionary of output paths. The allowed keys are 'home'
'temp' and the keys of OUTTMPL_TYPES (in utils.py)
outtmpl: Dictionary of templates for output names. Allowed keys
are 'default' and the keys of OUTTMPL_TYPES (in utils.py).
For compatibility with youtube-dl, a single string can also be used
outtmpl_na_placeholder: Placeholder for unavailable meta fields.
restrictfilenames: Do not allow "&" and spaces in file names
trim_file_name: Limit length of filename (extension excluded)
windowsfilenames: Force the filenames to be windows compatible
ignoreerrors: Do not stop on download/postprocessing errors.
Can be 'only_download' to ignore only download errors.
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
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
For compatibility with youtube-dl,
"nooverwrites" may also be used instead
playlist_items: Specific indices of playlist to download.
playlistrandom: Download playlist items in random order.
lazy_playlist: Process playlist entries as they are received.
matchtitle: Download only matching titles.
rejecttitle: Reject downloads for matching titles.
logger: Log messages to a logging.Logger instance.
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
getcomments: Extract video comments. This will not be written to disk
unless writeinfojson is also given
writeannotations: Write the video annotations to a .annotations.xml file
writethumbnail: Write the thumbnail image to a file
allow_playlist_files: Whether to write playlists' description, infojson etc
also to disk when using the 'write*' options
write_all_thumbnails: Write all thumbnail formats to files
writelink: Write an internet shortcut file, depending on the
current platform (.url/.webloc/.desktop)
writeurllink: Write a Windows internet shortcut file (.url)
writewebloclink: Write a macOS internet shortcut file (.webloc)
writedesktoplink: Write a Linux internet shortcut file (.desktop)
writesubtitles: Write the video subtitles to a file
writeautomaticsub: Write the automatically generated subtitles to a file
listsubtitles: Lists all available subtitles for the video
subtitlesformat: The format code for subtitles
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, 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
cachedir: Location of the cache files in the filesystem.
False to disable filesystem cache.
noplaylist: Download single video instead of a playlist if in doubt.
age_limit: An integer representing the user's age in years.
Unsuitable videos for the given age are skipped.
min_views: An integer representing the minimum view count the video
must have in order to not be skipped.
Videos without view count information are always
downloaded. None for no limit.
max_views: An integer representing the maximum view count.
Videos that are more popular than that are not
downloaded.
Videos without view count information are always
downloaded. None for no limit.
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
has been filtered out.
break_per_url: Whether break_on_reject and break_on_existing
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/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
client_certificate: Path to client certificate file in PEM format. May include the private key
client_certificate_key: Path to private key file for client certificate
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.
(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
on geo-restricted sites.
socket_timeout: Time to wait for unresponsive hosts, in seconds
bidi_workaround: Work around buggy terminals without bidirectional text
support, using fridibi
debug_printtraffic:Print out sent and received HTTP traffic
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: 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
postprocessors: A list of dictionaries, each with an entry
* key: The name of the postprocessor. See
yt_dlp/postprocessor/__init__.py for a list.
* when: When to run the postprocessor. Allowed values are
the entries of utils.POSTPROCESS_WHEN
Assumed to be 'post_process' if not given
progress_hooks: A list of functions that get called on download
progress, with a dictionary with the entries
* status: One of "downloading", "error", or "finished".
Check this first and ignore unknown values.
* info_dict: The extracted info_dict
If status is one of "downloading", or "finished", the
following properties may also be present:
* filename: The final filename (always present)
* tmpfilename: The filename we're currently writing to
* downloaded_bytes: Bytes on disk
* total_bytes: Size of the whole file, None if unknown
* total_bytes_estimate: Guess of the eventual file size,
None if unavailable.
* elapsed: The number of seconds since download started.
* eta: The estimated time in seconds, None if unknown
* speed: The download speed in bytes/second, None if
unknown
* fragment_index: The counter of the currently
downloaded video fragment.
* fragment_count: The number of fragments (= individual
files that will be merged)
Progress hooks are guaranteed to be called at least once
(with status "finished") if the download is successful.
postprocessor_hooks: A list of functions that get called on postprocessing
progress, with a dictionary with the entries
* status: One of "started", "processing", or "finished".
Check this first and ignore unknown values.
* postprocessor: Name of the postprocessor
* info_dict: The extracted info_dict
Progress hooks are guaranteed to be called at least twice
(with status "started" and "finished") if the processing is successful.
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.
One of:
- "never": do nothing
- "warn": only emit a warning
- "detect_or_warn": check whether we can do anything
about it, warn otherwise (default)
source_address: Client-side IP address to bind to.
sleep_interval_requests: Number of seconds to sleep between requests
during extraction
sleep_interval: Number of seconds to sleep before each download when
used alone or a lower bound of a range for randomized
sleep before each download (minimum possible number
of seconds to sleep) when used along with
max_sleep_interval.
max_sleep_interval:Upper bound of a range for randomized sleep before each
download (maximum possible number of seconds to sleep).
Must only be used along with sleep_interval.
Actual sleep time will be a random float from range
[sleep_interval; max_sleep_interval].
sleep_interval_subtitles: Number of seconds to sleep before each subtitle download
listformats: Print an overview of available video formats and exit.
list_thumbnails: Print a table of all thumbnails and exit.
match_filter: A function that gets called for every video with the signature
(info_dict, *, incomplete: bool) -> Optional[str]
For backward compatibility with youtube-dl, the signature
(info_dict) -> Optional[str] is also allowed.
- If it returns a message, the video is ignored.
- If it returns None, the video is downloaded.
- If it returns utils.NO_DEFAULT, the user is interactively
asked whether to download the video.
match_filter_func in utils.py is one example for this.
no_color: Do not emit color codes in output.
geo_bypass: Bypass geographic restriction via faking X-Forwarded-For
HTTP header
geo_bypass_country:
Two-letter ISO 3166-2 country code that will be used for
explicit geographic restriction bypassing via faking
X-Forwarded-For HTTP header
geo_bypass_ip_block:
IP range in CIDR notation that will be used similarly to
geo_bypass_country
external_downloader: A dictionary of protocol keys and the executable of the
external downloader to use for it. The allowed protocols
are default|http|ftp|m3u8|dash|rtsp|rtmp|mms.
Set the value to 'native' to use the native downloader
compat_opts: Compatibility options. See "Differences in default behavior".
The following options do not work when used through the API:
filename, abort-on-error, multistreams, no-live-chat, format-sort
no-clean-infojson, no-playlist-metafiles, no-keep-subs, no-attach-info-json.
Refer __init__.py for their implementation
progress_template: Dictionary of templates for progress outputs.
Allowed keys are 'download', 'postprocess',
'download-title' (console title) and 'postprocess-title'.
The template is mapped on a dictionary with keys 'progress' and 'info'
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 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, xattr_set_filesize, hls_use_mpegts, http_chunk_size,
external_downloader_args, concurrent_fragment_downloads.
The following options are used by the post processors:
ffmpeg_location: Location of the ffmpeg/avconv binary; either the path
to the binary or its containing directory.
postprocessor_args: A dictionary of postprocessor/executable keys (in lower case)
and a list of additional command-line arguments for the
postprocessor/executable. The dict can also have "PP+EXE" keys
which are used when the given exe is used by the given PP.
