# coding: utf-8
import collections
-import itertools
+import errno
import io
+import itertools
import json
-import netrc
+import os.path
import re
import socket
import string
import struct
import traceback
+import xml.etree.ElementTree
import zlib
from .common import InfoExtractor, SearchInfoExtractor
from .subtitles import SubtitlesInfoExtractor
from ..utils import (
+ compat_chr,
compat_http_client,
compat_parse_qs,
compat_urllib_error,
compat_urllib_parse,
compat_urllib_request,
+ compat_urlparse,
compat_str,
clean_html,
+ get_cachedir,
get_element_by_id,
ExtractorError,
unescapeHTML,
unified_strdate,
orderedSet,
+ write_json_file,
)
class YoutubeBaseInfoExtractor(InfoExtractor):
self._downloader.report_warning(u'unable to fetch login page: %s' % compat_str(err))
return False
- galx = None
- dsh = None
- match = re.search(re.compile(r'<input.+?name="GALX".+?value="(.+?)"', re.DOTALL), login_page)
- if match:
- galx = match.group(1)
- match = re.search(re.compile(r'<input.+?name="dsh".+?value="(.+?)"', re.DOTALL), login_page)
- if match:
- dsh = match.group(1)
+ galx = self._search_regex(r'(?s)<input.+?name="GALX".+?value="(.+?)"',
+ login_page, u'Login GALX parameter')
# Log in
login_form_strs = {
u'checkConnection': u'',
u'checkedDomains': u'youtube',
u'dnConn': u'',
- u'dsh': dsh,
u'pstMsg': u'0',
u'rmShown': u'1',
u'secTok': u'',
class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
IE_DESC = u'YouTube.com'
- _VALID_URL = r"""^
+ _VALID_URL = r"""(?x)^
(
- (?:https?://)? # http(s):// (optional)
- (?:(?:(?:(?:\w+\.)?youtube(?:-nocookie)?\.com/|
+ (?:https?://|//)? # http(s):// or protocol-independent URL (optional)
+ (?:(?:(?:(?:\w+\.)?[yY][oO][uU][tT][uU][bB][eE](?:-nocookie)?\.com/|
tube\.majestyc\.net/|
youtube\.googleapis\.com/) # the various hostnames, with wildcard subdomains
(?:.*?\#/)? # handle anchor (#/) redirect urls
'136': 'mp4',
'137': 'mp4',
'138': 'mp4',
- '139': 'mp4',
- '140': 'mp4',
- '141': 'mp4',
'160': 'mp4',
+ # Dash mp4 audio
+ '139': 'm4a',
+ '140': 'm4a',
+ '141': 'm4a',
+
# Dash webm
'171': 'webm',
'172': 'webm',
u"description": u"test chars: \"'/\\ä↭𝕐\n\nThis is a test video for youtube-dl.\n\nFor more information, contact phihag@phihag.de ."
}
},
- {
- u"url": u"http://www.youtube.com/watch?v=1ltcDfZMA3U",
- u"file": u"1ltcDfZMA3U.flv",
- u"note": u"Test VEVO video (#897)",
- u"info_dict": {
- u"upload_date": u"20070518",
- u"title": u"Maps - It Will Find You",
- u"description": u"Music video by Maps performing It Will Find You.",
- u"uploader": u"MuteUSA",
- u"uploader_id": u"MuteUSA"
- }
- },
{
u"url": u"http://www.youtube.com/watch?v=UxxajLWwzqY",
u"file": u"UxxajLWwzqY.mp4",
u"info_dict": {
u"upload_date": u"20120506",
u"title": u"Icona Pop - I Love It (feat. Charli XCX) [OFFICIAL VIDEO]",
- u"description": u"md5:3e2666e0a55044490499ea45fe9037b7",
+ u"description": u"md5:5b292926389560516e384ac437c0ec07",
u"uploader": u"Icona Pop",
u"uploader_id": u"IconaPop"
}
}
},
{
- u'url': u'https://www.youtube.com/watch?v=TGi3HqYrWHE',
- u'file': u'TGi3HqYrWHE.mp4',
- u'note': u'm3u8 video',
- u'info_dict': {
- u'title': u'Triathlon - Men - London 2012 Olympic Games',
- u'description': u'- Men - TR02 - Triathlon - 07 August 2012 - London 2012 Olympic Games',
- u'uploader': u'olympic',
- u'upload_date': u'20120807',
- u'uploader_id': u'olympic',
- },
- u'params': {
- u'skip_download': True,
- },
+ u"url": u"//www.YouTube.com/watch?v=yZIXLfi8CZQ",
+ u"file": u"yZIXLfi8CZQ.mp4",
+ u"note": u"Embed-only video (#1746)",
+ u"info_dict": {
+ u"upload_date": u"20120608",
+ u"title": u"Principal Sexually Assaults A Teacher - Episode 117 - 8th June 2012",
+ u"description": u"md5:09b78bd971f1e3e289601dfba15ca4f7",
+ u"uploader": u"SET India",
+ u"uploader_id": u"setindia"
+ }
},
]
def suitable(cls, url):
"""Receives a URL and returns True if suitable for this IE."""
