-#!/usr/bin/env python
+#!/usr/bin/env python3
# coding: utf-8
from __future__ import unicode_literals
'wav',
'f4f', 'f4m', 'm3u8', 'smil')
-REMUX_EXTENSIONS = ('mp4', 'mkv', 'flv', 'webm', 'mov', 'avi', 'mp3', 'mka', 'm4a', 'ogg', 'opus')
-
# needed for sanitizing filenames in restricted mode
ACCENT_CHARS = dict(zip('ÂÃÄÀÁÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖŐØŒÙÚÛÜŰÝÞßàáâãäåæçèéêëìíîïðñòóôõöőøœùúûüűýþÿ',
itertools.chain('AAAAAA', ['AE'], 'CEEEEIIIIDNOOOOOOO', ['OE'], 'UUUUUY', ['TH', 'ss'],
'%Y-%m-%d %H:%M',
'%Y-%m-%d %H:%M:%S',
'%Y-%m-%d %H:%M:%S.%f',
+ '%Y-%m-%d %H:%M:%S:%f',
'%d.%m.%Y %H:%M',
'%d.%m.%Y %H.%M',
'%Y-%m-%dT%H:%M:%SZ',
try:
with tf:
- json.dump(obj, tf, default=repr)
+ json.dump(obj, tf)
if sys.platform == 'win32':
# Need to remove existing file on Windows, else os.rename raises
# WindowsError or FileExistsError.
return '_'
return char
+ if s == '':
+ return ''
# Handle timestamps
s = re.sub(r'[0-9]+(?::[0-9]+)+', lambda m: m.group(0).replace(':', '_'), s)
result = ''.join(map(replace_insane, s))
return url
+def extract_basic_auth(url):
+ parts = compat_urlparse.urlsplit(url)
+ if parts.username is None:
+ return url, None
+ url = compat_urlparse.urlunsplit(parts._replace(netloc=(
+ parts.hostname if parts.port is None
+ else '%s:%d' % (parts.hostname, parts.port))))
+ auth_payload = base64.b64encode(
+ ('%s:%s' % (parts.username, parts.password or '')).encode('utf-8'))
+ return url, 'Basic ' + auth_payload.decode('utf-8')
+
+
def sanitized_Request(url, *args, **kwargs):
- return compat_urllib_request.Request(sanitize_url(url), *args, **kwargs)
+ url, auth_header = extract_basic_auth(escape_url(sanitize_url(url)))
+ if auth_header is not None:
+ headers = args[1] if len(args) >= 2 else kwargs.setdefault('headers', {})
+ headers['Authorization'] = auth_header
+ return compat_urllib_request.Request(url, *args, **kwargs)
def expand_path(s):
r'&([^&;]+;)', lambda m: _htmlentity_transform(m.group(1)), s)
+def escapeHTML(text):
+ return (
+ text
+ .replace('&', '&')
+ .replace('<', '<')
+ .replace('>', '>')
+ .replace('"', '"')
+ .replace("'", ''')
+ )
+
+
def process_communicate_or_kill(p, *args, **kwargs):
try:
return p.communicate(*args, **kwargs)
return optval
-def formatSeconds(secs, delim=':'):
+def formatSeconds(secs, delim=':', msec=False):
if secs > 3600:
- return '%d%s%02d%s%02d' % (secs // 3600, delim, (secs % 3600) // 60, delim, secs % 60)
+ ret = '%d%s%02d%s%02d' % (secs // 3600, delim, (secs % 3600) // 60, delim, secs % 60)
elif secs > 60:
- return '%d%s%02d' % (secs // 60, delim, secs % 60)
+ ret = '%d%s%02d' % (secs // 60, delim, secs % 60)
else:
- return '%d' % secs
+ ret = '%d' % secs
+ return '%s.%03d' % (ret, secs % 1) if msec else ret
def make_HTTPS_handler(params, **kwargs):
pass
+network_exceptions = [compat_urllib_error.URLError, compat_http_client.HTTPException, socket.error]
+if hasattr(ssl, 'CertificateError'):
+ network_exceptions.append(ssl.CertificateError)
+network_exceptions = tuple(network_exceptions)
+
+
class ExtractorError(YoutubeDLError):
"""Error during info extraction."""
If expected is set, this is a normal error message and most likely not a bug in yt-dlp.
