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.

3626 lines
165 KiB

#!/usr/bin/env python3
# coding: utf-8
from __future__ import absolute_import, unicode_literals
import collections
import contextlib
import copy
import datetime
import errno
import fileinput
import functools
import io
import itertools
import json
import locale
import operator
import os
import platform
import re
import shutil
import subprocess
import sys
import tempfile
import time
import tokenize
import traceback
import random
import unicodedata
from enum import Enum
from string import ascii_letters
from .compat import (
from .cookies import load_cookies
from .utils import (
from .cache import Cache
from .minicurses import format_text
from .extractor import (
_PLUGIN_CLASSES as plugin_extractors
from .extractor.openload import PhantomJSwrapper
from .downloader import (
from .downloader.rtmp import rtmpdump_version
from .postprocessor import (
_PLUGIN_CLASSES as plugin_postprocessors
from .update import detect_variant
from .version import __version__
if compat_os_name == 'nt':
import ctypes
class YoutubeDL(object):
"""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 list of templates to force print
forceurl: Force printing final URL. (Deprecated)
forcetitle: Force printing title. (Deprecated)
forceid: Force printing ID. (Deprecated)
forcethumbnail: Force printing thumbnail URL. (Deprecated)
forcedescription: Force printing description. (Deprecated)
forcefilename: Force printing final filename. (Deprecated)
forceduration: Force printing duration. (Deprecated)
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.
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
outtmpl: Dictionary of templates for output names. Allowed keys
are 'default' and the keys of OUTTMPL_TYPES (in
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
force_generic_extractor: Force downloader to use the generic extractor
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
playliststart: Playlist item to start at.
playlistend: Playlist item to end at.
playlist_items: Specific indices of playlist to download.
playlistreverse: Download playlist items in reverse order.
playlistrandom: Download playlist items in random order.
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.
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
allsubtitles: Deprecated - Use subtitleslangs = ['all']
Downloads all the subtitles of the video
(requires writesubtitles or writeautomaticsub)
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. Eg: ['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
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
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.
cookiefile: File name where cookies should be read from and dumped to
cookiesfrombrowser: A tuple containing the name of the browser and the profile
name/path from where cookies are loaded.
Eg: ('chrome', ) or ('vivaldi', 'default')
nocheckcertificate:Do not verify SSL certificates
prefer_insecure: Use HTTP instead of HTTPS to retrieve information.
At the moment, this is only supported by YouTube.
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
include_ads: Download ads as well
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.
postprocessors: A list of dictionaries, each with an entry
* key: The name of the postprocessor. See
yt_dlp/postprocessor/ for a list.
* when: When to run the postprocessor. Can be one of
Assumed to be 'post_process' if not given
post_hooks: Deprecated - Register a custom postprocessor instead
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.
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
* 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: Extension 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.
call_home: Boolean, true iff we are allowed to contact the
yt-dlp servers for debugging. (BROKEN)
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: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 with the info_dict of
every video.
If it returns a message, the video is ignored.
If it returns None, the video is downloaded.
match_filter_func in 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
Two-letter ISO 3166-2 country code that will be used for
explicit geographic restriction bypassing via faking
X-Forwarded-For HTTP header
IP range in CIDR notation that will be used similarly to
The following options determine which downloader is picked:
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
hls_prefer_native: Deprecated - Use external_downloader = {'m3u8': 'native'}
or {'m3u8': 'ffmpeg'} instead.
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.
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 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'
The following parameters are not used by YoutubeDL itself, they are used by
the downloader (see yt_dlp/downloader/
nopart, updatetime, buffersize, ratelimit, throttledratelimit, min_filesize,
max_filesize, test, noresizebuffer, retries, fragment_retries, continuedl,
noprogress, xattr_set_filesize, hls_use_mpegts, http_chunk_size,
external_downloader_args, concurrent_fragment_downloads.
The following options are used by the post processors:
prefer_ffmpeg: If False, use avconv instead of ffmpeg if both are available,
otherwise prefer ffmpeg. (avconv support is deprecated)
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.
Eg: {'youtube': {'skip': ['dash', 'hls']}}
youtube_include_dash_manifest: Deprecated - Use extractor_args instead.
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: Deprecated - Use extractor_args instead.