Use 'default' as the name for arguments to passed to all PP
For compatibility with youtube-dl, a single list of args
can also be used
The following options are used by the extractors:
extractor_retries: Number of times to retry for known errors
dynamic_mpd: Whether to process dynamic DASH manifests (default: True)
hls_split_discontinuity: Split HLS playlists to different formats at
discontinuities such as ad breaks (default: False)
extractor_args: A dictionary of arguments to be passed to the extractors.
See "EXTRACTOR ARGUMENTS" for details.
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
Playlist item to end at.
playlistreverse: - Use playlist_items
Download playlist items in reverse order.
forceurl: - Use forceprint
Force printing final URL.
forcetitle: - Use forceprint
Force printing title.
forceid: - Use forceprint
Force printing ID.
forcethumbnail: - Use forceprint
Force printing thumbnail URL.
forcedescription: - Use forceprint
Force printing description.
forcefilename: - Use forceprint
Force printing final filename.
forceduration: - Use forceprint
Force printing duration.
allsubtitles: - Use subtitleslangs = ['all']
Downloads all the subtitles of the video
(requires writesubtitles or writeautomaticsub)
include_ads: - Doesn't work
Download ads as well
call_home: - Not implemented
Boolean, true iff we are allowed to contact the
yt-dlp servers for debugging.
post_hooks: - Register a custom postprocessor
A list of functions that get called as the final step
for each video file, after all postprocessors have been
called. The filename will be passed as the only argument.
hls_prefer_native: - Use external_downloader = {'m3u8': 'native'} or {'m3u8': 'ffmpeg'}.
Use the native HLS downloader instead of ffmpeg/avconv
if True, otherwise use ffmpeg/avconv if False, otherwise
use downloader suggested by extractor if None.
prefer_ffmpeg: - avconv support is deprecated
If False, use avconv instead of ffmpeg if both are available,
otherwise prefer ffmpeg.
youtube_include_dash_manifest: - Use extractor_args
If True (default), DASH manifests and related
data will be downloaded and processed by extractor.
You can reduce network I/O by disabling it if you don't
care about DASH. (only for youtube)
youtube_include_hls_manifest: - Use extractor_args
If True (default), HLS manifests and related
data will be downloaded and processed by extractor.
You can reduce network I/O by disabling it if you don't
care about HLS. (only for youtube)
"""
_NUMERIC_FIELDS = {
'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',
'start_time', 'end_time',
'chapter_number', 'season_number', 'episode_number',
'track_number', 'disc_number', 'release_year',
}
_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', 'audio_channels',
'vbr', 'fps', 'vcodec', 'container', 'filesize', 'filesize_approx',
'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': 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):
"""Create a FileDownloader object with the given options.
@param auto_init Whether to load the default extractors and print header (if verbose).
Set to 'no_verbose_header' to not print the header
"""
if params is None:
params = {}
self.params = params
self._ies = {}
self._ies_instances = {}
self._pps = {k: [] for k in POSTPROCESS_WHEN}
self._printed_messages = set()
self._first_webpage_request = True
self._post_hooks = []
self._progress_hooks = []
self._postprocessor_hooks = []
self._download_retcode = 0
self._num_downloads = 0
self._num_videos = 0
self._playlist_level = 0
self._playlist_urls = set()
self.cache = Cache(self)
windows_enable_vt_mode()
stdout = sys.stderr if self.params.get('logtostderr') else sys.stdout
self._out_files = Namespace(
out=stdout,
error=sys.stderr,
screen=sys.stderr if self.params.get('quiet') else stdout,
console=None if compat_os_name == 'nt' else next(
filter(supports_terminal_sequences, (sys.stderr, sys.stdout)), None)
)
self._allow_colors = Namespace(**{
type_: not self.params.get('no_color') and supports_terminal_sequences(stream)
for type_, stream in self._out_files.items_ if type_ != 'console'
})
# 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. '
'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(
f'{msg}! Please update to Python %d.%d or above' % (*current_version, *MIN_RECOMMENDED))
if self.params.get('allow_unplayable_formats'):
self.report_warning(
f'You have asked for {self._format_err("UNPLAYABLE", self.Styles.EMPHASIS)} formats to be listed/downloaded. '
'This is a developer option intended for debugging. \n'
' If you experience any issues while using this option, '
f'{self._format_err("DO NOT", self.Styles.ERROR)} open a bug report')
def check_deprecated(param, option, suggestion):
if self.params.get(param) is not None:
self.report_warning(f'{option} is deprecated. Use {suggestion} instead')
return True
return False
if check_deprecated('cn_verification_proxy', '--cn-verification-proxy', '--geo-verification-proxy'):
if self.params.get('geo_verification_proxy') is None:
self.params['geo_verification_proxy'] = self.params['cn_verification_proxy']
check_deprecated('autonumber', '--auto-number', '-o "%(autonumber)s-%(title)s.%(ext)s"')
check_deprecated('usetitle', '--title', '-o "%(title)s-%(id)s.%(ext)s"')
check_deprecated('useid', '--id', '-o "%(id)s.%(ext)s"')
for msg in self.params.get('_warnings', []):
self.report_warning(msg)
for msg in self.params.get('_deprecation_warnings', []):
self.deprecated_feature(msg)
self.params['compat_opts'] = set(self.params.get('compat_opts', ()))
if 'list-formats' in self.params['compat_opts']:
self.params['listformats_table'] = False
if 'overwrites' not in self.params and self.params.get('nooverwrites') is not None:
# nooverwrites was unnecessarily changed to overwrites
# in 0c3d0f51778b153f65c21906031c2e091fcfb641
# This ensures compatibility with both keys
self.params['overwrites'] = not self.params['nooverwrites']
elif self.params.get('overwrites') is None:
self.params.pop('overwrites', None)
else:
self.params['nooverwrites'] = not self.params['overwrites']
self.params.setdefault('forceprint', {})
self.params.setdefault('print_to_file', {})
# Compatibility with older syntax
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'
and sys.getfilesystemencoding() in ['ascii', 'ANSI_X3.4-1968']
and not self.params.get('restrictfilenames', False)):
# Unicode filesystem API will throw errors (#1474, #13027)
self.report_warning(
'Assuming --restrict-filenames since file system encoding '
'cannot encode all characters. '
'Set the LC_ALL environment variable to fix this.')
self.params['restrictfilenames'] = True
self._parse_outtmpl()
# Creating format selector here allows us to catch syntax errors before the extraction
self.format_selector = (
self.params.get('format') if self.params.get('format') in (None, '-')
else self.params['format'] if callable(self.params['format'])
else self.build_format_selector(self.params['format']))
# Set http_headers defaults according to std_headers
self.params['http_headers'] = merge_headers(std_headers, self.params.get('http_headers', {}))
hooks = {
'post_hooks': self.add_post_hook,
'progress_hooks': self.add_progress_hook,
'postprocessor_hooks': self.add_postprocessor_hook,
}
for opt, fn in hooks.items():
for ph in self.params.get(opt, []):
fn(ph)
for pp_def_raw in self.params.get('postprocessors', []):
pp_def = dict(pp_def_raw)
when = pp_def.pop('when', 'post_process')
self.add_post_processor(
get_postprocessor(pp_def.pop('key'))(self, **pp_def),
when=when)
self._setup_opener()
register_socks_protocols()
def preload_download_archive(fn):
"""Preload the archive, if any is specified"""
archive = set()
if fn is None:
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:
archive.add(line.strip())
except OSError as ioe:
if ioe.errno != errno.ENOENT:
raise
return archive
self.archive = preload_download_archive(self.params.get('download_archive'))
def warn_if_short_id(self, argv):
# short YouTube ID starting with dash?
idxs = [
i for i, a in enumerate(argv)
if re.match(r'^-[0-9A-Za-z_-]{10}$', a)]
if idxs:
correct_argv = (
['yt-dlp']
+ [a for i, a in enumerate(argv) if i not in idxs]
+ ['--'] + [argv[i] for i in idxs]
)
self.report_warning(
'Long argument string detected. '
'Use -- to separate parameters and URLs, like this:\n%s' %
args_to_str(correct_argv))
def add_info_extractor(self, ie):
"""Add an InfoExtractor object to the end of the list."""
ie_key = ie.ie_key()
self._ies[ie_key] = ie
if not isinstance(ie, type):
self._ies_instances[ie_key] = ie
ie.set_downloader(self)
def get_info_extractor(self, ie_key):
"""
Get an instance of an IE with name ie_key, it will try to get one from
the _ies list, if there's no instance it will create a new one and add
it to the extractor list.