if YoutubePlaylistIE.suitable(url): return False
- return re.match(cls._VALID_URL, url, re.VERBOSE) is not None
+ return re.match(cls._VALID_URL, url) is not None
def __init__(self, *args, **kwargs):
super(YoutubeIE, self).__init__(*args, **kwargs)
- self._jsplayer_cache = {}
+ self._player_cache = {}
def report_video_webpage_download(self, video_id):
"""Report attempt to download video webpage."""
"""Indicate the download will use the RTMP protocol."""
self.to_screen(u'RTMP download detected')
- def _extract_signature_function(self, video_id, player_url):
- id_m = re.match(r'.*-(?P<id>[^.]+)\.(?P<ext>[^.]+)$', player_url)
+ def _extract_signature_function(self, video_id, player_url, slen):
+ id_m = re.match(r'.*-(?P<id>[a-zA-Z0-9_-]+)\.(?P<ext>[a-z]+)$',
+ player_url)
player_type = id_m.group('ext')
player_id = id_m.group('id')
+ # Read from filesystem cache
+ func_id = '%s_%s_%d' % (player_type, player_id, slen)
+ assert os.path.basename(func_id) == func_id
+ cache_dir = get_cachedir(self._downloader.params)
+
+ cache_enabled = cache_dir is not None
+ if cache_enabled:
+ cache_fn = os.path.join(os.path.expanduser(cache_dir),
+ u'youtube-sigfuncs',
+ func_id + '.json')
+ try:
+ with io.open(cache_fn, 'r', encoding='utf-8') as cachef:
+ cache_spec = json.load(cachef)
+ return lambda s: u''.join(s[i] for i in cache_spec)
+ except IOError:
+ pass # No cache available
+
if player_type == 'js':
code = self._download_webpage(
player_url, video_id,
- note=u'Downloading %s player %s' % (player_type, jsplayer_id),
+ note=u'Downloading %s player %s' % (player_type, player_id),
errnote=u'Download of %s failed' % player_url)
- return self._parse_sig_js(code)
- elif player_tpye == 'swf':
+ res = self._parse_sig_js(code)
+ elif player_type == 'swf':
urlh = self._request_webpage(
player_url, video_id,
- note=u'Downloading %s player %s' % (player_type, jsplayer_id),
+ note=u'Downloading %s player %s' % (player_type, player_id),
errnote=u'Download of %s failed' % player_url)
code = urlh.read()
- return self._parse_sig_swf(code)
+ res = self._parse_sig_swf(code)
else:
assert False, 'Invalid player type %r' % player_type
+ if cache_enabled:
+ try:
+ test_string = u''.join(map(compat_chr, range(slen)))
+ cache_res = res(test_string)
+ cache_spec = [ord(c) for c in cache_res]
+ try:
+ os.makedirs(os.path.dirname(cache_fn))
+ except OSError as ose:
+ if ose.errno != errno.EEXIST:
+ raise
+ write_json_file(cache_spec, cache_fn)
+ except Exception:
+ tb = traceback.format_exc()
+ self._downloader.report_warning(
+ u'Writing cache to %r failed: %s' % (cache_fn, tb))
+
+ return res
+
+ def _print_sig_code(self, func, slen):
+ def gen_sig_code(idxs):
+ def _genslice(start, end, step):
+ starts = u'' if start == 0 else str(start)
+ ends = (u':%d' % (end+step)) if end + step >= 0 else u':'
+ steps = u'' if step == 1 else (u':%d' % step)
+ return u's[%s%s%s]' % (starts, ends, steps)
+
+ step = None
+ start = '(Never used)' # Quelch pyflakes warnings - start will be
+ # set as soon as step is set
+ for i, prev in zip(idxs[1:], idxs[:-1]):
+ if step is not None:
+ if i - prev == step:
+ continue
+ yield _genslice(start, prev, step)
+ step = None
+ continue
+ if i - prev in [-1, 1]:
+ step = i - prev
+ start = prev
+ continue
+ else:
+ yield u's[%d]' % prev
+ if step is None:
+ yield u's[%d]' % i
+ else:
+ yield _genslice(start, i, step)
+
+ test_string = u''.join(map(compat_chr, range(slen)))
+ cache_res = func(test_string)
+ cache_spec = [ord(c) for c in cache_res]
+ expr_code = u' + '.join(gen_sig_code(cache_spec))
+ code = u'if len(s) == %d:\n return %s\n' % (slen, expr_code)
+ self.to_screen(u'Extracted signature function:\n' + code)
+
def _parse_sig_js(self, jscode):
funcname = self._search_regex(
r'signature=([a-zA-Z]+)', jscode,
def interpret_statement(stmt, local_vars, allow_recursion=20):
if allow_recursion < 0:
- raise ExctractorError(u'Recursion limit reached')
+ raise ExtractorError(u'Recursion limit reached')
if stmt.startswith(u'var '):
stmt = stmt[len(u'var '):]
for tag_code, tag in extract_tags(content)
if tag_code == 82)
p = code_tag.index(b'\0', 4) + 1
+ code_reader = io.BytesIO(code_tag[p:])
# Parse ABC (AVM2 ByteCode)
- def read_int(data=None, pos=None):
- if hasattr(data, 'read'):
- assert pos is None
-
- res = 0
- shift = 0
- for _ in range(5):
- buf = data.read(1)
- assert len(buf) == 1
- b = struct.unpack('<B', buf)[0]
- res = res | ((b & 0x7f) << shift)
- if b & 0x80 == 0:
- break
- shift += 7
- return res
-
- if data is None:
- data = code_tag
- if pos is None:
- pos = p
+ def read_int(reader=None):
+ if reader is None:
+ reader = code_reader
res = 0
shift = 0
for _ in range(5):
- b = struct.unpack('<B', data[pos:pos+1])[0]
- pos += 1
+ buf = reader.read(1)
+ assert len(buf) == 1
+ b = struct.unpack('<B', buf)[0]
res = res | ((b & 0x7f) << shift)
if b & 0x80 == 0:
break
shift += 7
- return (res, pos)
- assert read_int(b'\x00', 0) == (0, 1)
- assert read_int(b'\x10', 0) == (16, 1)
- assert read_int(b'\x34', 0) == (0x34, 1)
- assert read_int(b'\xb4\x12', 0) == (0x12 * 0x80 + 0x34, 2)
- assert read_int(b'\xff\xff\xff\x00', 0) == (0x1fffff, 4)
-
- def u30(*args, **kwargs):
- res = read_int(*args, **kwargs)
- if isinstance(res, tuple):
- assert res[0] & 0xf0000000 == 0
- else:
- assert res & 0xf0000000 == 0
+ return res
+
+ def u30(reader=None):
+ res = read_int(reader)
+ assert res & 0xf0000000 == 0
return res
u32 = read_int
- def s32(data=None, pos=None):
- v, pos = read_int(data, pos)
+ def s32(reader=None):
+ v = read_int(reader)
if v & 0x80000000 != 0:
v = - ((v ^ 0xffffffff) + 1)
- return (v, pos)
- assert s32(b'\xff\xff\xff\xff\x0f', 0) == (-1, 5)
-
- def string():
- slen, p = u30()
- return (code_tag[p:p+slen].decode('utf-8'), p + slen)
-
- def read_byte(data=None, pos=None):
- if data is None:
- data = code_tag
- if pos is None:
- pos = p
- res = struct.unpack('<B', data[pos:pos+1])[0]
- return (res, pos + 1)
+ return v
+
+ def read_string(reader=None):
+ if reader is None:
+ reader = code_reader
+ slen = u30(reader)
+ resb = reader.read(slen)
+ assert len(resb) == slen
+ return resb.decode('utf-8')
+
+ def read_bytes(count, reader=None):
+ if reader is None:
+ reader = code_reader