"""
- if sys.exc_info()[0] in (compat_urllib_error.URLError, socket.timeout, UnavailableVideoError):
+ if sys.exc_info()[0] in network_exceptions:
expected = True
if video_id is not None:
msg = video_id + ': ' + msg
pass
+class ThrottledDownload(YoutubeDLError):
+ """ Download speed below --throttled-rate. """
+ pass
+
+
class MaxDownloadsReached(YoutubeDLError):
""" --max-downloads limit has been reached. """
pass
return unrecognized
-class PagedList(object):
+class LazyList(collections.abc.Sequence):
+ ''' Lazy immutable list from an iterable
+ Note that slices of a LazyList are lists and not LazyList'''
+
+ def __init__(self, iterable):
+ self.__iterable = iter(iterable)
+ self.__cache = []
+ self.__reversed = False
+
+ def __iter__(self):
+ if self.__reversed:
+ # We need to consume the entire iterable to iterate in reverse
+ yield from self.exhaust()
+ return
+ yield from self.__cache
+ for item in self.__iterable:
+ self.__cache.append(item)
+ yield item
+
+ def __exhaust(self):
+ self.__cache.extend(self.__iterable)
+ return self.__cache
+
+ def exhaust(self):
+ ''' Evaluate the entire iterable '''
+ return self.__exhaust()[::-1 if self.__reversed else 1]
+
+ @staticmethod
+ def __reverse_index(x):
+ return None if x is None else -(x + 1)
+
+ def __getitem__(self, idx):
+ if isinstance(idx, slice):
+ if self.__reversed:
+ idx = slice(self.__reverse_index(idx.start), self.__reverse_index(idx.stop), -(idx.step or 1))
+ start, stop, step = idx.start, idx.stop, idx.step or 1
+ elif isinstance(idx, int):
+ if self.__reversed:
+ idx = self.__reverse_index(idx)
+ start, stop, step = idx, idx, 0
+ else:
+ raise TypeError('indices must be integers or slices')
+ if ((start or 0) < 0 or (stop or 0) < 0
+ or (start is None and step < 0)
+ or (stop is None and step > 0)):
+ # We need to consume the entire iterable to be able to slice from the end
+ # Obviously, never use this with infinite iterables
+ return self.__exhaust()[idx]
+
+ n = max(start or 0, stop or 0) - len(self.__cache) + 1
+ if n > 0:
+ self.__cache.extend(itertools.islice(self.__iterable, n))
+ return self.__cache[idx]
+
+ def __bool__(self):
+ try:
+ self[-1] if self.__reversed else self[0]
+ except IndexError:
+ return False
+ return True
+
+ def __len__(self):
+ self.exhaust()
+ return len(self.__cache)
+
+ def reverse(self):
+ self.__reversed = not self.__reversed
+ return self
+
+ def __repr__(self):
+ # repr and str should mimic a list. So we exhaust the iterable
+ return repr(self.exhaust())
+
+ def __str__(self):
+ return repr(self.exhaust())
+
+
+class PagedList:
def __len__(self):
# This is only useful for tests
return len(self.getslice())
-
-class OnDemandPagedList(PagedList):
def __init__(self, pagefunc, pagesize, use_cache=True):
self._pagefunc = pagefunc
self._pagesize = pagesize
self._use_cache = use_cache
- if use_cache:
- self._cache = {}
+ self._cache = {}
+
+ def getpage(self, pagenum):
+ page_results = self._cache.get(pagenum) or list(self._pagefunc(pagenum))
+ if self._use_cache:
+ self._cache[pagenum] = page_results
+ return page_results
def getslice(self, start=0, end=None):
- res = []
+ return list(self._getslice(start, end))
+
+ def _getslice(self, start, end):
+ raise NotImplementedError('This method must be implemented by subclasses')
+
+ def __getitem__(self, idx):
+ # NOTE: cache must be enabled if this is used
+ if not isinstance(idx, int) or idx < 0:
+ raise TypeError('indices must be non-negative integers')
+ entries = self.getslice(idx, idx + 1)
+ return entries[0] if entries else None
+
+
+class OnDemandPagedList(PagedList):
+ def _getslice(self, start, end):
for pagenum in itertools.count(start // self._pagesize):
firstid = pagenum * self._pagesize
nextfirstid = pagenum * self._pagesize + self._pagesize
if start >= nextfirstid:
continue
- page_results = None
- if self._use_cache:
- page_results = self._cache.get(pagenum)
- if page_results is None:
- page_results = list(self._pagefunc(pagenum))
- if self._use_cache:
- self._cache[pagenum] = page_results
-
startv = (
start % self._pagesize
if firstid <= start < nextfirstid
else 0)
-
endv = (
((end - 1) % self._pagesize) + 1
if (end is not None and firstid <= end <= nextfirstid)
else None)
+ page_results = self.getpage(pagenum)
if startv != 0 or endv is not None:
page_results = page_results[startv:endv]