"""
ie = self._ies_instances.get(ie_key)
if ie is None:
ie = get_info_extractor(ie_key)()
self.add_info_extractor(ie)
return ie
def add_default_info_extractors(self):
"""
Add the InfoExtractors returned by gen_extractors to the end of the list
"""
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."""
assert when in POSTPROCESS_WHEN, f'Invalid when={when}'
self._pps[when].append(pp)
pp.set_downloader(self)
def add_post_hook(self, ph):
"""Add the post hook"""
self._post_hooks.append(ph)
def add_progress_hook(self, ph):
"""Add the download progress hook"""
self._progress_hooks.append(ph)
def add_postprocessor_hook(self, ph):
"""Add the postprocessing progress hook"""
self._postprocessor_hooks.append(ph)
for pps in self._pps.values():
for pp in pps:
pp.add_progress_hook(ph)
def _bidi_workaround(self, message):
if not hasattr(self, '_output_channel'):
return message
assert hasattr(self, '_output_process')
assert isinstance(message, str)
line_count = message.count('\n') + 1
self._output_process.stdin.write((message + '\n').encode())
self._output_process.stdin.flush()
res = ''.join(self._output_channel.readline().decode()
for _ in range(line_count))
return res[:-len('\n')]
def _write_string(self, message, out=None, only_once=False):
if only_once:
if message in self._printed_messages:
return
self._printed_messages.add(message)
write_string(message, out=out, encoding=self.params.get('encoding'))
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')
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._write_string(f'{self._bidi_workaround(message)}\n', self._out_files.out)
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)
return
if (self.params.get('quiet') if quiet is None else quiet) and not self.params.get('verbose'):
return
self._write_string(
'%s%s' % (self._bidi_workaround(message), ('' if skip_eol else '\n')),
self._out_files.screen, only_once=only_once)
def to_stderr(self, message, only_once=False):
"""Print message to stderr"""
assert isinstance(message, str)
if self.params.get('logger'):
self.params['logger'].error(message)
else:
self._write_string(f'{self._bidi_workaround(message)}\n', self._out_files.error, only_once=only_once)
def _send_console_code(self, code):
if compat_os_name == 'nt' or not self._out_files.console:
return
self._write_string(code, self._out_files.console)
def to_console_title(self, message):
if not self.params.get('consoletitle', False):
return
message = remove_terminal_sequences(message)
if compat_os_name == 'nt':
if ctypes.windll.kernel32.GetConsoleWindow():
# c_wchar_p() might not be necessary if `message` is
# already of type unicode()
ctypes.windll.kernel32.SetConsoleTitleW(ctypes.c_wchar_p(message))
else:
self._send_console_code(f'\033]0;{message}\007')
def save_console_title(self):
if not self.params.get('consoletitle') or self.params.get('simulate'):
return
self._send_console_code('\033[22;0t') # Save the title on stack
def restore_console_title(self):
if not self.params.get('consoletitle') or self.params.get('simulate'):
return
self._send_console_code('\033[23;0t') # Restore the title from stack
def __enter__(self):
self.save_console_title()
return self
def __exit__(self, *args):
self.restore_console_title()
if self.params.get('cookiefile') is not None:
self.cookiejar.save(ignore_discard=True, ignore_expires=True)
def trouble(self, message=None, tb=None, is_error=True):
"""Determine action to take when a download problem appears.
Depending on if the downloader has been configured to ignore
download errors or not, this method may throw an exception or
not when errors are found, after printing the message.
@param tb If given, is additional traceback information
@param is_error Whether to raise error according to ignorerrors
"""
if message is not None:
self.to_stderr(message)
if self.params.get('verbose'):
if tb is None:
if sys.exc_info()[0]: # if .trouble has been called from an except block
tb = ''
if hasattr(sys.exc_info()[1], 'exc_info') and sys.exc_info()[1].exc_info[0]:
tb += ''.join(traceback.format_exception(*sys.exc_info()[1].exc_info))
tb += encode_compat_str(traceback.format_exc())
else:
tb_data = traceback.format_list(traceback.extract_stack())
tb = ''.join(tb_data)
if tb:
self.to_stderr(tb)
if not is_error:
return
if not self.params.get('ignoreerrors'):
if sys.exc_info()[0] and hasattr(sys.exc_info()[1], 'exc_info') and sys.exc_info()[1].exc_info[0]:
exc_info = sys.exc_info()[1].exc_info
else:
exc_info = sys.exc_info()
raise DownloadError(message, exc_info)
self._download_retcode = 1
Styles = Namespace(
HEADERS='yellow',
EMPHASIS='light blue',
FILENAME='green',
ID='green',
DELIM='blue',
ERROR='red',
WARNING='yellow',
SUPPRESS='light black',
)
def _format_text(self, handle, allow_colors, text, f, fallback=None, *, test_encoding=False):
text = str(text)
if test_encoding:
original_text = text
# handle.encoding can be None. See https://github.com/yt-dlp/yt-dlp/issues/2711
encoding = self.params.get('encoding') or getattr(handle, 'encoding', None) or 'ascii'
text = text.encode(encoding, 'ignore').decode(encoding)
if fallback is not None and text != original_text:
text = fallback
return format_text(text, f) if allow_colors else text if fallback is None else fallback
def _format_out(self, *args, **kwargs):
return self._format_text(self._out_files.out, self._allow_colors.out, *args, **kwargs)
def _format_screen(self, *args, **kwargs):
return self._format_text(self._out_files.screen, self._allow_colors.screen, *args, **kwargs)
def _format_err(self, *args, **kwargs):
return self._format_text(self._out_files.error, self._allow_colors.error, *args, **kwargs)
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 self.params.get('logger') is not None:
self.params['logger'].warning(message)
else:
if self.params.get('no_warnings'):
return
self.to_stderr(f'{self._format_err("WARNING:", self.Styles.WARNING)} {message}', only_once)
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'Deprecated Feature: {message}')
self.to_stderr(f'{self._format_err("Deprecated Feature:", self.Styles.ERROR)} {message}', True)
def report_error(self, message, *args, **kwargs):
'''
Do the same as trouble, but prefixes the message with 'ERROR:', colored
in red if stderr is a tty file.