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)
'width', 'height', 'tbr', 'abr', 'asr', 'vbr', 'fps', '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_selection_exts = {
'audio': {'m4a', 'mp3', 'ogg', 'aac'},
'video': {'mp4', 'flv', 'webm', '3gp'},
'storyboards': {'mhtml'},
params = None
_ies = {}
_pps = {'pre_process': [], 'before_dl': [], 'after_move': [], 'post_process': []}
_printed_messages = set()
_first_webpage_request = True
_download_retcode = None
_num_downloads = None
_playlist_level = 0
_playlist_urls = set()
_screen_file = None
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._ies = {}
self._ies_instances = {}
self._pps = {'pre_process': [], 'before_dl': [], 'after_move': [], 'post_process': []}
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._screen_file = [sys.stdout, sys.stderr][params.get('logtostderr', False)]
self._err_file = sys.stderr
self.params = params
self.cache = Cache(self)
self._allow_colors = {
'screen': not self.params.get('no_color') and supports_terminal_sequences(self._screen_file),
'err': not self.params.get('no_color') and supports_terminal_sequences(self._err_file),
if sys.version_info < (3, 6):
'Python version %d.%d is not supported! Please update to Python 3.6 or above' % sys.version_info[:2])
if self.params.get('allow_unplayable_formats'):
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('%s is deprecated. Use %s instead' % (option, suggestion))
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', []):
if 'list-formats' in self.params.get('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)
self.params['nooverwrites'] = not self.params['overwrites']
if params.get('bidi_workaround', False):
import pty
master, slave = pty.openpty()
width = compat_get_terminal_size().columns
if width is None:
width_args = []
width_args = ['-w', str(width)]
sp_kwargs = dict(
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:
'Could not find fribidi executable, ignoring --bidi-workaround. '
'Make sure that fribidi is an executable file in one of the directories in your $PATH.')
if (sys.platform != 'win32'
and sys.getfilesystemencoding() in ['ascii', 'ANSI_X3.4-1968']
and not params.get('restrictfilenames', False)):
# Unicode filesystem API will throw errors (#1474, #13027)
'Assuming --restrict-filenames since file system encoding '
'cannot encode all characters. '
'Set the LC_ALL environment variable to fix this.')
self.params['restrictfilenames'] = True
self.outtmpl_dict = self.parse_outtmpl()
# Creating format selector here allows us to catch syntax errors before the extraction
self.format_selector = (
None if self.params.get('format') is None
else self.params['format'] if callable(self.params['format'])
else self.build_format_selector(self.params['format']))
if auto_init:
if auto_init != 'no_verbose_header':
for pp_def_raw in self.params.get('postprocessors', []):
pp_def = dict(pp_def_raw)
when = pp_def.pop('when', 'post_process')
pp_class = get_postprocessor(pp_def.pop('key'))
pp = pp_class(self, **compat_kwargs(pp_def))
self.add_post_processor(pp, when=when)
hooks = {
'post_hooks': self.add_post_hook,
'progress_hooks': self.add_progress_hook,
'postprocessor_hooks': self.add_postprocessor_hook,
for opt, fn in hooks.items():
for ph in self.params.get(opt, []):
def preload_download_archive(fn):
"""Preload the archive, if any is specified"""
if fn is None:
return False
self.write_debug(f'Loading archive file {fn!r}')
with locked_file(fn, 'r', encoding='utf-8') as archive_file:
for line in archive_file:
except IOError as ioe:
if ioe.errno != errno.ENOENT:
return False
return True
self.archive = set()
def warn_if_short_id(self, argv):
# short YouTube ID starting with dash?
idxs = [
i for i, a in enumerate(argv)
if re.match(r'^-[0-9A-Za-z_-]{10}$', a)]
if idxs:
correct_argv = (
+ [a for i, a in enumerate(argv) if i not in idxs]
+ ['--'] + [argv[i] for i in idxs]
'Long argument string detected. '
'Use -- to separate parameters and URLs, like this:\n%s' %
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
def _get_info_extractor_class(self, ie_key):
ie = self._ies.get(ie_key)
if ie is None:
ie = get_info_extractor(ie_key)
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
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)()
return ie
def add_default_info_extractors(self):
Add the InfoExtractors returned by gen_extractors to the end of the list
for ie in gen_extractor_classes():
def add_post_processor(self, pp, when='post_process'):
"""Add a PostProcessor object to the end of the chain."""
