import collections
import contextlib
+import copy
import datetime
import errno
import fileinput
-import functools
+import http.cookiejar
import io
import itertools
import json
import tokenize
import traceback
import unicodedata
-import urllib.request
from .cache import Cache
-from .compat import compat_os_name, compat_shlex_quote
-from .cookies import load_cookies
+from .compat import functools, urllib # isort: split
+from .compat import compat_os_name, compat_shlex_quote, urllib_req_to_req
+from .cookies import LenientSimpleCookie, 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 .networking import HEADRequest, Request, RequestDirector
+from .networking.common import _REQUEST_HANDLERS, _RH_PREFERENCES
+from .networking.exceptions import (
+ HTTPError,
+ NoSupportingHandlers,
+ RequestError,
+ SSLError,
+ _CompatHTTPError,
+ network_exceptions,
+)
from .plugins import directories as plugin_directories
from .postprocessor import _PLUGIN_CLASSES as plugin_pps
from .postprocessor import (
get_postprocessor,
)
from .postprocessor.ffmpeg import resolve_mapping as resolve_recode_mapping
-from .update import REPOSITORY, current_git_head, detect_variant
+from .update import REPOSITORY, _get_system_deprecation, _make_label, current_git_head, detect_variant
from .utils import (
DEFAULT_OUTTMPL,
IDENTITY,
ExtractorError,
FormatSorter,
GeoRestrictedError,
- HEADRequest,
ISO3166Utils,
LazyList,
MaxDownloadsReached,
Namespace,
PagedList,
- PerRequestProxyHandler,
PlaylistEntries,
Popen,
PostProcessingError,
SameFileError,
UnavailableVideoError,
UserNotLive,
- YoutubeDLCookieProcessor,
- YoutubeDLHandler,
- YoutubeDLRedirectHandler,
age_restricted,
args_to_str,
bug_reports_message,
error_to_compat_str,
escapeHTML,
expand_path,
+ extract_basic_auth,
filter_dict,
float_or_none,
format_bytes,
locked_file,
make_archive_id,
make_dir,
- make_HTTPS_handler,
- merge_headers,
- network_exceptions,
number_of_digits,
orderedSet,
orderedSet_from_options,
sanitize_filename,
sanitize_path,
sanitize_url,
- sanitized_Request,
- std_headers,
str_or_none,
strftime_or_none,
subtitles_filename,
write_json_file,
write_string,
)
-from .version import CHANNEL, RELEASE_GIT_HEAD, VARIANT, __version__
+from .utils._utils import _YDLLogger
+from .utils.networking import (
+ HTTPHeaderDict,
+ clean_headers,
+ clean_proxies,
+ std_headers,
+)
+from .version import CHANNEL, ORIGIN, RELEASE_GIT_HEAD, VARIANT, __version__
if compat_os_name == 'nt':
import ctypes
ap_password: Multiple-system operator account password.
usenetrc: Use netrc for authentication instead.
netrc_location: Location of the netrc file. Defaults to ~/.netrc.
+ netrc_cmd: Use a shell command to get credentials
verbose: Print additional info to stdout.
quiet: Do not print messages to stdout.
no_warnings: Do not print out anything for warnings.
'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)
+ 'temp' and the keys of OUTTMPL_TYPES (in utils/_utils.py)
outtmpl: Dictionary of templates for output names. Allowed keys
- are 'default' and the keys of OUTTMPL_TYPES (in utils.py).
+ are 'default' and the keys of OUTTMPL_TYPES (in utils/_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
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.
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
+ clean_infojson: Remove internal metadata 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
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.
+ daterange: A utils.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.
'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)
+ * False: Always process. Default for API
* 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)
+ playlists (not multi_video). Default for CLI
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
asked whether to download the video.
- Raise utils.DownloadCancelled(msg) to abort remaining
downloads when a video is rejected.
- match_filter_func in utils.py is one example for this.
- no_color: Do not emit color codes in output.
+ match_filter_func in utils/_utils.py is one example for this.
+ color: A Dictionary with output stream names as keys
+ and their respective color policy as values.
+ Can also just be a single color policy,
+ in which case it applies to all outputs.
+ Valid stream names are 'stdout' and 'stderr'.
+ Valid color policies are one of 'always', 'auto', 'no_color' or 'never'.