'''
self.trouble(f'{self._format_err("ERROR:", self.Styles.ERROR)} {message}', *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 = f'[debug] {message}'
if self.params.get('logger'):
self.params['logger'].debug(message)
else:
self.to_stderr(message, only_once)
def report_file_already_downloaded(self, file_name):
"""Report file has already been fully downloaded."""
try:
self.to_screen('[download] %s has already been downloaded' % file_name)
except UnicodeEncodeError:
self.to_screen('[download] The file has already been downloaded')
def report_file_delete(self, file_name):
"""Report that existing file will be deleted."""
try:
self.to_screen('Deleting existing file %s' % file_name)
except UnicodeEncodeError:
self.to_screen('Deleting existing file')
def raise_no_formats(self, info, forced=False, *, msg=None):
has_drm = info.get('_has_drm')
ignored, expected = self.params.get('ignore_no_formats_error'), bool(msg)
msg = msg or has_drm and 'This video is DRM protected' or 'No video formats found!'
if forced or not ignored:
raise ExtractorError(msg, video_id=info['id'], ie=info['extractor'],
expected=has_drm or ignored or expected)
else:
self.report_warning(msg)
def parse_outtmpl(self):
self.deprecation_warning('"YoutubeDL.parse_outtmpl" is deprecated and may be removed in a future version')
self._parse_outtmpl()
return self.params['outtmpl']
def _parse_outtmpl(self):
sanitize = IDENTITY
if self.params.get('restrictfilenames'): # Remove spaces in the default template
sanitize = lambda x: x.replace(' - ', ' ').replace(' ', '-')
outtmpl = self.params.setdefault('outtmpl', {})
if not isinstance(outtmpl, dict):
self.params['outtmpl'] = outtmpl = {'default': outtmpl}
outtmpl.update({k: sanitize(v) for k, v in DEFAULT_OUTTMPL.items() if outtmpl.get(k) is None})
def get_output_path(self, dir_type='', filename=None):
paths = self.params.get('paths', {})
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 '',
filename or '')
return sanitize_path(path, force=self.params.get('windowsfilenames'))
@staticmethod
def _outtmpl_expandpath(outtmpl):
# expand_path translates '%%' into '%' and '$$' into '$'
# correspondingly that is not what we want since we need to keep
# '%%' intact for template dict substitution step. Working around
# with boundary-alike separator hack.
sep = ''.join([random.choice(ascii_letters) for _ in range(32)])
outtmpl = outtmpl.replace('%%', f'%{sep}%').replace('$$', f'${sep}$')
# outtmpl should be expand_path'ed before template dict substitution
# because meta fields may contain env variables we don't want to
# 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, '')
@staticmethod
def escape_outtmpl(outtmpl):
''' Escape any remaining strings like %s, %abc% etc. '''
return re.sub(
STR_FORMAT_RE_TMPL.format('', '(?![%(\0])'),
lambda mobj: ('' if mobj.group('has_key') else '%') + mobj.group(0),
outtmpl)
@classmethod
def validate_outtmpl(cls, outtmpl):
''' @return None or Exception object '''
outtmpl = re.sub(
STR_FORMAT_RE_TMPL.format('[^)]*', '[ljhqBUDS]'),
lambda mobj: f'{mobj.group(0)[:-1]}s',
cls._outtmpl_expandpath(outtmpl))
try:
cls.escape_outtmpl(outtmpl) % collections.defaultdict(int)
return None
except ValueError as err:
return err
@staticmethod
def _copy_infodict(info_dict):
info_dict = dict(info_dict)
info_dict.pop('__postprocessors', None)
info_dict.pop('__pending_error', None)
return info_dict
def prepare_outtmpl(self, outtmpl, info_dict, sanitize=False):
""" Make the outtmpl and info_dict suitable for substitution: ydl.escape_outtmpl(outtmpl) % info_dict
@param sanitize Whether to sanitize the output as a filename.
For backward compatibility, a function can also be passed
"""
info_dict.setdefault('epoch', int(time.time())) # keep epoch consistent once set
info_dict = self._copy_infodict(info_dict)
info_dict['duration_string'] = ( # %(duration>%H-%M-%S)s is wrong if duration > 24hrs
formatSeconds(info_dict['duration'], '-' if sanitize else ':')
if info_dict.get('duration', None) is not None
else None)
info_dict['autonumber'] = int(self.params.get('autonumber_start', 1) - 1 + self._num_downloads)
info_dict['video_autonumber'] = self._num_videos
if info_dict.get('resolution') is None:
info_dict['resolution'] = self.format_resolution(info_dict, default=None)
# For fields playlist_index, playlist_autonumber and autonumber convert all occurrences
# of %(field)s to %(field)0Nd for backward compatibility
field_size_compat_map = {
'playlist_index': number_of_digits(info_dict.get('__last_playlist_index') or 0),
'playlist_autonumber': number_of_digits(info_dict.get('n_entries') or 0),
'autonumber': self.params.get('autonumber_size') or 5,
}
TMPL_DICT = {}
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, 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)
(?P<negate>-)?
(?P<fields>{FIELD_RE})
(?P<maths>(?:{MATH_OPERATORS_RE}{MATH_FIELD_RE})*)
(?:>(?P<strf_format>.+?))?
(?P<remaining>
(?P<alternate>(?<!\\),[^|&)]+)?
(?:&(?P<replacement>.*?))?
(?:\|(?P<default>.*?))?
)$''')
def _traverse_infodict(fields):
fields = [f for x in re.split(r'\.({.+?})\.?', fields)
for f in ([x] if x.startswith('{') else x.split('.'))]