def add_post_hook(self, ph):
"""Add the post hook"""
def add_progress_hook(self, ph):
"""Add the download progress hook"""
def add_postprocessor_hook(self, ph):
"""Add the postprocessing progress hook"""
def _bidi_workaround(self, message):
if not hasattr(self, '_output_channel'):
return message
assert hasattr(self, '_output_process')
assert isinstance(message, compat_str)
line_count = message.count('\n') + 1
self._output_process.stdin.write((message + '\n').encode('utf-8'))
res = ''.join(self._output_channel.readline().decode('utf-8')
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:
write_string(message, out=out, encoding=self.params.get('encoding'))
def to_stdout(self, message, skip_eol=False, quiet=False):
"""Print message to stdout"""
if self.params.get('logger'):
elif not quiet or self.params.get('verbose'):
'%s%s' % (self._bidi_workaround(message), ('' if skip_eol else '\n')),
self._err_file if quiet else self._screen_file)
def to_stderr(self, message, only_once=False):
"""Print message to stderr"""
assert isinstance(message, compat_str)
if self.params.get('logger'):
self._write_string('%s\n' % self._bidi_workaround(message), self._err_file, only_once=only_once)
def to_console_title(self, message):
if not self.params.get('consoletitle', False):
if compat_os_name == 'nt':
if ctypes.windll.kernel32.GetConsoleWindow():
# c_wchar_p() might not be necessary if `message` is
# already of type unicode()
elif 'TERM' in os.environ:
self._write_string('\033]0;%s\007' % message, self._screen_file)
def save_console_title(self):
if not self.params.get('consoletitle', False):
if self.params.get('simulate'):
if compat_os_name != 'nt' and 'TERM' in os.environ:
# Save the title on stack
self._write_string('\033[22;0t', self._screen_file)
def restore_console_title(self):
if not self.params.get('consoletitle', False):
if self.params.get('simulate'):
if compat_os_name != 'nt' and 'TERM' in os.environ:
# Restore the title from stack
self._write_string('\033[23;0t', self._screen_file)
def __enter__(self):
return self
def __exit__(self, *args):
if self.params.get('cookiefile') is not None:, ignore_expires=True)
def trouble(self, message=None, tb=None):
"""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.
tb, if given, is additional traceback information.
if message is not None:
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())
tb_data = traceback.format_list(traceback.extract_stack())
tb = ''.join(tb_data)
if tb:
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
exc_info = sys.exc_info()
raise DownloadError(message, exc_info)
self._download_retcode = 1
def to_screen(self, message, skip_eol=False):
"""Print message to stdout if not in quiet mode"""
message, skip_eol, quiet=self.params.get('quiet', False))
class Styles(Enum):
HEADERS = 'yellow'
EMPHASIS = 'blue'
ID = 'green'
DELIM = 'blue'
ERROR = 'red'
WARNING = 'yellow'
SUPPRESS = 'light black'
def __format_text(self, out, text, f, fallback=None, *, test_encoding=False):
assert out in ('screen', 'err')
if test_encoding:
original_text = text
handle = self._screen_file if out == 'screen' else self._err_file
encoding = self.params.get('encoding') or getattr(handle, 'encoding', 'ascii')
text = text.encode(encoding, 'ignore').decode(encoding)
if fallback is not None and text != original_text:
text = fallback
if isinstance(f, self.Styles):
f = f._value_
return format_text(text, f) if self._allow_colors[out] else text if fallback is None else fallback
def _format_screen(self, *args, **kwargs):
return self.__format_text('screen', *args, **kwargs)
def _format_err(self, *args, **kwargs):
return self.__format_text('err', *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:
if self.params.get('no_warnings'):
self.to_stderr(f'{self._format_err("WARNING:", self.Styles.WARNING)} {message}', only_once)
def report_error(self, message, tb=None):
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}', tb)
def write_debug(self, message, only_once=False):
'''Log debug message or Print message to stderr'''
if not self.params.get('verbose', False):
message = '[debug] %s' % message
if self.params.get('logger'):
self.to_stderr(message, only_once)
def report_file_already_downloaded(self, file_name):
"""Report file has already been fully downloaded."""
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."""
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):
has_drm = info.get('__has_drm')
msg = 'This video is DRM protected' if has_drm else 'No video formats found!'
expected = self.params.get('ignore_no_formats_error')
if forced or not expected:
raise ExtractorError(msg, video_id=info['id'], ie=info['extractor'],
expected=has_drm or expected)
def parse_outtmpl(self):
outtmpl_dict = self.params.get('outtmpl', {})
if not isinstance(outtmpl_dict, dict):
outtmpl_dict = {'default': outtmpl_dict}
# Remove spaces in the default template
if self.params.get('restrictfilenames'):
sanitize = lambda x: x.replace(' - ', ' ').replace(' ', '-')
sanitize = lambda x: x
k: sanitize(v) for k, v in DEFAULT_OUTTMPL.items()
if outtmpl_dict.get(k) is None})
for key, val in outtmpl_dict.items():
if isinstance(val, bytes):
'Parameter outtmpl is bytes, but should be a unicode string. '
'Put from __future__ import unicode_literals at the top of your code file or consider switching to Python 3.x.')