geo_bypass: Bypass geographic restriction via faking X-Forwarded-For
HTTP header
geo_bypass_country:
can also be used
The following options are used by the extractors:
- extractor_retries: Number of times to retry for known errors
+ extractor_retries: Number of times to retry for known errors (default: 3)
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)
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)
+ no_color: Same as `color='no_color'`
+ no_overwrites: Same as `overwrites=False`
"""
_NUMERIC_FIELDS = {
'width', 'height', 'aspect_ratio', 'resolution', 'dynamic_range', 'tbr', 'abr', 'acodec', 'asr', 'audio_channels',
'vbr', 'fps', 'vcodec', 'container', 'filesize', 'filesize_approx', 'rows', 'columns',
'player_url', 'protocol', 'fragment_base_url', 'fragments', 'is_from_start',
- 'preference', 'language', 'language_preference', 'quality', 'source_preference',
+ 'preference', 'language', 'language_preference', 'quality', 'source_preference', 'cookies',
'http_headers', 'stretched_ratio', 'no_resume', 'has_drm', 'extra_param_to_segment_url', 'hls_aes', 'downloader_options',
'page_url', 'app', 'play_path', 'tc_url', 'flash_version', 'rtmp_live', 'rtmp_conn', 'rtmp_protocol', 'rtmp_real_time'
}
self._playlist_level = 0
self._playlist_urls = set()
self.cache = Cache(self)
+ self.__header_cookies = []
stdout = sys.stderr if self.params.get('logtostderr') else sys.stdout
self._out_files = Namespace(
except Exception as e:
self.write_debug(f'Failed to enable VT mode: {e}')
+ if self.params.get('no_color'):
+ if self.params.get('color') is not None:
+ self.params.setdefault('_warnings', []).append(
+ 'Overwriting params from "color" with "no_color"')
+ self.params['color'] = 'no_color'
+
+ term_allow_color = os.environ.get('TERM', '').lower() != 'dumb'
+
+ def process_color_policy(stream):
+ stream_name = {sys.stdout: 'stdout', sys.stderr: 'stderr'}[stream]
+ policy = traverse_obj(self.params, ('color', (stream_name, None), {str}), get_all=False)
+ if policy in ('auto', None):
+ return term_allow_color and supports_terminal_sequences(stream)
+ assert policy in ('always', 'never', 'no_color'), policy
+ return {'always': True, 'never': False}.get(policy, policy)
+
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'
+ name: process_color_policy(stream)
+ for name, stream in self._out_files.items_ if name != '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.deprecated_feature(
- f'{msg}! Please update to Python %d.%d or above' % (*current_version, *MIN_RECOMMENDED))
+ system_deprecation = _get_system_deprecation()
+ if system_deprecation:
+ self.deprecated_feature(system_deprecation.replace('\n', '\n '))
if self.params.get('allow_unplayable_formats'):
self.report_warning(
raise
self.params['compat_opts'] = set(self.params.get('compat_opts', ()))
+ self.params['http_headers'] = HTTPHeaderDict(std_headers, self.params.get('http_headers'))
+ self._load_cookies(self.params['http_headers'].get('Cookie')) # compat
+ self.params['http_headers'].pop('Cookie', None)
+ self._request_director = self.build_request_director(_REQUEST_HANDLERS.values(), _RH_PREFERENCES)
+
if auto_init and auto_init != 'no_verbose_header':
self.print_debug_header()
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,
get_postprocessor(pp_def.pop('key'))(self, **pp_def),
when=when)
- self._setup_opener()
-
def preload_download_archive(fn):
"""Preload the archive, if any is specified"""
archive = set()
self.save_console_title()
return self
+ def save_cookies(self):
+ if self.params.get('cookiefile') is not None:
+ self.cookiejar.save()
+
def __exit__(self, *args):
self.restore_console_title()
+ self.close()
- if self.params.get('cookiefile') is not None:
- self.cookiejar.save(ignore_discard=True, ignore_expires=True)
+ def close(self):
+ self.save_cookies()
+ self._request_director.close()
def trouble(self, message=None, tb=None, is_error=True):
"""Determine action to take when a download problem appears.
ID='green',
DELIM='blue',
ERROR='red',
+ BAD_FORMAT='light red',
WARNING='yellow',
SUPPRESS='light black',
)
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
+ return format_text(text, f) if allow_colors is True 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)
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
+ value, replacement, default, last_field = 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']
+ last_field, replacement = mobj['fields'], 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'
-
- if value is None:
- value = default
- elif replacement is not None:
+ if None not in (value, replacement):
try:
value = replacement_formatter.format(replacement, value)
except ValueError:
- value = na
+ value, default = None, na
+
+ fmt = outer_mobj.group('format')
+ if fmt == 's' and last_field in field_size_compat_map.keys() and isinstance(value, int):
+ fmt = f'0{field_size_compat_map[last_field]:d}d'
flags = outer_mobj.group('conversion') or ''
str_fmt = f'{fmt[:-1]}s'
- if fmt[-1] == 'l': # list
+ if value is None:
+ value, fmt = default, 's'
+ elif 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 = 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
+ value, fmt = filename_sanitizer(last_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
+ elif fmt[-1] not in 'rsa': # numeric
value = float_or_none(value)
if value is None:
value, fmt = default, 's'
if sanitize:
+ # If value is an object, sanitize might convert it to a string
+ # So we convert it to repr first
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)
+ elif fmt[-1] == 'a':
+ value, fmt = ascii(value), str_fmt
+ if fmt[-1] in 'csra':
+ value = sanitizer(last_field, value)
key = '%s\0%s' % (key.replace('%', '%\0'), outer_mobj.group('format'))
TMPL_DICT[key] = value
def _match_entry(self, info_dict, incomplete=False, silent=False):
"""Returns None if the file should be downloaded"""
- _type = info_dict.get('_type', 'video')
+ _type = 'video' if 'playlist-match-filter' in self.params['compat_opts'] else info_dict.get('_type', 'video')
assert incomplete or _type == 'video', 'Only video result can be considered complete'
video_title = info_dict.get('title', info_dict.get('id', 'entry'))
return ret
if self.in_download_archive(info_dict):
- reason = '%s has already been recorded in the archive' % video_title
+ reason = ''.join((
+ format_field(info_dict, 'id', f'{self._format_screen("%s", self.Styles.ID)}: '),
+ format_field(info_dict, 'title', f'{self._format_screen("%s", self.Styles.EMPHASIS)} '),
+ 'has already been recorded in the archive'))
break_opt, break_err = 'break_on_existing', ExistingVideoReached
else:
try:
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')
+ self.to_screen(f'[download] {self._format_screen(temp_id, self.Styles.ID)}: '
+ 'has already been recorded in the archive')
if self.params.get('break_on_existing', False):
raise ExistingVideoReached()
break
self.to_screen('')
raise
+ def _load_cookies(self, data, *, autoscope=True):
+ """Loads cookies from a `Cookie` header
+
+ This tries to work around the security vulnerability of passing cookies to every domain.