for i in (0, -1):
if fields and not fields[i]:
fields.pop(i)
for i, f in enumerate(fields):
if not f.startswith('{'):
continue
assert f.endswith('}'), f'No closing brace for {f} in {fields}'
fields[i] = {k: k.split('.') for k in f[1:-1].split(',')}
return traverse_obj(info_dict, fields, is_user_input=True, traverse_string=True)
def get_value(mdict):
# Object traversal
value = _traverse_infodict(mdict['fields'])
# Negative
if mdict['negate']:
value = float_or_none(value)
if value is not None:
value *= -1
# Do maths
offset_key = mdict['maths']
if offset_key:
value = float_or_none(value)
operator = None
while offset_key:
item = re.match(
MATH_FIELD_RE if operator else MATH_OPERATORS_RE,
offset_key).group(0)
offset_key = offset_key[len(item):]
if operator is None:
operator = MATH_FUNCTIONS[item]
continue
item, multiplier = (item[1:], -1) if item[0] == '-' else (item, 1)
offset = float_or_none(item)
if offset is None:
offset = float_or_none(_traverse_infodict(item))
try:
value = operator(value, multiplier * offset)
except (TypeError, ZeroDivisionError):
return None
operator = None
# Datetime formatting
if mdict['strf_format']:
value = strftime_or_none(value, mdict['strf_format'].replace('\\,', ','))
# XXX: Workaround for https://github.com/yt-dlp/yt-dlp/issues/4485
if sanitize and value == '':
value = None
return value
na = self.params.get('outtmpl_na_placeholder', 'NA')
def filename_sanitizer(key, value, restricted=self.params.get('restrictfilenames')):
return sanitize_filename(str(value), restricted=restricted, is_id=(
bool(re.search(r'(^|[_.])id(\.|$)', key))
if 'filename-sanitization' in self.params['compat_opts']
else NO_DEFAULT))
sanitizer = sanitize if callable(sanitize) else filename_sanitizer
sanitize = bool(sanitize)
def _dumpjson_default(obj):
if isinstance(obj, (set, LazyList)):
return list(obj)
return repr(obj)
def create_key(outer_mobj):
if not outer_mobj.group('has_key'):
return outer_mobj.group(0)
key = outer_mobj.group('key')
mobj = re.match(INTERNAL_FORMAT_RE, key)
initial_field = mobj.group('fields') if mobj else ''
value, replacement, default = None, None, na
while mobj:
mobj = mobj.groupdict()
default = mobj['default'] if mobj['default'] is not None else default
value = get_value(mobj)
replacement = mobj['replacement']
if value is None and mobj['alternate']:
mobj = re.match(INTERNAL_FORMAT_RE, mobj['remaining'][1:])
else:
break
fmt = outer_mobj.group('format')
if fmt == 's' and value is not None and key in field_size_compat_map.keys():
fmt = f'0{field_size_compat_map[key]:d}d'
value = default if value is None else value if replacement is None else replacement
flags = outer_mobj.group('conversion') or ''
str_fmt = f'{fmt[:-1]}s'
if fmt[-1] == 'l': # list
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, ensure_ascii=False), 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
elif fmt[-1] == 'B': # bytes
value = f'%{str_fmt}'.encode() % str(value).encode()
value, fmt = value.decode('utf-8', 'ignore'), 's'
elif fmt[-1] == 'U': # unicode normalized
value, fmt = unicodedata.normalize(
# "+" = compatibility equivalence, "#" = NFD
'NF%s%s' % ('K' if '+' in flags else '', 'D' if '#' in flags else 'C'),
value), str_fmt
elif fmt[-1] == 'D': # decimal suffix
num_fmt, fmt = fmt[:-1].replace('#', ''), 's'
value = format_decimal_suffix(value, f'%{num_fmt}f%s' if num_fmt else '%d%s',
factor=1024 if '#' in flags else 1000)
elif fmt[-1] == 'S': # filename sanitization
value, fmt = filename_sanitizer(initial_field, value, restricted='#' in flags), str_fmt
elif fmt[-1] == 'c':
if value:
value = str(value)[0]
else:
fmt = str_fmt
elif fmt[-1] not in 'rs': # numeric
value = float_or_none(value)
if value is None:
value, fmt = default, 's'
if sanitize:
if fmt[-1] == 'r':
# If value is an object, sanitize might convert it to a string
# So we convert it to repr first
value, fmt = repr(value), str_fmt
if fmt[-1] in 'csr':
value = sanitizer(initial_field, value)
key = '%s\0%s' % (key.replace('%', '%\0'), outer_mobj.group('format'))
TMPL_DICT[key] = value
return '{prefix}%({key}){fmt}'.format(key=key, fmt=fmt, prefix=outer_mobj.group('prefix'))
return EXTERNAL_FORMAT_RE.sub(create_key, outtmpl), TMPL_DICT
def evaluate_outtmpl(self, outtmpl, info_dict, *args, **kwargs):
outtmpl, info_dict = self.prepare_outtmpl(outtmpl, info_dict, *args, **kwargs)
return self.escape_outtmpl(outtmpl) % info_dict
def _prepare_filename(self, info_dict, *, outtmpl=None, tmpl_type=None):
assert None in (outtmpl, tmpl_type), 'outtmpl and tmpl_type are mutually exclusive'
if outtmpl is None:
outtmpl = self.params['outtmpl'].get(tmpl_type or 'default', self.params['outtmpl']['default'])
try:
outtmpl = self._outtmpl_expandpath(outtmpl)
filename = self.evaluate_outtmpl(outtmpl, info_dict, True)
if not filename:
return None
if tmpl_type in ('', 'temp'):
final_ext, ext = self.params.get('final_ext'), info_dict.get('ext')
if final_ext and ext and final_ext != ext and filename.endswith(f'.{final_ext}'):
filename = replace_extension(filename, ext, final_ext)
elif tmpl_type:
force_ext = OUTTMPL_TYPES[tmpl_type]
if force_ext:
filename = replace_extension(filename, force_ext, info_dict.get('ext'))
# https://github.com/blackjack4494/youtube-dlc/issues/85
trim_file_name = self.params.get('trim_file_name', False)
if trim_file_name:
no_ext, *ext = filename.rsplit('.', 2)
filename = join_nonempty(no_ext[:trim_file_name], *ext, delim='.')
return filename
except ValueError as err:
self.report_error('Error in output template: ' + str(err) + ' (encoding: ' + repr(preferredencoding()) + ')')
return None
def prepare_filename(self, info_dict, dir_type='', *, outtmpl=None, warn=False):
"""Generate the output filename"""
if outtmpl:
assert not dir_type, 'outtmpl and dir_type are mutually exclusive'
dir_type = None
filename = self._prepare_filename(info_dict, tmpl_type=dir_type, outtmpl=outtmpl)
if not filename and dir_type not in ('', 'temp'):
return ''
if warn:
if not self.params.get('paths'):
pass
elif filename == '-':
self.report_warning('--paths is ignored when an outputting to stdout', only_once=True)
elif os.path.isabs(filename):
self.report_warning('--paths is ignored since an absolute path is given in output template', only_once=True)
if filename == '-' or not filename:
return filename
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 """
video_title = info_dict.get('title', info_dict.get('id', 'entry'))
def check_filter():
if 'title' in info_dict:
# This can happen when we're just evaluating the playlist
title = info_dict['title']
matchtitle = self.params.get('matchtitle', False)
if matchtitle:
if not re.search(matchtitle, title, re.IGNORECASE):
return '"' + title + '" title did not match pattern "' + matchtitle + '"'
rejecttitle = self.params.get('rejecttitle', False)
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())
if date not in dateRange:
return f'{date_from_str(date).isoformat()} upload date is not in range {dateRange}'
view_count = info_dict.get('view_count')
if view_count is not None:
min_views = self.params.get('min_views')
if min_views is not None and view_count < min_views:
return 'Skipping %s, because it has not reached minimum view count (%d/%d)' % (video_title, view_count, min_views)
max_views = self.params.get('max_views')
if max_views is not None and view_count > max_views:
return 'Skipping %s, because it has exceeded the maximum view count (%d/%d)' % (video_title, view_count, max_views)
if age_restricted(info_dict.get('age_limit'), self.params.get('age_limit')):
return 'Skipping "%s" because it is age restricted' % video_title
match_filter = self.params.get('match_filter')
if match_filter is not None:
try:
ret = match_filter(info_dict, incomplete=incomplete)
except TypeError:
# For backward compatibility
ret = None if incomplete else match_filter(info_dict)
if ret is NO_DEFAULT:
while True:
filename = self._format_screen(self.prepare_filename(info_dict), self.Styles.FILENAME)
reply = input(self._format_screen(
f'Download "{filename}"? (Y/n): ', self.Styles.EMPHASIS)).lower().strip()
if reply in {'y', ''}:
return None
elif reply == 'n':
return f'Skipping {video_title}'
elif ret is not None:
return ret
return None
if self.in_download_archive(info_dict):
reason = '%s has already been recorded in the archive' % video_title
break_opt, break_err = 'break_on_existing', ExistingVideoReached
else:
reason = check_filter()
break_opt, break_err = 'break_on_reject', RejectedVideoReached
if reason is not None:
if not silent:
self.to_screen('[download] ' + reason)
if self.params.get(break_opt, False):
raise break_err()
return reason
@staticmethod
def add_extra_info(info_dict, extra_info):
'''Set the keys from extra_info in info dict if they are missing'''
for key, value in extra_info.items():
info_dict.setdefault(key, value)
def extract_info(self, url, download=True, ie_key=None, extra_info=None,
process=True, force_generic_extractor=False):
"""
Extract and return the information dictionary of the URL
Arguments:
@param url URL to extract
Keyword arguments:
@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:
extra_info = {}
if not ie_key and force_generic_extractor:
ie_key = 'Generic'
if ie_key:
ies = {ie_key: self._ies[ie_key]} if ie_key in self._ies else {}
else:
ies = self._ies
for key, ie in ies.items():
if not ie.suitable(url):
continue
if not ie.working():
self.report_warning('The program functionality for this site has been marked as broken, '
'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': 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(key), download, extra_info, process)
else:
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)
def wrapper(self, *args, **kwargs):
while True:
try:
return func(self, *args, **kwargs)
except (DownloadCancelled, LazyList.IndexError, PagedList.IndexError):
raise
except ReExtractInfo as e:
if e.expected:
self.to_screen(f'{e}; Re-extracting data')
else:
self.to_stderr('\r')
self.report_warning(f'{e}; Re-extracting data')
continue
except GeoRestrictedError as e:
msg = e.msg
if e.countries:
msg += '\nThis video is available in %s.' % ', '.join(
map(ISO3166Utils.short2full, e.countries))
msg += '\nYou might want to use a VPN or a proxy server (with --proxy) to workaround.'