return outtmpl_dict
def get_output_path(self, dir_type='', filename=None):
paths = self.params.get('paths', {})
assert isinstance(paths, dict)
path = os.path.join(
expand_path(paths.get('home', '').strip()),
expand_path(paths.get(dir_type, '').strip()) if dir_type else '',
filename or '')
# Temporary fix for #4787
# 'Treat' all problem characters by passing filename through preferredencoding
# to workaround encoding issues with subprocess on python2 @ Windows
if sys.version_info < (3, 0) and sys.platform == 'win32':
path = encodeFilename(path, True).decode(preferredencoding())
return sanitize_path(path, force=self.params.get('windowsfilenames'))
def _outtmpl_expandpath(outtmpl):
# expand_path translates '%%' into '%' and '$$' into '$'
# correspondingly that is not what we want since we need to keep
# '%%' intact for template dict substitution step. Working around
# with boundary-alike separator hack.
sep = ''.join([random.choice(ascii_letters) for _ in range(32)])
outtmpl = outtmpl.replace('%%', '%{0}%'.format(sep)).replace('$$', '${0}$'.format(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. For example, for outtmpl "%(title)s.%(ext)s" and
# title "Hello $PATH", we don't want `$PATH` to be expanded.
return expand_path(outtmpl).replace(sep, '')
def escape_outtmpl(outtmpl):
''' Escape any remaining strings like %s, %abc% etc. '''
return re.sub(
STR_FORMAT_RE_TMPL.format('', '(?![%(\0])'),
lambda mobj: ('' if'has_key') else '%') +,
def validate_outtmpl(cls, outtmpl):
''' @return None or Exception object '''
outtmpl = re.sub(
STR_FORMAT_RE_TMPL.format('[^)]*', '[ljqBU]'),
lambda mobj: f'{[:-1]}s',
cls.escape_outtmpl(outtmpl) % collections.defaultdict(int)
return None
except ValueError as err:
return err
def _copy_infodict(info_dict):
info_dict = dict(info_dict)
for key in ('__original_infodict', '__postprocessors'):
info_dict.pop(key, None)
return info_dict
def prepare_outtmpl(self, outtmpl, info_dict, sanitize=None):
""" Make the outtmpl and info_dict suitable for substitution: ydl.escape_outtmpl(outtmpl) % info_dict """
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'] = self.params.get('autonumber_start', 1) - 1 + self._num_downloads
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,
EXTERNAL_FORMAT_RE = re.compile(STR_FORMAT_RE_TMPL.format('[^)]*', f'[{STR_FORMAT_TYPES}ljqBU]'))
'+': 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+)')
MATH_FIELD_RE = r'''{field}|{num}'''.format(field=FIELD_RE, num=r'-?\d+(?:.\d+)?')
MATH_OPERATORS_RE = r'(?:%s)' % '|'.join(map(re.escape, MATH_FUNCTIONS.keys()))
INTERNAL_FORMAT_RE = re.compile(r'''(?x)
$'''.format(field=FIELD_RE, math_op=MATH_OPERATORS_RE, math_field=MATH_FIELD_RE))
def _traverse_infodict(k):
k = k.split('.')
if k[0] == '':
return traverse_obj(info_dict, k, 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(
offset_key = offset_key[len(item):]
if operator is None:
operator = MATH_FUNCTIONS[item]
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))
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('\\,', ','))
return value
na = self.params.get('outtmpl_na_placeholder', 'NA')
def _dumpjson_default(obj):
if isinstance(obj, (set, LazyList)):
return list(obj)
raise TypeError(f'Object of type {type(obj).__name__} is not JSON serializable')
def create_key(outer_mobj):
if not'has_key'):
key ='key')
mobj = re.match(INTERNAL_FORMAT_RE, key)
initial_field ='fields').split('.')[-1] if mobj else ''
value, default = None, na
while mobj:
mobj = mobj.groupdict()
default = mobj['default'] if mobj['default'] is not None else default
value = get_value(mobj)
if value is None and mobj['alternate']:
mobj = re.match(INTERNAL_FORMAT_RE, mobj['alternate'][1:])
fmt ='format')
if fmt == 's' and value is not None and key in field_size_compat_map.keys():
fmt = '0{:d}d'.format(field_size_compat_map[key])
value = default if value is None else value
flags ='conversion') or ''
str_fmt = f'{fmt[:-1]}s'
if fmt[-1] == 'l': # list
delim = '\n' if '#' in flags else ', '
value, fmt = delim.join(variadic(value)), str_fmt
elif fmt[-1] == 'j': # json
value, fmt = json.dumps(value, default=_dumpjson_default, indent=4 if '#' in flags else None), 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('utf-8') % str(value).encode('utf-8')
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] == 'c':
if value:
value = str(value)[0]
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 = sanitize(initial_field, value)
key = '%s\0%s' % (key.replace('%', '%\0'),'format'))
TMPL_DICT[key] = value
return '{prefix}%({key}){fmt}'.format(key=key, fmt=fmt,'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, tmpl_type='default'):
sanitize = lambda k, v: sanitize_filename(
is_id=(k == 'id' or k.endswith('_id')))
outtmpl = self._outtmpl_expandpath(self.outtmpl_dict.get(tmpl_type, self.outtmpl_dict['default']))
filename = self.evaluate_outtmpl(outtmpl, info_dict, sanitize)
force_ext = OUTTMPL_TYPES.get(tmpl_type)
if filename and force_ext is not None:
filename = replace_extension(filename, force_ext, info_dict.get('ext'))
trim_file_name = self.params.get('trim_file_name', False)
if trim_file_name:
fn_groups = filename.rsplit('.')