+ See: https://github.com/yt-dlp/yt-dlp/security/advisories/GHSA-v8mc-9377-rwjj
+
+ @param data The Cookie header as string to load the cookies from
+ @param autoscope If `False`, scope cookies using Set-Cookie syntax and error for cookie without domains
+ If `True`, save cookies for later to be stored in the jar with a limited scope
+ If a URL, save cookies in the jar with the domain of the URL
+ """
+ for cookie in LenientSimpleCookie(data).values():
+ if autoscope and any(cookie.values()):
+ raise ValueError('Invalid syntax in Cookie Header')
+
+ domain = cookie.get('domain') or ''
+ expiry = cookie.get('expires')
+ if expiry == '': # 0 is valid
+ expiry = None
+ prepared_cookie = http.cookiejar.Cookie(
+ cookie.get('version') or 0, cookie.key, cookie.value, None, False,
+ domain, True, True, cookie.get('path') or '', bool(cookie.get('path')),
+ cookie.get('secure') or False, expiry, False, None, None, {})
+
+ if domain:
+ self.cookiejar.set_cookie(prepared_cookie)
+ elif autoscope is True:
+ self.deprecated_feature(
+ 'Passing cookies as a header is a potential security risk; '
+ 'they will be scoped to the domain of the downloaded urls. '
+ 'Please consider loading cookies from a file or browser instead.')
+ self.__header_cookies.append(prepared_cookie)
+ elif autoscope:
+ self.report_warning(
+ 'The extractor result contains an unscoped cookie as an HTTP header. '
+ f'If you are using yt-dlp with an input URL{bug_reports_message(before=",")}',
+ only_once=True)
+ self._apply_header_cookies(autoscope, [prepared_cookie])
+ else:
+ self.report_error('Unscoped cookies are not allowed; please specify some sort of scoping',
+ tb=False, is_error=False)
+
+ def _apply_header_cookies(self, url, cookies=None):
+ """Applies stray header cookies to the provided url
+
+ This loads header cookies and scopes them to the domain provided in `url`.
+ While this is not ideal, it helps reduce the risk of them being sent
+ to an unintended destination while mostly maintaining compatibility.
+ """
+ parsed = urllib.parse.urlparse(url)
+ if not parsed.hostname:
+ return
+
+ for cookie in map(copy.copy, cookies or self.__header_cookies):
+ cookie.domain = f'.{parsed.hostname}'
+ self.cookiejar.set_cookie(cookie)
+
@_handle_extraction_exceptions
def __extract_info(self, url, ie, download, extra_info, process):
+ self._apply_header_cookies(url)
+
try:
ie_result = ie.extract(url)
except UserNotLive as e:
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', []):
+ if not lazy and 'playlist-index' in self.params['compat_opts']:
playlist_index = ie_result['requested_entries'][i]
entry_copy = collections.ChainMap(entry, {
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:
return new_dict
def _check_formats(formats):
- if not check_formats:
+ if self.params.get('check_formats') == 'selected':
+ yield from self._check_formats(formats)
+ return
+ elif (self.params.get('check_formats') is not None
+ or self.params.get('allow_unplayable_formats')):
yield from formats
return
- yield from self._check_formats(formats)
+
+ for f in formats:
+ if f.get('has_drm') or f.get('__needs_testing'):
+ yield from self._check_formats([f])
+ else:
+ yield f
def _build_selector_function(selector):
if isinstance(selector, list): # ,
parsed_selector = _parse_format_selection(iter(TokenIterator(tokens)))
return _build_selector_function(parsed_selector)
- def _calc_headers(self, info_dict):
- res = merge_headers(self.params['http_headers'], info_dict.get('http_headers') or {})
- if 'Youtubedl-No-Compression' in res: # deprecated
- res.pop('Youtubedl-No-Compression', None)
- res['Accept-Encoding'] = 'identity'
- cookies = self._calc_cookies(info_dict['url'])
+ def _calc_headers(self, info_dict, load_cookies=False):
+ res = HTTPHeaderDict(self.params['http_headers'], info_dict.get('http_headers'))
+ clean_headers(res)
+
+ if load_cookies: # For --load-info-json
+ self._load_cookies(res.get('Cookie'), autoscope=info_dict['url']) # compat
+ self._load_cookies(info_dict.get('cookies'), autoscope=False)
+ # The `Cookie` header is removed to prevent leaks and unscoped cookies.