- res.extend(page_results)
+ yield from page_results
# A little optimization - if current page is not "full", ie. does
# not contain page_size videos then we can assume that this page
# break out early as well
if end == nextfirstid:
break
- return res
class InAdvancePagedList(PagedList):
def __init__(self, pagefunc, pagecount, pagesize):
- self._pagefunc = pagefunc
self._pagecount = pagecount
- self._pagesize = pagesize
+ PagedList.__init__(self, pagefunc, pagesize, True)
- def getslice(self, start=0, end=None):
- res = []
+ def _getslice(self, start, end):
start_page = start // self._pagesize
end_page = (
self._pagecount if end is None else (end // self._pagesize + 1))
skip_elems = start - start_page * self._pagesize
only_more = None if end is None else end - start
for pagenum in range(start_page, end_page):
- page = list(self._pagefunc(pagenum))
+ page_results = self.getpage(pagenum)
if skip_elems:
- page = page[skip_elems:]
+ page_results = page_results[skip_elems:]
skip_elems = None
if only_more is not None:
- if len(page) < only_more:
- only_more -= len(page)
+ if len(page_results) < only_more:
+ only_more -= len(page_results)
else:
- page = page[:only_more]
- res.extend(page)
+ yield from page_results[:only_more]
break
- res.extend(page)
- return res
+ yield from page_results
def uppercase_escape(s):
def try_get(src, getter, expected_type=None):
- if not isinstance(getter, (list, tuple)):
- getter = [getter]
- for get in getter:
+ for get in variadic(getter):
try:
v = get(src)
except (AttributeError, KeyError, TypeError, IndexError):
def js_to_json(code, vars={}):
# vars is a dict of var, val pairs to substitute
- COMMENT_RE = r'/\*(?:(?!\*/).)*?\*/|//[^\n]*'
+ COMMENT_RE = r'/\*(?:(?!\*/).)*?\*/|//[^\n]*\n'
SKIP_RE = r'\s*(?:{comment})?\s*'.format(comment=COMMENT_RE)
INTEGER_TABLE = (
(r'(?s)^(0[xX][0-9a-fA-F]+){skip}:?$'.format(skip=SKIP_RE), 16),
'description': 'description',
'annotation': 'annotations.xml',
'infojson': 'info.json',
+ 'pl_thumbnail': None,
'pl_description': 'description',
'pl_infojson': 'info.json',
}
# As of [1] format syntax is:
# %[mapping_key][conversion_flags][minimum_width][.precision][length_modifier]type
# 1. https://docs.python.org/2/library/stdtypes.html#string-formatting
-FORMAT_RE = r'''(?x)
- (?<!%)
+STR_FORMAT_RE_TMPL = r'''(?x)
+ (?<!%)(?P<prefix>(?:%%)*)
%
- \({0}\) # mapping key
- (?:[#0\-+ ]+)? # conversion flags (optional)
- (?:\d+)? # minimum field width (optional)
- (?:\.\d+)? # precision (optional)
- [hlL]? # length modifier (optional)
- (?P<type>[diouxXeEfFgGcrs%]) # conversion type
+ (?P<has_key>\((?P<key>{0})\))? # mapping key
+ (?P<format>
+ (?:[#0\-+ ]+)? # conversion flags (optional)
+ (?:\d+)? # minimum field width (optional)
+ (?:\.\d+)? # precision (optional)
+ [hlL]? # length modifier (optional)
+ {1} # conversion type
+ )
'''
+STR_FORMAT_TYPES = 'diouxXeEfFgGcrs'
+
+
def limit_length(s, length):
""" Add ellipses to overly long strings """
if s is None:
if not codecs_str:
return {}
split_codecs = list(filter(None, map(
- lambda str: str.strip(), codecs_str.strip().strip(',').split(','))))
+ str.strip, codecs_str.strip().strip(',').split(','))))
vcodec, acodec = None, None
for full_codec in split_codecs:
codec = full_codec.split('.')[0]
return '\n'.join(format_str % tuple(row) for row in table)
-def _match_one(filter_part, dct):
+def _match_one(filter_part, dct, incomplete):
+ # TODO: Generalize code with YoutubeDL._build_format_filter
+ STRING_OPERATORS = {
+ '*=': operator.contains,
+ '^=': lambda attr, value: attr.startswith(value),
+ '$=': lambda attr, value: attr.endswith(value),
+ '~=': lambda attr, value: re.search(value, attr),
+ }
COMPARISON_OPERATORS = {
+ **STRING_OPERATORS,
+ '<=': operator.le, # "<=" must be defined above "<"
'<': operator.lt,
- '<=': operator.le,
- '>': operator.gt,
'>=': operator.ge,
+ '>': operator.gt,
'=': operator.eq,
- '!=': operator.ne,
}
+
operator_rex = re.compile(r'''(?x)\s*
(?P<key>[a-z_]+)
- \s*(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
+ \s*(?P<negation>!\s*)?(?P<op>%s)(?P<none_inclusive>\s*\?)?\s*
(?:
(?P<intval>[0-9.]+(?:[kKmMgGtTpPeEzZyY]i?[Bb]?)?)|
- (?P<quote>["\'])(?P<quotedstrval>(?:\\.|(?!(?P=quote)|\\).)+?)(?P=quote)|
- (?P<strval>(?![0-9.])[a-z0-9A-Z]*)
+ (?P<quote>["\'])(?P<quotedstrval>.+?)(?P=quote)|
+ (?P<strval>.+?)