self.report_error(msg)
except ExtractorError as e: # An error we somewhat expected
self.report_error(str(e), e.format_traceback())
except Exception as e:
if self.params.get('ignoreerrors'):
self.report_error(str(e), tb=encode_compat_str(traceback.format_exc()))
else:
raise
break
return wrapper
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')):
return
format_dur = lambda dur: '%02d:%02d:%02d' % timetuple_from_msec(dur * 1000)[:-1]
last_msg = ''
def progress(msg):
nonlocal last_msg
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')
diff = try_get(ie_result, lambda x: x['release_timestamp'] - time.time())
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 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')
wait_till = time.time() + diff
try:
while True:
diff = wait_till - time.time()
if diff <= 0:
progress('')
raise ReExtractInfo('[wait] Wait period ended', expected=True)
progress(f'[wait] Remaining time until next attempt: {self._format_screen(format_dur(diff), self.Styles.EMPHASIS)}')
time.sleep(1)
except KeyboardInterrupt:
progress('')
raise ReExtractInfo('[wait] Interrupted by user', expected=True)
except BaseException as e:
if not isinstance(e, ReExtractInfo):
self.to_screen('')
raise
@_handle_extraction_exceptions
def __extract_info(self, url, ie, download, extra_info, process):
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
ie_result = {
'_type': 'compat_list',
'entries': ie_result,
}
if extra_info.get('original_url'):
ie_result.setdefault('original_url', extra_info['original_url'])
self.add_default_extra_info(ie_result, ie, url)
if process:
self._wait_for_video(ie_result)
return self.process_ie_result(ie_result, download, extra_info)
else:
return ie_result
def add_default_extra_info(self, ie_result, ie, url):
if url is not None:
self.add_extra_info(ie_result, {
'webpage_url': url,
'original_url': url,
})
webpage_url = ie_result.get('webpage_url')
if webpage_url:
self.add_extra_info(ie_result, {
'webpage_url_basename': url_basename(webpage_url),
'webpage_url_domain': get_domain(webpage_url),
})
if ie is not None:
self.add_extra_info(ie_result, {
'extractor': ie.IE_NAME,
'extractor_key': ie.ie_key(),
})
def process_ie_result(self, ie_result, download=True, extra_info=None):
"""
Take the result of the ie(may be modified) and resolve all unresolved
references (URLs, playlist items).
It will also download the videos if 'download'.
Returns the resolved ie_result.
"""
if extra_info is None:
extra_info = {}
result_type = ie_result.get('_type', 'video')
if result_type in ('url', 'url_transparent'):
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'])
extract_flat = self.params.get('extract_flat', False)
if ((extract_flat == 'in_playlist' and 'playlist' in extra_info)
or extract_flat is True):
info_copy = ie_result.copy()
ie = try_get(ie_result.get('ie_key'), self.get_info_extractor)
if ie and not ie_result.get('id'):
info_copy['id'] = ie.get_temp_id(ie_result['url'])
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.__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):
self.record_download_archive(info_copy)
return ie_result
if result_type == 'video':
self.add_extra_info(ie_result, extra_info)
ie_result = self.process_video_result(ie_result, download=download)
self._raise_pending_errors(ie_result)
additional_urls = (ie_result or {}).get('additional_urls')
if additional_urls:
# TODO: Improve MetadataParserPP to allow setting a list
if isinstance(additional_urls, str):
additional_urls = [additional_urls]
self.to_screen(
'[info] %s: %d additional URL(s) requested' % (ie_result['id'], len(additional_urls)))
self.write_debug('Additional URLs: "%s"' % '", "'.join(additional_urls))
ie_result['additional_entries'] = [
self.extract_info(
url, download, extra_info=extra_info,
force_generic_extractor=self.params.get('force_generic_extractor'))
for url in additional_urls
]
return ie_result
elif result_type == 'url':
# We have to add extra_info to the results because it may be
# contained in a playlist
return self.extract_info(
ie_result['url'], download,
ie_key=ie_result.get('ie_key'),
extra_info=extra_info)
elif result_type == 'url_transparent':
# Use the information from the embedding page
info = self.extract_info(
ie_result['url'], ie_key=ie_result.get('ie_key'),
extra_info=extra_info, download=False, process=False)
# extract_info may return None when ignoreerrors is enabled and
# extraction failed with an error, don't crash and return early
# in this case
if not info:
return info
exempted_fields = {'_type', 'url', 'ie_key'}
if not ie_result.get('section_end') and ie_result.get('section_start') is None:
# For video clips, the id etc of the clip extractor should be used
exempted_fields |= {'id', 'extractor', 'extractor_key'}
new_result = info.copy()
new_result.update(filter_dict(ie_result, lambda k, v: v is not None and k not in exempted_fields))