+ # See: https://github.com/yt-dlp/yt-dlp/security/advisories/GHSA-v8mc-9377-rwjj
+ res.pop('Cookie', None)
+ cookies = self.cookiejar.get_cookies_for_url(info_dict['url'])
if cookies:
- res['Cookie'] = cookies
+ encoder = LenientSimpleCookie()
+ values = []
+ for cookie in cookies:
+ _, value = encoder.value_encode(cookie.value)
+ values.append(f'{cookie.name}={value}')
+ if cookie.domain:
+ values.append(f'Domain={cookie.domain}')
+ if cookie.path:
+ values.append(f'Path={cookie.path}')
+ if cookie.secure:
+ values.append('Secure')
+ if cookie.expires:
+ values.append(f'Expires={cookie.expires}')
+ if cookie.version:
+ values.append(f'Version={cookie.version}')
+ info_dict['cookies'] = '; '.join(values)
if 'X-Forwarded-For' not in res:
x_forwarded_for_ip = info_dict.get('__x_forwarded_for_ip')
return res
def _calc_cookies(self, url):
- pr = sanitized_Request(url)
- self.cookiejar.add_cookie_header(pr)
- return pr.get_header('Cookie')
+ self.deprecation_warning('"YoutubeDL._calc_cookies" is deprecated and may be removed in a future version')
+ return self.cookiejar.get_cookie_header(url)
def _sort_thumbnails(self, thumbnails):
thumbnails.sort(key=lambda t: (
# Working around out-of-range timestamp values (e.g. negative ones on Windows,
# see http://bugs.python.org/issue1646728)
with contextlib.suppress(ValueError, OverflowError, OSError):
- upload_date = datetime.datetime.utcfromtimestamp(info_dict[ts_key])
+ upload_date = datetime.datetime.fromtimestamp(info_dict[ts_key], datetime.timezone.utc)
info_dict[date_key] = upload_date.strftime('%Y%m%d')
live_keys = ('is_live', 'was_live')
if field_preference:
info_dict['_format_sort_fields'] = field_preference
- # or None ensures --clean-infojson removes it
- info_dict['_has_drm'] = any(f.get('has_drm') for f in formats) or None
+ info_dict['_has_drm'] = any( # or None ensures --clean-infojson removes it
+ f.get('has_drm') and f['has_drm'] != 'maybe' for f in formats) or None
if not self.params.get('allow_unplayable_formats'):
- formats = [f for f in formats if not f.get('has_drm')]
+ formats = [f for f in formats if not f.get('has_drm') or f['has_drm'] == 'maybe']
if formats and all(f.get('acodec') == f.get('vcodec') == 'none' for f in formats):
self.report_warning(
format['dynamic_range'] = 'SDR'
if format.get('aspect_ratio') is None:
format['aspect_ratio'] = try_call(lambda: round(format['width'] / format['height'], 2))
- if (info_dict.get('duration') and format.get('tbr')
+ # For fragmented formats, "tbr" is often max bitrate and not average
+ if (('manifest-filesize-approx' in self.params['compat_opts'] or not format.get('manifest_url'))
+ and info_dict.get('duration') and format.get('tbr')
and not format.get('filesize') and not format.get('filesize_approx')):
format['filesize_approx'] = int(info_dict['duration'] * format['tbr'] * (1024 / 8))
- format['http_headers'] = self._calc_headers(collections.ChainMap(format, info_dict))
+ format['http_headers'] = self._calc_headers(collections.ChainMap(format, info_dict), load_cookies=True)
+
+ # Safeguard against old/insecure infojson when using --load-info-json
+ if info_dict.get('http_headers'):
+ info_dict['http_headers'] = HTTPHeaderDict(info_dict['http_headers'])
+ info_dict['http_headers'].pop('Cookie', None)
# This is copied to http_headers by the above _calc_headers and can now be removed
if '__x_forwarded_for_ip' in info_dict:
return info_dict
format_selector = self.format_selector
- if format_selector is None:
- req_format = self._default_format_spec(info_dict, download=download)
- self.write_debug('Default format spec: %s' % req_format)
- format_selector = self.build_format_selector(req_format)
-
while True:
if interactive_format_selection:
- req_format = input(
- self._format_screen('\nEnter format selector: ', self.Styles.EMPHASIS))
+ req_format = input(self._format_screen('\nEnter format selector ', self.Styles.EMPHASIS)
+ + '(Press ENTER for default, or Ctrl+C to quit)'
+ + self._format_screen(': ', self.Styles.EMPHASIS))
try:
- format_selector = self.build_format_selector(req_format)
+ format_selector = self.build_format_selector(req_format) if req_format else None
except SyntaxError as err:
self.report_error(err, tb=False, is_error=False)
continue
+ if format_selector is None:
+ req_format = self._default_format_spec(info_dict, download=download)
+ self.write_debug(f'Default format spec: {req_format}')
+ format_selector = self.build_format_selector(req_format)
+
formats_to_download = list(format_selector({
'formats': formats,
'has_merged_format': any('none' not in (f.get('acodec'), f.get('vcodec')) for f in formats),
- 'incomplete_formats': (
- # All formats are video-only or
- all(f.get('vcodec') != 'none' and f.get('acodec') == 'none' for f in formats)
- # all formats are audio-only
- or all(f.get('vcodec') == 'none' and f.get('acodec') != 'none' for f in formats)),
+ 'incomplete_formats': (all(f.get('vcodec') == 'none' for f in formats) # No formats with video
+ or all(f.get('acodec') == 'none' for f in formats)), # OR, No formats with audio
}))
if interactive_format_selection and not formats_to_download:
self.report_error('Requested format is not available', tb=False, is_error=False)
new_info.update(fmt)
offset, duration = info_dict.get('section_start') or 0, info_dict.get('duration') or float('inf')
end_time = offset + min(chapter.get('end_time', duration), duration)
+ # duration may not be accurate. So allow deviations <1sec
+ if end_time == float('inf') or end_time > offset + duration + 1:
+ end_time = None
if chapter or offset:
new_info.update({
'section_start': offset + chapter.get('start_time', 0),
- # duration may not be accurate. So allow deviations <1sec
- 'section_end': end_time if end_time <= offset + duration + 1 else None,
+ 'section_end': end_time,
'section_title': chapter.get('title'),
'section_number': chapter.get('index'),
})
print_field('url', 'urls')
print_field('thumbnail', optional=True)
print_field('description', optional=True)
- if filename:
- print_field('filename')
+ print_field('filename')
if self.params.get('forceduration') and info_copy.get('duration') is not None:
self.to_stdout(formatSeconds(info_copy['duration']))
print_field('format')
fd, success = None, True
if info_dict.get('protocol') or info_dict.get('url'):
fd = get_suitable_downloader(info_dict, self.params, to_stdout=temp_filename == '-')
- if fd is not FFmpegFD and 'no-direct-merge' not in self.params['compat_opts'] and (
+ if fd != FFmpegFD and 'no-direct-merge' not in self.params['compat_opts'] and (
info_dict.get('section_start') or info_dict.get('section_end')):
msg = ('This format cannot be partially downloaded' if FFmpegFD.available()
else 'You have requested downloading the video partially, but ffmpeg is not installed')
return
if info_dict.get('requested_formats') is not None:
- requested_formats = info_dict['requested_formats']
old_ext = info_dict['ext']
if self.params.get('merge_output_format') is None:
if (info_dict['ext'] == 'webm'
full_filename = correct_ext(full_filename)
temp_filename = correct_ext(temp_filename)
dl_filename = existing_video_file(full_filename, temp_filename)
+
info_dict['__real_download'] = False
+ # NOTE: Copy so that original format dicts are not modified
+ info_dict['requested_formats'] = list(map(dict, info_dict['requested_formats']))
merger = FFmpegMergerPP(self)
downloaded = []
if dl_filename is not None:
self.report_file_already_downloaded(dl_filename)
elif fd:
- for f in requested_formats if fd != FFmpegFD else []:
+ for f in info_dict['requested_formats'] if fd != FFmpegFD else []:
f['filepath'] = fname = prepend_extension(
correct_ext(temp_filename, info_dict['ext']),
'f%s' % f['format_id'], info_dict['ext'])
downloaded.append(fname)
- info_dict['url'] = '\n'.join(f['url'] for f in requested_formats)
+ info_dict['url'] = '\n'.join(f['url'] for f in info_dict['requested_formats'])
success, real_download = self.dl(temp_filename, info_dict)
info_dict['__real_download'] = real_download
else:
f'You have requested downloading multiple formats to stdout {reason}. '
'The formats will be streamed one after the other')
fname = temp_filename
- for f in requested_formats:
+ for f in info_dict['requested_formats']:
new_info = dict(info_dict)
del new_info['requested_formats']
new_info.update(f)
) for pp in self._pps['post_process'])
if not postprocessed_by_ffmpeg:
- ffmpeg_fixup(ext == 'm4a' and info_dict.get('container') == 'm4a_dash',
+ ffmpeg_fixup(fd != FFmpegFD and ext == 'm4a'
+ and info_dict.get('container') == 'm4a_dash',
'writing DASH m4a. Only some players support this container',
FFmpegFixupM4aPP)
ffmpeg_fixup(downloader == 'hlsnative' and not self.params.get('hls_use_mpegts')
'version': __version__,
'current_git_head': current_git_head(),
'release_git_head': RELEASE_GIT_HEAD,
- 'repository': REPOSITORY,
+ 'repository': ORIGIN,
})
if remove_private_keys:
reject = lambda k, v: v is None or k.startswith('__') or k in {
'requested_downloads', 'requested_formats', 'requested_subtitles', 'requested_entries',
'entries', 'filepath', '_filename', 'filename', 'infojson_filename', 'original_url',
- 'playlist_autonumber', '_format_sort_fields',
+ 'playlist_autonumber',
}
else:
reject = lambda k, v: False
def simplified_codec(f, field):
assert field in ('acodec', 'vcodec')
- codec = f.get(field, 'unknown')
+ codec = f.get(field)
if not codec:
return 'unknown'
elif codec != 'none':
format_field(f, 'fps', '\t%d', func=round),
format_field(f, 'dynamic_range', '%s', ignore=(None, 'SDR')).replace('HDR', ''),
format_field(f, 'audio_channels', '\t%s'),
- delim,
- format_field(f, 'filesize', ' \t%s', func=format_bytes) + format_field(f, 'filesize_approx', '~\t%s', func=format_bytes),
+ delim, (
+ format_field(f, 'filesize', ' \t%s', func=format_bytes)
+ or format_field(f, 'filesize_approx', '≈\t%s', func=format_bytes)
+ or format_field(try_call(lambda: format_bytes(int(info_dict['duration'] * f['tbr'] * (1024 / 8)))),
+ None, self._format_out('~\t%s', self.Styles.SUPPRESS))),
format_field(f, 'tbr', '\t%dk', func=round),
shorten_protocol_name(f.get('protocol', '')),
delim,
simplified_codec(f, 'acodec'),
format_field(f, 'abr', '\t%dk', func=round),
format_field(f, 'asr', '\t%s', func=format_decimal_suffix),
- join_nonempty(
- self._format_out('UNSUPPORTED', 'light red') if f.get('ext') in ('f4f', 'f4m') else None,
- self._format_out('DRM', 'light red') if f.get('has_drm') else None,
- format_field(f, 'language', '[%s]'),
- join_nonempty(format_field(f, 'format_note'),
- format_field(f, 'container', ignore=(None, f.get('ext'))),
- delim=', '),
- delim=' '),
+ join_nonempty(format_field(f, 'language', '[%s]'), join_nonempty(
+ self._format_out('UNSUPPORTED', self.Styles.BAD_FORMAT) if f.get('ext') in ('f4f', 'f4m') else None,
+ (self._format_out('Maybe DRM', self.Styles.WARNING) if f.get('has_drm') == 'maybe'
+ else self._format_out('DRM', self.Styles.BAD_FORMAT) if f.get('has_drm') else None),
+ format_field(f, 'format_note'),
+ format_field(f, 'container', ignore=(None, f.get('ext'))),
+ delim=', '), delim=' '),
] for f in formats if f.get('preference') is None or f['preference'] >= -1000]
header_line = self._list_format_headers(
'ID', 'EXT', 'RESOLUTION', '\tFPS', 'HDR', 'CH', delim, '\tFILESIZE', '\tTBR', 'PROTO',
def list_subtitles(self, video_id, subtitles, name='subtitles'):
self.__list_table(video_id, name, self.render_subtitles_table, video_id, subtitles)
- def urlopen(self, req):
- """ Start an HTTP download """
- if isinstance(req, str):
- req = sanitized_Request(req)
- return self._opener.open(req, timeout=self._socket_timeout)
-
def print_debug_header(self):
if not self.params.get('verbose'):
return
def get_encoding(stream):
ret = str(getattr(stream, 'encoding', 'missing (%s)' % type(stream).__name__))
+ additional_info = []
+ if os.environ.get('TERM', '').lower() == 'dumb':
+ additional_info.append('dumb')
if not supports_terminal_sequences(stream):
from .utils import WINDOWS_VT_MODE # Must be imported locally
- ret += ' (No VT)' if WINDOWS_VT_MODE is False else ' (No ANSI)'
+ additional_info.append('No VT' if WINDOWS_VT_MODE is False else 'No ANSI')
+ if additional_info:
+ ret = f'{ret} ({",".join(additional_info)})'
return ret
encoding_str = 'Encodings: locale %s, fs %s, pref %s, %s' % (
source += '*'
klass = type(self)
write_debug(join_nonempty(
- f'{"yt-dlp" if REPOSITORY == "yt-dlp/yt-dlp" else REPOSITORY} version',
- f'{CHANNEL}@{__version__}',
+ f'{REPOSITORY.rpartition("/")[2]} version',
+ _make_label(ORIGIN, CHANNEL.partition('@')[2] or __version__, __version__),