)
\s*$
''' % '|'.join(map(re.escape, COMPARISON_OPERATORS.keys())))
m = operator_rex.search(filter_part)
if m:
- op = COMPARISON_OPERATORS[m.group('op')]
+ unnegated_op = COMPARISON_OPERATORS[m.group('op')]
+ if m.group('negation'):
+ op = lambda attr, value: not unnegated_op(attr, value)
+ else:
+ op = unnegated_op
actual_value = dct.get(m.group('key'))
if (m.group('quotedstrval') is not None
or m.group('strval') is not None
# https://github.com/ytdl-org/youtube-dl/issues/11082).
or actual_value is not None and m.group('intval') is not None
and isinstance(actual_value, compat_str)):
- if m.group('op') not in ('=', '!='):
- raise ValueError(
- 'Operator %s does not support string values!' % m.group('op'))
comparison_value = m.group('quotedstrval') or m.group('strval') or m.group('intval')
quote = m.group('quote')
if quote is not None:
comparison_value = comparison_value.replace(r'\%s' % quote, quote)
else:
+ if m.group('op') in STRING_OPERATORS:
+ raise ValueError('Operator %s only supports string values!' % m.group('op'))
try:
comparison_value = int(m.group('intval'))
except ValueError:
'Invalid integer value %r in filter part %r' % (
m.group('intval'), filter_part))
if actual_value is None:
- return m.group('none_inclusive')
+ return incomplete or m.group('none_inclusive')
return op(actual_value, comparison_value)
UNARY_OPERATORS = {
if m:
op = UNARY_OPERATORS[m.group('op')]
actual_value = dct.get(m.group('key'))
+ if incomplete and actual_value is None:
+ return True
return op(actual_value)
raise ValueError('Invalid filter part %r' % filter_part)
-def match_str(filter_str, dct):
- """ Filter a dictionary with a simple string syntax. Returns True (=passes filter) or false """
-
+def match_str(filter_str, dct, incomplete=False):
+ """ Filter a dictionary with a simple string syntax. Returns True (=passes filter) or false
+ When incomplete, all conditions passes on missing fields
+ """
return all(
- _match_one(filter_part, dct) for filter_part in filter_str.split('&'))
+ _match_one(filter_part.replace(r'\&', '&'), dct, incomplete)
+ for filter_part in re.split(r'(?<!\\)&', filter_str))
def match_filter_func(filter_str):
- def _match_func(info_dict):
- if match_str(filter_str, info_dict):
+ def _match_func(info_dict, *args, **kwargs):
+ if match_str(filter_str, info_dict, *args, **kwargs):
return None
else:
video_title = info_dict.get('title', info_dict.get('id', 'video'))
assert isinstance(keys, (list, tuple))
for key_list in keys:
- if isinstance(key_list, compat_str):
- key_list = (key_list,)
arg_list = list(filter(
lambda x: x is not None,
- [argdict.get(key.lower()) for key in key_list]))
+ [argdict.get(key.lower()) for key in variadic(key_list)]))
if arg_list:
return [arg for args in arg_list for arg in args]
return default
return path
-def format_field(obj, field, template='%s', ignore=(None, ''), default='', func=None):
- val = obj.get(field, default)
+def format_field(obj, field=None, template='%s', ignore=(None, ''), default='', func=None):
+ if field is None:
+ val = obj if obj is not None else default
+ else:
+ val = obj.get(field, default)
if func and val not in ignore:
val = func(val)
return template % val if val not in ignore else default
return os.path.abspath(path)
-def load_plugins(name, type, namespace):
+def load_plugins(name, suffix, namespace):
plugin_info = [None]
classes = []
try:
name, [os.path.join(get_executable_path(), 'ytdlp_plugins')])
plugins = imp.load_module(name, *plugin_info)
for name in dir(plugins):
- if not name.endswith(type):
+ if name in namespace:
+ continue
+ if not name.endswith(suffix):
continue
klass = getattr(plugins, name)
classes.append(klass)
return classes
-def traverse_dict(dictn, keys, casesense=True):
- keys = list(keys)[::-1]
- while keys:
- key = keys.pop()
- if isinstance(dictn, dict):
- if not casesense:
- dictn = {k.lower(): v for k, v in dictn.items()}
- key = key.lower()
- dictn = dictn.get(key)
- elif isinstance(dictn, (list, tuple, compat_str)):
- key, n = int_or_none(key), len(dictn)
- if key is not None and -n <= key < n:
- dictn = dictn[key]
- else:
- dictn = None
- else:
+def traverse_obj(
+ obj, *path_list, default=None, expected_type=None, get_all=True,
+ casesense=True, is_user_input=False, traverse_string=False):
+ ''' Traverse nested list/dict/tuple
+ @param path_list A list of paths which are checked one by one.