ext = fn_groups[-1]
sub_ext = ''
if len(fn_groups) > 2:
sub_ext = fn_groups[-2]
filename = join_nonempty(fn_groups[0][:trim_file_name], sub_ext, 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='', warn=False):
"""Generate the output filename."""
filename = self._prepare_filename(info_dict, dir_type or 'default')
if not filename and dir_type not in ('', 'temp'):
return ''
if warn:
if not self.params.get('paths'):
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', 'video'))
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, title, re.IGNORECASE):
return '"' + title + '" title did not match pattern "' + matchtitle + '"'
rejecttitle = self.params.get('rejecttitle', False)
if rejecttitle:
if, 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 '%s upload date is not in range %s' % (date_from_str(date).isoformat(), 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:
ret = match_filter(info_dict, incomplete=incomplete)
except TypeError:
# For backward compatibility
ret = None if incomplete else match_filter(info_dict)
if 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
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
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):
Return a list with a dictionary for each video extracted.
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
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._get_info_extractor_class(ie_key)}
ies = self._ies
for ie_key, ie in ies.items():
if not ie.suitable(url):
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': ie_key}):
self.to_screen("[%s] %s: has already been recorded in archive" % (
ie_key, temp_id))
return self.__extract_info(url, self.get_info_extractor(ie_key), download, extra_info, process)
self.report_error('no suitable InfoExtractor for URL %s' % url)
def __handle_extraction_exceptions(func):
def wrapper(self, *args, **kwargs):
return func(self, *args, **kwargs)
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.'
except ExtractorError as e: # An error we somewhat expected
self.report_error(compat_str(e), e.format_traceback())
except ThrottledDownload as e:
self.report_warning(f'{e}; Re-extracting data')
return wrapper(self, *args, **kwargs)
except (DownloadCancelled, LazyList.IndexError, PagedList.IndexError):
except Exception as e:
if self.params.get('ignoreerrors'):
self.report_error(error_to_compat_str(e), tb=encode_compat_str(traceback.format_exc()))
return wrapper
def __extract_info(self, url, ie, download, extra_info, process):
ie_result = ie.extract(url)
if ie_result is None: # Finished already (backwards compatibility; listformats and friends should be moved here)
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:
return self.process_ie_result(ie_result, download, extra_info)
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_basename': url_basename(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'])
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)
self.__forced_printings(info_copy, self.prepare_filename(info_copy), incomplete=True)
if self.params.get('force_write_download_archive', False):
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)
additional_urls = (ie_result or {}).get('additional_urls')
if additional_urls:
# TODO: Improve MetadataParserPP to allow setting a list
if isinstance(additional_urls, compat_str):
additional_urls = [additional_urls]
'[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'] = [
url, download, extra_info,
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,
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
force_properties = dict(
(k, v) for k, v in ie_result.items() if v is not None)
for f in ('_type', 'url', 'id', 'extractor', 'extractor_key', 'ie_key'):
if f in force_properties:
del force_properties[f]
new_result = info.copy()
# 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
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
webpage_url = ie_result['webpage_url']
if webpage_url in self._playlist_urls:
'[download] Skipping already downloaded playlist: %s'
% ie_result.get('title') or ie_result.get('id'))
self._playlist_level += 1
return self.__process_playlist(ie_result, download)
self._playlist_level -= 1
if not self._playlist_level:
elif result_type == 'compat_list':
'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']),
'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
raise Exception('Invalid result type: %s' % result_type)
def _ensure_dir_exists(self, path):
return make_dir(path, self.report_error)
def __process_playlist(self, ie_result, download):
# We process each entry in the playlist
playlist = ie_result.get('title') or ie_result.get('id')
self.to_screen('[download] Downloading playlist: %s' % playlist)