f'[{RELEASE_GIT_HEAD[:9]}]' if RELEASE_GIT_HEAD else '',
'' if source == 'unknown' else f'({source})',
'' if _IN_CLI else 'API' if klass == YoutubeDL else f'API:{self.__module__}.{klass.__qualname__}',
join_nonempty(*get_package_info(m)) for m in available_dependencies.values()
})) or 'none'))
- self._setup_opener()
- proxy_map = {}
- for handler in self._opener.handlers:
- if hasattr(handler, 'proxies'):
- proxy_map.update(handler.proxies)
- write_debug(f'Proxy map: {proxy_map}')
-
+ write_debug(f'Proxy map: {self.proxies}')
+ write_debug(f'Request Handlers: {", ".join(rh.RH_NAME for rh in self._request_director.handlers.values())}')
for plugin_type, plugins in {'Extractor': plugin_ies, 'Post-Processor': plugin_pps}.items():
display_list = ['%s%s' % (
klass.__name__, '' if klass.__name__ == name else f' as {name}')
'See https://yt-dl.org/update if you need help updating.' %
latest_version)
- def _setup_opener(self):
- if hasattr(self, '_opener'):
- return
- timeout_val = self.params.get('socket_timeout')
- self._socket_timeout = 20 if timeout_val is None else float(timeout_val)
-
- opts_cookiesfrombrowser = self.params.get('cookiesfrombrowser')
- opts_cookiefile = self.params.get('cookiefile')
+ @functools.cached_property
+ def proxies(self):
+ """Global proxy configuration"""
opts_proxy = self.params.get('proxy')
-
- self.cookiejar = load_cookies(opts_cookiefile, opts_cookiesfrombrowser, self)
-
- cookie_processor = YoutubeDLCookieProcessor(self.cookiejar)
if opts_proxy is not None:
if opts_proxy == '':
- proxies = {}
- else:
- proxies = {'http': opts_proxy, 'https': opts_proxy}
+ opts_proxy = '__noproxy__'
+ proxies = {'all': opts_proxy}
else:
proxies = urllib.request.getproxies()
- # Set HTTPS proxy to HTTP one if given (https://github.com/ytdl-org/youtube-dl/issues/805)
+ # compat. Set HTTPS_PROXY to __noproxy__ to revert
if 'http' in proxies and 'https' not in proxies:
proxies['https'] = proxies['http']
- proxy_handler = PerRequestProxyHandler(proxies)
-
- debuglevel = 1 if self.params.get('debug_printtraffic') else 0
- https_handler = make_HTTPS_handler(self.params, debuglevel=debuglevel)
- ydlh = YoutubeDLHandler(self.params, debuglevel=debuglevel)
- redirect_handler = YoutubeDLRedirectHandler()
- data_handler = urllib.request.DataHandler()
-
- # When passing our own FileHandler instance, build_opener won't add the
- # default FileHandler and allows us to disable the file protocol, which
- # can be used for malicious purposes (see
- # https://github.com/ytdl-org/youtube-dl/issues/8227)
- file_handler = urllib.request.FileHandler()
-
- if not self.params.get('enable_file_urls'):
- def file_open(*args, **kwargs):
- raise urllib.error.URLError(
- 'file:// URLs are explicitly disabled in yt-dlp for security reasons. '
- 'Use --enable-file-urls to enable at your own risk.')
- file_handler.file_open = file_open
-
- opener = urllib.request.build_opener(
- proxy_handler, https_handler, cookie_processor, ydlh, redirect_handler, data_handler, file_handler)
-
- # Delete the default user-agent header, which would otherwise apply in
- # cases where our custom HTTP handler doesn't come into play
- # (See https://github.com/ytdl-org/youtube-dl/issues/1309 for details)
- opener.addheaders = []
- self._opener = opener
+
+ return proxies
+
+ @functools.cached_property
+ def cookiejar(self):
+ """Global cookiejar instance"""
+ return load_cookies(
+ self.params.get('cookiefile'), self.params.get('cookiesfrombrowser'), self)
+
+ @property
+ def _opener(self):
+ """
+ Get a urllib OpenerDirector from the Urllib handler (deprecated).
+ """
+ self.deprecation_warning('YoutubeDL._opener is deprecated, use YoutubeDL.urlopen()')
+ handler = self._request_director.handlers['Urllib']
+ return handler._get_instance(cookiejar=self.cookiejar, proxies=self.proxies)
+
+ def urlopen(self, req):
+ """ Start an HTTP download """
+ if isinstance(req, str):
+ req = Request(req)
+ elif isinstance(req, urllib.request.Request):
+ self.deprecation_warning(
+ 'Passing a urllib.request.Request object to YoutubeDL.urlopen() is deprecated. '
+ 'Use yt_dlp.networking.common.Request instead.')