+ Each path is a list of keys where each key is a string,
+ a tuple of strings or "...". When a tuple is given,
+ all the keys given in the tuple are traversed, and
+ "..." traverses all the keys in the object
+ @param default Default value to return
+ @param expected_type Only accept final value of this type (Can also be any callable)
+ @param get_all Return all the values obtained from a path or only the first one
+ @param casesense Whether to consider dictionary keys as case sensitive
+ @param is_user_input Whether the keys are generated from user input. If True,
+ strings are converted to int/slice if necessary
+ @param traverse_string Whether to traverse inside strings. If True, any
+ non-compatible object will also be converted into a string
+ # TODO: Write tests
+ '''
+ if not casesense:
+ _lower = lambda k: (k.lower() if isinstance(k, str) else k)
+ path_list = (map(_lower, variadic(path)) for path in path_list)
+
+ def _traverse_obj(obj, path, _current_depth=0):
+ nonlocal depth
+ if obj is None:
return None
- return dictn
+ path = tuple(variadic(path))
+ for i, key in enumerate(path):
+ if isinstance(key, (list, tuple)):
+ obj = [_traverse_obj(obj, sub_key, _current_depth) for sub_key in key]
+ key = ...
+ if key is ...:
+ obj = (obj.values() if isinstance(obj, dict)
+ else obj if isinstance(obj, (list, tuple, LazyList))
+ else str(obj) if traverse_string else [])
+ _current_depth += 1
+ depth = max(depth, _current_depth)
+ return [_traverse_obj(inner_obj, path[i + 1:], _current_depth) for inner_obj in obj]
+ elif isinstance(obj, dict) and not (is_user_input and key == ':'):
+ obj = (obj.get(key) if casesense or (key in obj)
+ else next((v for k, v in obj.items() if _lower(k) == key), None))
+ else:
+ if is_user_input:
+ key = (int_or_none(key) if ':' not in key
+ else slice(*map(int_or_none, key.split(':'))))
+ if key == slice(None):
+ return _traverse_obj(obj, (..., *path[i + 1:]), _current_depth)
+ if not isinstance(key, (int, slice)):
+ return None
+ if not isinstance(obj, (list, tuple, LazyList)):
+ if not traverse_string:
+ return None
+ obj = str(obj)
+ try:
+ obj = obj[key]
+ except IndexError:
+ return None
+ return obj
+
+ if isinstance(expected_type, type):
+ type_test = lambda val: val if isinstance(val, expected_type) else None
+ elif expected_type is not None:
+ type_test = expected_type
+ else:
+ type_test = lambda val: val
+
+ for path in path_list:
+ depth = 0
+ val = _traverse_obj(obj, path)
+ if val is not None:
+ if depth:
+ for _ in range(depth - 1):
+ val = itertools.chain.from_iterable(v for v in val if v is not None)
+ val = [v for v in map(type_test, val) if v is not None]
+ if val:
+ return val if get_all else val[0]
+ else:
+ val = type_test(val)
+ if val is not None:
+ return val
+ return default
+
+
+def traverse_dict(dictn, keys, casesense=True):
+ ''' For backward compatibility. Do not use '''
+ return traverse_obj(dictn, keys, casesense=casesense,
+ is_user_input=True, traverse_string=True)
+
+
+def variadic(x, allowed_types=(str, bytes)):
+ return x if isinstance(x, collections.abc.Iterable) and not isinstance(x, allowed_types) else (x,)
projects / yt-dlp.git / blobdiff