# Extracted info may not be a video result (i.e.
# info.get('_type', 'video') != video) but rather an url or
# url_transparent. In such cases outer metadata (from ie_result)
# should be propagated to inner one (info). For this to happen
# _type of info should be overridden with url_transparent. This
# fixes issue from https://github.com/ytdl-org/youtube-dl/pull/11163.
if new_result.get('_type') == 'url':
new_result['_type'] = 'url_transparent'
return self.process_ie_result(
new_result, download=download, extra_info=extra_info)
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.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'))
return
self._playlist_level += 1
self._playlist_urls.add(webpage_url)
self._fill_common_fields(ie_result, False)
self._sanitize_thumbnails(ie_result)
try:
return self.__process_playlist(ie_result, download)
finally:
self._playlist_level -= 1
if not self._playlist_level:
self._playlist_urls.clear()
elif result_type == 'compat_list':
self.report_warning(
'Extractor %s returned a compat_list result. '
'It needs to be updated.' % ie_result.get('extractor'))
def _fixup(r):
self.add_extra_info(r, {
'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'],
})
return r
ie_result['entries'] = [
self.process_ie_result(_fixup(r), download, extra_info)
for r in ie_result['entries']
]
return ie_result
else:
raise Exception('Invalid result type: %s' % result_type)
def _ensure_dir_exists(self, path):
return make_dir(path, self.report_error)
@staticmethod
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'),
**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"""
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)
lazy = self.params.get('lazy_playlist')
if lazy:
resolved_entries, n_entries = [], 'N/A'
ie_result['requested_entries'], ie_result['entries'] = None, None
else:
entries = resolved_entries = list(entries)
n_entries = len(resolved_entries)
ie_result['requested_entries'], ie_result['entries'] = tuple(zip(*resolved_entries)) or ([], [])
if not ie_result.get('playlist_count'):
# 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'):
_infojson_written = self._write_info_json(
'playlist', ie_result, self.prepare_filename(ie_copy, 'pl_infojson'))
if _infojson_written is None:
return
if self._write_description('playlist', ie_result,
self.prepare_filename(ie_copy, 'pl_description')) is None:
return
# TODO: This should be passed to ThumbnailsConvertor if necessary
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'):
self.report_warning('playlistreverse and playlistrandom are not supported with lazy_playlist', only_once=True)
elif self.params.get('playlistreverse'):
entries.reverse()
elif self.params.get('playlistrandom'):
random.shuffle(entries)
self.to_screen(f'[{ie_result["extractor"]}] Playlist {title}: Downloading {n_entries} videos'
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))
if not entry:
continue
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_copy = collections.ChainMap(entry, {
**common_info,
'n_entries': int_or_none(n_entries),
'playlist_index': playlist_index,
'playlist_autonumber': i + 1,
})
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 video %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
if keep_resolved_entries:
resolved_entries[i] = (playlist_index, entry_result)
# Update with processed data
ie_result['requested_entries'] = [i for i, e in resolved_entries if e is not NO_DEFAULT]
ie_result['entries'] = [e for _, e in resolved_entries if e is not NO_DEFAULT]
# Write the updated info to json
if _infojson_written is True and self._write_info_json(
'updated playlist', ie_result,
self.prepare_filename(ie_copy, 'pl_infojson'), overwrite=True) is None:
return
ie_result = self.run_all_pps('playlist', ie_result)
self.to_screen(f'[download] Finished downloading playlist: {title}')
return ie_result
@_handle_extraction_exceptions
def __process_iterable_entry(self, entry, download, extra_info):
return self.process_ie_result(
entry, download=download, extra_info=extra_info)
def _build_format_filter(self, filter_spec):
" Returns a function to filter the formats according to the filter_spec "
OPERATORS = {
'<': operator.lt,
'<=': operator.le,
'>': operator.gt,
'>=': operator.ge,
'=': operator.eq,
'!=': operator.ne,
}
operator_rex = re.compile(r'''(?x)\s*
(?P<key>width|height|tbr|abr|vbr|asr|filesize|filesize_approx|fps)\s*
(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
(?P<value>[0-9.]+(?:[kKmMgGtTpPeEzZyY]i?[Bb]?)?)\s*
''' % '|'.join(map(re.escape, OPERATORS.keys())))
m = operator_rex.fullmatch(filter_spec)
if m:
try:
comparison_value = int(m.group('value'))
except ValueError:
comparison_value = parse_filesize(m.group('value'))
if comparison_value is None:
comparison_value = parse_filesize(m.group('value') + 'B')
if comparison_value is None:
raise ValueError(
'Invalid value %r in format specification %r' % (
m.group('value'), filter_spec))
op = OPERATORS[m.group('op')]
if not m:
STR_OPERATORS = {
'=': operator.eq,
'^=': lambda attr, value: attr.startswith(value),
'$=': lambda attr, value: attr.endswith(value),
'*=': lambda attr, value: value in attr,
'~=': lambda attr, value: value.search(attr) is not None
}
str_operator_rex = re.compile(r'''(?x)\s*
(?P<key>[a-zA-Z0-9._-]+)\s*
(?P<negation>!\s*)?(?P<op>%s)\s*(?P<none_inclusive>\?\s*)?
(?P<quote>["'])?
(?P<value>(?(quote)(?:(?!(?P=quote))[^\\]|\\.)+|[\w.-]+))
(?(quote)(?P=quote))\s*
''' % '|'.join(map(re.escape, STR_OPERATORS.keys())))
m = str_operator_rex.fullmatch(filter_spec)
if m:
if m.group('op') == '~=':
comparison_value = re.compile(m.group('value'))
else:
comparison_value = re.sub(r'''\\([\\"'])''', r'\1', m.group('value'))
str_op = STR_OPERATORS[m.group('op')]
if m.group('negation'):
op = lambda attr, value: not str_op(attr, value)
else:
op = str_op
if not m:
raise SyntaxError('Invalid filter specification %r' % filter_spec)
def _filter(f):
actual_value = f.get(m.group('key'))
if actual_value is None:
return m.group('none_inclusive')
return op(actual_value, comparison_value)
return _filter
def _check_formats(self, formats):
for f in formats:
self.to_screen('[info] Testing format %s' % f['format_id'])
path = self.get_output_path('temp')
if not self._ensure_dir_exists(f'{path}/'):
continue
temp_file = tempfile.NamedTemporaryFile(suffix='.tmp', delete=False, dir=path or None)
temp_file.close()
try:
success, _ = self.dl(temp_file.name, f, test=True)
except (DownloadError, OSError, ValueError) + network_exceptions:
success = False
finally:
if os.path.exists(temp_file.name):
try:
os.remove(temp_file.name)
except OSError:
self.report_warning('Unable to delete temporary file "%s"' % temp_file.name)
if success:
yield f
else:
self.to_screen('[info] Unable to download format %s. Skipping...' % f['format_id'])
def _default_format_spec(self, info_dict, download=True):
def can_merge():
merger = FFmpegMergerPP(self)
return merger.available and merger.can_merge()
prefer_best = (
not self.params.get('simulate')
and download
and (
not can_merge()
or info_dict.get('is_live') and not self.params.get('live_from_start')
or self.params['outtmpl']['default'] == '-'))
compat = (
prefer_best
or self.params.get('allow_multiple_audio_streams', False)
or 'format-spec' in self.params['compat_opts'])
return (
'best/bestvideo+bestaudio' if prefer_best
else 'bestvideo*+bestaudio/best' if not compat
else 'bestvideo+bestaudio/best')
def build_format_selector(self, format_spec):
def syntax_error(note, start):
message = (
'Invalid format specification: '
'{}\n\t{}\n\t{}^'.format(note, format_spec, ' ' * start[1]))
return SyntaxError(message)
PICKFIRST = 'PICKFIRST'
MERGE = 'MERGE'
SINGLE = 'SINGLE'
GROUP = 'GROUP'
FormatSelector = collections.namedtuple('FormatSelector', ['type', 'selector', 'filters'])
allow_multiple_streams = {'audio': self.params.get('allow_multiple_audio_streams', False),