+ req = urllib_req_to_req(req)
+ assert isinstance(req, Request)
+
+ # compat: Assume user:pass url params are basic auth
+ url, basic_auth_header = extract_basic_auth(req.url)
+ if basic_auth_header:
+ req.headers['Authorization'] = basic_auth_header
+ req.url = sanitize_url(url)
+
+ clean_proxies(proxies=req.proxies, headers=req.headers)
+ clean_headers(req.headers)
+
+ try:
+ return self._request_director.send(req)
+ except NoSupportingHandlers as e:
+ for ue in e.unsupported_errors:
+ if not (ue.handler and ue.msg):
+ continue
+ if ue.handler.RH_KEY == 'Urllib' and 'unsupported url scheme: "file"' in ue.msg.lower():
+ raise RequestError(
+ 'file:// URLs are disabled by default in yt-dlp for security reasons. '
+ 'Use --enable-file-urls to enable at your own risk.', cause=ue) from ue
+ if 'unsupported proxy type: "https"' in ue.msg.lower():
+ raise RequestError(
+ 'To use an HTTPS proxy for this request, one of the following dependencies needs to be installed: requests')
+ raise
+ except SSLError as e:
+ if 'UNSAFE_LEGACY_RENEGOTIATION_DISABLED' in str(e):
+ raise RequestError('UNSAFE_LEGACY_RENEGOTIATION_DISABLED: Try using --legacy-server-connect', cause=e) from e
+ elif 'SSLV3_ALERT_HANDSHAKE_FAILURE' in str(e):
+ raise RequestError(
+ 'SSLV3_ALERT_HANDSHAKE_FAILURE: The server may not support the current cipher list. '
+ 'Try using --legacy-server-connect', cause=e) from e
+ raise
+ except HTTPError as e: # TODO: Remove in a future release
+ raise _CompatHTTPError(e) from e
+
+ def build_request_director(self, handlers, preferences=None):
+ logger = _YDLLogger(self)
+ headers = self.params['http_headers'].copy()
+ proxies = self.proxies.copy()
+ clean_headers(headers)
+ clean_proxies(proxies, headers)
+
+ director = RequestDirector(logger=logger, verbose=self.params.get('debug_printtraffic'))
+ for handler in handlers:
+ director.add_handler(handler(
+ logger=logger,
+ headers=headers,
+ cookiejar=self.cookiejar,
+ proxies=proxies,
+ prefer_system_certs='no-certifi' in self.params['compat_opts'],
+ verify=not self.params.get('nocheckcertificate'),
+ **traverse_obj(self.params, {
+ 'verbose': 'debug_printtraffic',
+ 'source_address': 'source_address',
+ 'timeout': 'socket_timeout',
+ 'legacy_ssl_support': 'legacyserverconnect',
+ 'enable_file_urls': 'enable_file_urls',
+ 'client_cert': {
+ 'client_certificate': 'client_certificate',
+ 'client_certificate_key': 'client_certificate_key',
+ 'client_certificate_password': 'client_certificate_password',
+ },
+ }),
+ ))
+ director.preferences.update(preferences or [])
+ if 'prefer-legacy-http-handler' in self.params['compat_opts']:
+ director.preferences.add(lambda rh, _: 500 if rh.RH_KEY == 'Urllib' else 0)
+ return director
def encode(self, s):
if isinstance(s, bytes):
return ret
def _write_thumbnails(self, label, info_dict, filename, thumb_filename_base=None):
- ''' Write thumbnails to file and return list of (thumb_filename, final_thumb_filename) '''
+ ''' Write thumbnails to file and return list of (thumb_filename, final_thumb_filename); or None if error '''
write_all = self.params.get('write_all_thumbnails', False)
thumbnails, ret = [], []
if write_all or self.params.get('writethumbnail', False):
self.write_debug(f'Skipping writing {label} thumbnail')
return ret
+ if thumbnails and not self._ensure_dir_exists(filename):
+ return None
+
for idx, t in list(enumerate(thumbnails))[::-1]:
thumb_ext = (f'{t["id"]}.' if multiple else '') + determine_ext(t['url'], 'jpg')
thumb_display_id = f'{label} thumbnail {t["id"]}'
else:
self.to_screen(f'[info] Downloading {thumb_display_id} ...')
try:
- uf = self.urlopen(sanitized_Request(t['url'], headers=t.get('http_headers', {})))
+ uf = self.urlopen(Request(t['url'], headers=t.get('http_headers', {})))
self.to_screen(f'[info] Writing {thumb_display_id} to: {thumb_filename}')
with open(encodeFilename(thumb_filename), 'wb') as thumbf:
shutil.copyfileobj(uf, thumbf)
ret.append((thumb_filename, thumb_filename_final))
t['filepath'] = thumb_filename
except network_exceptions as err:
+ if isinstance(err, HTTPError) and err.status == 404:
+ self.to_screen(f'[info] {thumb_display_id.title()} does not exist')
+ else:
+ self.report_warning(f'Unable to download {thumb_display_id}: {err}')
thumbnails.pop(idx)
- self.report_warning(f'Unable to download {thumb_display_id}: {err}')
if ret and not write_all:
break
return ret