if 'entries' not in ie_result:
raise EntryNotInPlaylist('There are no entries')
MissingEntry = object()
incomplete_entries = bool(ie_result.get('requested_entries'))
if incomplete_entries:
def fill_missing_entries(entries, indices):
ret = [MissingEntry] * max(indices)
for i, entry in zip(indices, entries):
ret[i - 1] = entry
return ret
ie_result['entries'] = fill_missing_entries(ie_result['entries'], ie_result['requested_entries'])
playlist_results = []
playliststart = self.params.get('playliststart', 1)
playlistend = self.params.get('playlistend')
# For backwards compatibility, interpret -1 as whole list
if playlistend == -1:
playlistend = None
playlistitems_str = self.params.get('playlist_items')
playlistitems = None
if playlistitems_str is not None:
def iter_playlistitems(format):
for string_segment in format.split(','):
if '-' in string_segment:
start, end = string_segment.split('-')
for item in range(int(start), int(end) + 1):
yield int(item)
yield int(string_segment)
playlistitems = orderedSet(iter_playlistitems(playlistitems_str))
ie_entries = ie_result['entries']
msg = (
'Downloading %d videos' if not isinstance(ie_entries, list)
else 'Collected %d videos; downloading %%d of them' % len(ie_entries))
if isinstance(ie_entries, list):
def get_entry(i):
return ie_entries[i - 1]
if not isinstance(ie_entries, (PagedList, LazyList)):
ie_entries = LazyList(ie_entries)
def get_entry(i):
return YoutubeDL.__handle_extraction_exceptions(
lambda self, i: ie_entries[i - 1]
)(self, i)
entries = []
items = playlistitems if playlistitems is not None else itertools.count(playliststart)
for i in items:
if i == 0:
if playlistitems is None and playlistend is not None and playlistend < i:
entry = None
entry = get_entry(i)
if entry is MissingEntry:
raise EntryNotInPlaylist()
except (IndexError, EntryNotInPlaylist):
if incomplete_entries:
raise EntryNotInPlaylist(f'Entry {i} cannot be found')
elif not playlistitems:
if entry is not None:
self._match_entry(entry, incomplete=True, silent=True)
except (ExistingVideoReached, RejectedVideoReached):
ie_result['entries'] = entries
# Save playlist_index before re-ordering
entries = [
((playlistitems[i - 1] if playlistitems else i + playliststart - 1), entry)
for i, entry in enumerate(entries, 1)
if entry is not None]
n_entries = len(entries)
if not playlistitems and (playliststart != 1 or playlistend):
playlistitems = list(range(playliststart, playliststart + n_entries))
ie_result['requested_entries'] = playlistitems
_infojson_written = False
if not self.params.get('simulate') and self.params.get('allow_playlist_files', True):
ie_copy = {
'playlist': playlist,
'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,
'n_entries': n_entries,
_infojson_written = self._write_info_json(
'playlist', ie_result, self.prepare_filename(ie_copy, 'pl_infojson'))
if _infojson_written is None:
if self._write_description('playlist', ie_result,
self.prepare_filename(ie_copy, 'pl_description')) is None:
# TODO: This should be passed to ThumbnailsConvertor if necessary
self._write_thumbnails('playlist', ie_copy, self.prepare_filename(ie_copy, 'pl_thumbnail'))
if self.params.get('playlistreverse', False):
entries = entries[::-1]
if self.params.get('playlistrandom', False):
x_forwarded_for = ie_result.get('__x_forwarded_for_ip')
self.to_screen('[%s] playlist %s: %s' % (ie_result['extractor'], playlist, msg % n_entries))
failures = 0
max_failures = self.params.get('skip_playlist_after_errors') or float('inf')
for i, entry_tuple in enumerate(entries, 1):
playlist_index, entry = entry_tuple
if 'playlist-index' in self.params.get('compat_opts', []):
playlist_index = playlistitems[i - 1] if playlistitems else i + playliststart - 1
self.to_screen('[download] Downloading video %s of %s' % (i, n_entries))
# This __x_forwarded_for_ip thing is a bit ugly but requires
# minimal changes
if x_forwarded_for:
entry['__x_forwarded_for_ip'] = x_forwarded_for
extra = {
'n_entries': n_entries,
'_last_playlist_index': max(playlistitems) if playlistitems else (playlistend or n_entries),
'playlist_index': playlist_index,
'playlist_autonumber': i,
'playlist': playlist,
'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']),
'extractor_key': ie_result['extractor_key'],
if self._match_entry(entry, incomplete=True) is not None:
entry_result = self.__process_iterable_entry(entry, download, extra)
if not entry_result:
failures += 1
if failures >= max_failures:
'Skipping the remaining entries in playlist "%s" since %d items failed extraction' % (playlist, failures))
ie_result['entries'] = playlist_results
# Write the updated info to json
if _infojson_written and self._write_info_json(