'video': self.params.get('allow_multiple_video_streams', False)}
check_formats = self.params.get('check_formats') == 'selected'
def _parse_filter(tokens):
filter_parts = []
for type, string, start, _, _ in tokens:
if type == tokenize.OP and string == ']':
return ''.join(filter_parts)
else:
filter_parts.append(string)
def _remove_unused_ops(tokens):
# 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:
if type == tokenize.OP and string == '[':
if last_string:
yield tokenize.NAME, last_string, last_start, last_end, last_line
last_string = None
yield type, string, start, end, line
# everything inside brackets will be handled by _parse_filter
for type, string, start, end, line in tokens:
yield type, string, start, end, line
if type == tokenize.OP and string == ']':
break
elif type == tokenize.OP and string in ALLOWED_OPS:
if last_string:
yield tokenize.NAME, last_string, last_start, last_end, last_line
last_string = None
yield type, string, start, end, line
elif type in [tokenize.NAME, tokenize.NUMBER, tokenize.OP]:
if not last_string:
last_string = string
last_start = start
last_end = end
else:
last_string += string
if last_string:
yield tokenize.NAME, last_string, last_start, last_end, last_line
def _parse_format_selection(tokens, inside_merge=False, inside_choice=False, inside_group=False):
selectors = []
current_selector = None
for type, string, start, _, _ in tokens:
# ENCODING is only defined in python 3.x
if type == getattr(tokenize, 'ENCODING', None):
continue
elif type in [tokenize.NAME, tokenize.NUMBER]:
current_selector = FormatSelector(SINGLE, string, [])
elif type == tokenize.OP:
if string == ')':
if not inside_group:
# ')' will be handled by the parentheses group
tokens.restore_last_token()
break
elif inside_merge and string in ['/', ',']:
tokens.restore_last_token()
break
elif inside_choice and string == ',':
tokens.restore_last_token()
break
elif string == ',':
if not current_selector:
raise syntax_error('"," must follow a format selector', start)
selectors.append(current_selector)
current_selector = None
elif string == '/':
if not current_selector:
raise syntax_error('"/" must follow a format selector', start)
first_choice = current_selector
second_choice = _parse_format_selection(tokens, inside_choice=True)
current_selector = FormatSelector(PICKFIRST, (first_choice, second_choice), [])
elif string == '[':
if not current_selector:
current_selector = FormatSelector(SINGLE, 'best', [])
format_filter = _parse_filter(tokens)
current_selector.filters.append(format_filter)
elif string == '(':
if current_selector:
raise syntax_error('Unexpected "("', start)
group = _parse_format_selection(tokens, inside_group=True)
current_selector = FormatSelector(GROUP, group, [])
elif string == '+':
if not current_selector:
raise syntax_error('Unexpected "+"', start)
selector_1 = current_selector
selector_2 = _parse_format_selection(tokens, inside_merge=True)
if not selector_2:
raise syntax_error('Expected a selector', start)
current_selector = FormatSelector(MERGE, (selector_1, selector_2), [])
else:
raise syntax_error(f'Operator not recognized: "{string}"', start)
elif type == tokenize.ENDMARKER:
break
if current_selector:
selectors.append(current_selector)
return selectors
def _merge(formats_pair):
format_1, format_2 = formats_pair
formats_info = []
formats_info.extend(format_1.get('requested_formats', (format_1,)))
formats_info.extend(format_2.get('requested_formats', (format_2,)))
if not allow_multiple_streams['video'] or not allow_multiple_streams['audio']:
get_no_more = {'video': False, 'audio': False}
for (i, fmt_info) in enumerate(formats_info):
if fmt_info.get('acodec') == fmt_info.get('vcodec') == 'none':
formats_info.pop(i)
continue
for aud_vid in ['audio', 'video']:
if not allow_multiple_streams[aud_vid] and fmt_info.get(aud_vid[0] + 'codec') != 'none':
if get_no_more[aud_vid]:
formats_info.pop(i)
break
get_no_more[aud_vid] = True
if len(formats_info) == 1:
return formats_info[0]
video_fmts = [fmt_info for fmt_info in formats_info if fmt_info.get('vcodec') != 'none']
audio_fmts = [fmt_info for fmt_info in formats_info if fmt_info.get('acodec') != 'none']
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 = 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))
new_dict = {
'requested_formats': formats_info,
'format': '+'.join(filtered('format')),
'format_id': '+'.join(filtered('format_id')),
'ext': output_ext,
'protocol': '+'.join(map(determine_protocol, formats_info)),
'language': '+'.join(orderedSet(filtered('language'))) or None,
'format_note': '+'.join(orderedSet(filtered('format_note'))) or None,
'filesize_approx': sum(filtered('filesize', 'filesize_approx')) or None,
'tbr': sum(filtered('tbr', 'vbr', 'abr')),
}
if the_only_video:
new_dict.update({
'width': the_only_video.get('width'),
'height': the_only_video.get('height'),
'resolution': the_only_video.get('resolution') or self.format_resolution(the_only_video),
'fps': the_only_video.get('fps'),
'dynamic_range': the_only_video.get('dynamic_range'),
'vcodec': the_only_video.get('vcodec'),
'vbr': the_only_video.get('vbr'),
'stretched_ratio': the_only_video.get('stretched_ratio'),
})
if the_only_audio:
new_dict.update({
'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
def _check_formats(formats):
if not check_formats:
yield from formats
return
yield from self._check_formats(formats)
def _build_selector_function(selector):
if isinstance(selector, list): # ,
fs = [_build_selector_function(s) for s in selector]
def selector_function(ctx):
for f in fs:
yield from f(ctx)
return selector_function
elif selector.type == GROUP: # ()
selector_function = _build_selector_function(selector.selector)
elif selector.type == PICKFIRST: # /
fs = [_build_selector_function(s) for s in selector.selector]
def selector_function(ctx):
for f in fs:
picked_formats = list(f(ctx))
if picked_formats:
return picked_formats
return []
elif selector.type == MERGE: # +
selector_1, selector_2 = map(_build_selector_function, selector.selector)
def selector_function(ctx):
for pair in itertools.product(selector_1(ctx), selector_2(ctx)):
yield _merge(pair)
elif selector.type == SINGLE: # atom
format_spec = selector.selector or 'best'
# TODO: Add allvideo, allaudio etc by generalizing the code with best/worst selector
if format_spec == 'all':
def selector_function(ctx):
yield from _check_formats(ctx['formats'][::-1])
elif format_spec == 'mergeall':
def selector_function(ctx):
formats = list(_check_formats(
f for f in ctx['formats'] if f.get('vcodec') != 'none' or f.get('acodec') != 'none'))
if not formats:
return
merged_format = formats[-1]
for f in formats[-2::-1]:
merged_format = _merge((merged_format, f))
yield merged_format
else:
format_fallback, seperate_fallback, format_reverse, format_idx = False, None, True, 1
mobj = re.match(
r'(?P<bw>best|worst|b|w)(?P<type>video|audio|v|a)?(?P<mod>\*)?(?:\.(?P<n>[1-9]\d*))?$',
format_spec)
if mobj is not None:
format_idx = int_or_none(mobj.group('n'), default=1)
format_reverse = mobj.group('bw')[0] == 'b'
format_type = (mobj.group('type') or [None])[0]
not_format_type = {'v': 'a', 'a': 'v'}.get(format_type)
format_modified = mobj.group('mod') is not None
format_fallback = not format_type and not format_modified # for b, w
_filter_f = (
(lambda f: f.get('%scodec' % format_type) != 'none')
if format_type and format_modified # bv*, ba*, wv*, wa*
else (lambda f: f.get('%scodec' % not_format_type) == 'none')
if format_type # bv, ba, wv, wa
else (lambda f: f.get('vcodec') != 'none' and f.get('acodec') != 'none')
if not format_modified # b, w
else lambda f: True) # b*, w*
filter_f = lambda f: _filter_f(f) and (
f.get('vcodec') != 'none' or f.get('acodec') != 'none')
else:
if format_spec in self._format_selection_exts['audio']:
filter_f = lambda f: f.get('ext') == format_spec and f.get('acodec') != 'none'
elif format_spec in self._format_selection_exts['video