'updated playlist', ie_result,
self.prepare_filename(ie_copy, 'pl_infojson'), overwrite=True) is None:
self.to_screen('[download] Finished downloading playlist: %s' % playlist)
return ie_result
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 "
'<=': operator.le,
'=': operator.eq,
operator_rex = re.compile(r'''(?x)\s*
''' % '|'.join(map(re.escape, OPERATORS.keys())))
m = operator_rex.fullmatch(filter_spec)
if m:
comparison_value = int('value'))
except ValueError:
comparison_value = parse_filesize('value'))
if comparison_value is None:
comparison_value = parse_filesize('value') + 'B')
if comparison_value is None:
raise ValueError(
'Invalid value %r in format specification %r' % ('value'), filter_spec))
op = OPERATORS['op')]
if not m:
'=': operator.eq,
'^=': lambda attr, value: attr.startswith(value),
'$=': lambda attr, value: attr.endswith(value),
'*=': lambda attr, value: value in attr,
str_operator_rex = re.compile(r'''(?x)\s*
''' % '|'.join(map(re.escape, STR_OPERATORS.keys())))
m = str_operator_rex.fullmatch(filter_spec)
if m:
comparison_value ='value')
str_op = STR_OPERATORS['op')]
op = lambda attr, value: not str_op(attr, value)
op = str_op
if not m:
raise SyntaxError('Invalid filter specification %r' % filter_spec)
def _filter(f):
actual_value = f.get('key'))
if actual_value is None:
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}/'):
temp_file = tempfile.NamedTemporaryFile(suffix='.tmp', delete=False, dir=path or None)
success, _ = self.dl(, f, test=True)
except (DownloadError, IOError, OSError, ValueError) + network_exceptions:
success = False
if os.path.exists(
except OSError:
self.report_warning('Unable to delete temporary file "%s"' %
if success:
yield f
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', False)
or self.outtmpl_dict['default'] == '-'))
compat = (
or self.params.get('allow_multiple_audio_streams', False)
or 'format-spec' in self.params.get('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: '
'{0}\n\t{1}\n\t{2}^'.format(note, format_spec, ' ' * start[1]))
return SyntaxError(message)
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)
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'
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 == ']':
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
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):
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
elif inside_merge and string in ['/', ',']:
elif inside_choice and string == ',':
elif string == ',':
if not current_selector:
raise syntax_error('"," must follow a format selector', start)
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)
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), [])
raise syntax_error('Operator not recognized: "{0}"'.format(string), start)
elif type == tokenize.ENDMARKER:
if 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':
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]:
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 = 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']
output_ext = '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:
'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:
'acodec': the_only_audio.get('acodec'),
'abr': the_only_audio.get('abr'),
'asr': the_only_audio.get('asr'),
return new_dict
def _check_formats(formats):
if not check_formats:
yield from formats
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(copy.deepcopy(ctx)), selector_2(copy.deepcopy(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(ctx['formats']))
if not formats:
merged_format = formats[-1]
for f in formats[-2::-1]:
merged_format = _merge((merged_format, f))
yield merged_format
format_fallback, format_reverse, format_idx = False, True, 1
mobj = re.match(
if mobj is not None:
format_idx = int_or_none('n'), default=1)
format_reverse ='bw')[0] == 'b'
format_type = ('type') or [None])[0]
not_format_type = {'v': 'a', 'a': 'v'}.get(format_type)
format_modified ='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')
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']:
filter_f = lambda f: f.get('ext') == format_spec and f.get('acodec') != 'none' and f.get('vcodec') != 'none'
elif format_spec in self._format_selection_exts['storyboards']:
filter_f = lambda f: f.get('ext') == format_spec and f.get('acodec') == 'none' and f.get('vcodec') == 'none'
filter_f = lambda f: f.get('format_id') == format_spec # id
def selector_function(ctx):
formats = list(ctx['formats'])
matches = list(filter(filter_f, formats)) if filter_f is not None else formats
if format_fallback and ctx['incomplete_formats'] and not matches:
# for extractors with incomplete formats (audio only (soundcloud)
# or video only (imgur)) best/worst will fallback to
# best/worst {video,audio}-only format
matches = formats
matches = LazyList(_check_formats(matches[::-1 if format_reverse else 1]))
yield matches[format_idx - 1]
except IndexError:
filters = [self._build_format_filter(f) for f in selector.filters]
def final_selector(ctx):
ctx_copy = copy.deepcopy(ctx)
for _filter in filters:
ctx_copy['formats'] = list(filter(_filter, ctx_copy['formats']))
return selector_function(ctx_copy)
return final_selector
stream = io.BytesIO(format_spec.encode('utf-8'))
tokens = list(_remove_unused_ops(compat_tokenize_tokenize(stream.readline)))
except tokenize.TokenError:
raise syntax_error('Missing closing/opening brackets or parenthesis', (0, len(format_spec)))
class TokenIterator(object):
def __init__(self, tokens):
self.tokens = tokens
self.counter = 0
def __iter__(self):
return self
def __next__(self):
if self.counter >= len(self.tokens):
raise StopIteration()
value = self.tokens[self.counter]
self.counter += 1
return value
next = __next__
def restore_last_token(self):
self.counter -= 1
parsed_selector = _parse_format_selection(iter(TokenIterator(tokens)))
return _build_selector_function(parsed_selector)
def _calc_headers(self, info_dict):
res = std_headers.copy()
add_headers = info_dict.get('http_headers')
if add_headers:
cookies = self._calc_cookies(info_dict)
if cookies:
res['Cookie'] = cookies
if 'X-Forwarded-For' not in res:
x_forwarded_for_ip = info_dict.get('__x_forwarded_for_ip')
if x_forwarded_for_ip:
res['X-Forwarded-For'] = x_forwarded_for_ip
return res
def _calc_cookies(self, info_dict):
pr = sanitized_Request(info_dict['url'])
return pr.get_header('Cookie')
def _sort_thumbnails(self, thumbnails):
thumbnails.sort(key=lambda t: (
t.get('preference') if t.get('preference') is not None else -1,
t.get('width') if t.get('width') is not None else -1,
t.get('height') if t.get('height') is not None else -1,
t.get('id') if t.get('id') is not None else '',
def _sanitize_thumbnails(self, info_dict):
thumbnails = info_dict.get('thumbnails')
if thumbnails is None:
thumbnail = info_dict.get('thumbnail')
if thumbnail:
info_dict['thumbnails'] = thumbnails = [{'url': thumbnail}]
if not thumbnails:
def check_thumbnails(thumbnails):
for t in thumbnails:
self.to_screen(f'[info] Testing thumbnail {t["id"]}')
except network_exceptions as err:
self.to_screen(f'[info] Unable to connect to thumbnail {t["id"]} URL {t["url"]!r} - {err}. Skipping...')
yield t
for i, t in enumerate(thumbnails):
if t.get('id') is None:
t['id'] = '%d' % i
if t.get('width') and t.get('height'):
t['resolution'] = '%dx%d' % (t['width'], t['height'])
t['url'] = sanitize_url(t['url'])
if self.params.get('check_formats') is True:
info_dict['thumbnails'] = LazyList(check_thumbnails(thumbnails[::-1]), reverse=True)
info_dict['thumbnails'] = thumbnails
def process_video_result(self, info_dict, download=True):
assert info_dict.get('_type', 'video') == 'video'
if 'id' not in info_dict:
raise ExtractorError('Missing "id" field in extractor result')
if 'title' not in info_dict:
raise ExtractorError('Missing "title" field in extractor result',
video_id=info_dict['id'], ie=info_dict['extractor'])
def report_force_conversion(field, field_not, conversion):
'"%s" field is not %s - forcing %s conversion, there is an error in extractor'
% (field, field_not, conversion))
def sanitize_string_field(info, string_field):
field = info.get(string_field)
if field is None or isinstance(field, compat_str):
report_force_conversion(string_field, 'a string', 'string')
info[string_field] = compat_str(field)
def sanitize_numeric_fields(info):
for numeric_field in self._NUMERIC_FIELDS:
field = info.get(numeric_field)
if field is None or isinstance(field, compat_numeric_types):
report_force_conversion(numeric_field, 'numeric', 'int')
info[numeric_field] = int_or_none(field)
sanitize_string_field(info_dict, 'id')
if 'playlist' not in info_dict:
# It isn't part of a playlist
info_dict['playlist'] = None
info_dict['playlist_index'] = None
thumbnail = info_dict.get('thumbnail')
thumbnails = info_dict.get('thumbnails')
if thumbnail:
info_dict['thumbnail'] = sanitize_url(thumbnail)
elif thumbnails:
info_dict['thumbnail'] = thumbnails[-1]['url']
if info_dict.get('display_id') is None and 'id' in info_dict:
info_dict['display_id'] = info_dict['id']
if info_dict.get('duration') is not None:
info_dict['duration_string'] = formatSeconds(info_dict['duration'])
for ts_key, date_key in (
('timestamp', 'upload_date'),
('release_timestamp', 'release_date'),
if info_dict.get(date_key) is None and info_dict.get(ts_key) is not None:
# Working around out-of-range timestamp values (e.g. negative ones on Windows,
# see