+ resb = reader.read(count)
+ assert len(resb) == count
+ return resb
+
+ def read_byte(reader=None):
+ resb = read_bytes(1, reader=reader)
+ res = struct.unpack('<B', resb)[0]
+ return res
# minor_version + major_version
- p += 2 + 2
+ read_bytes(2 + 2)
# Constant pool
- int_count, p = u30()
+ int_count = u30()
for _c in range(1, int_count):
- _, p = s32()
- uint_count, p = u30()
+ s32()
+ uint_count = u30()
for _c in range(1, uint_count):
- _, p = u32()
- double_count, p = u30()
- p += (double_count-1) * 8
- string_count, p = u30()
+ u32()
+ double_count = u30()
+ read_bytes((double_count-1) * 8)
+ string_count = u30()
constant_strings = [u'']
for _c in range(1, string_count):
- s, p = string()
+ s = read_string()
constant_strings.append(s)
- namespace_count, p = u30()
+ namespace_count = u30()
for _c in range(1, namespace_count):
- p += 1 # kind
- _, p = u30() # name
- ns_set_count, p = u30()
+ read_bytes(1) # kind
+ u30() # name
+ ns_set_count = u30()
for _c in range(1, ns_set_count):
- count, p = u30()
+ count = u30()
for _c2 in range(count):
- _, p = u30()
- multiname_count, p = u30()
+ u30()
+ multiname_count = u30()
MULTINAME_SIZES = {
0x07: 2, # QName
0x0d: 2, # QNameA
}
multinames = [u'']
for _c in range(1, multiname_count):
- kind, p = u30()
+ kind = u30()
assert kind in MULTINAME_SIZES, u'Invalid multiname kind %r' % kind
if kind == 0x07:
- namespace_idx, p = u30()
- name_idx, p = u30()
+ u30() # namespace_idx
+ name_idx = u30()
multinames.append(constant_strings[name_idx])
else:
multinames.append('[MULTINAME kind: %d]' % kind)
for _c2 in range(MULTINAME_SIZES[kind]):
- _, p = u30()
+ u30()
# Methods
- method_count, p = u30()
+ method_count = u30()
MethodInfo = collections.namedtuple(
'MethodInfo',
['NEED_ARGUMENTS', 'NEED_REST'])
method_infos = []
for method_id in range(method_count):
- param_count, p = u30()
- _, p = u30() # return type
+ param_count = u30()
+ u30() # return type
for _ in range(param_count):
- _, p = u30() # param type
- _, p = u30() # name index (always 0 for youtube)
- flags, p = read_byte()
+ u30() # param type
+ u30() # name index (always 0 for youtube)
+ flags = read_byte()
if flags & 0x08 != 0:
# Options present
- option_count, p = u30()
+ option_count = u30()
for c in range(option_count):
- _, p = u30() # val
- p += 1 # kind
+ u30() # val
+ read_bytes(1) # kind
if flags & 0x80 != 0:
# Param names present
for _ in range(param_count):
- _, p = u30() # param name
+ u30() # param name
mi = MethodInfo(flags & 0x01 != 0, flags & 0x04 != 0)
method_infos.append(mi)
# Metadata
- metadata_count, p = u30()
+ metadata_count = u30()
for _c in range(metadata_count):
- _, p = u30() # name
- item_count, p = u30()
+ u30() # name
+ item_count = u30()
for _c2 in range(item_count):
- _, p = u30() # key
- _, p = u30() # value
-
- def parse_traits_info(pos=None):
- if pos is None:
- pos = p
- trait_name_idx, pos = u30(pos=pos)
- kind_full, pos = read_byte(pos=pos)
+ u30() # key
+ u30() # value
+
+ def parse_traits_info():
+ trait_name_idx = u30()
+ kind_full = read_byte()
kind = kind_full & 0x0f
attrs = kind_full >> 4
methods = {}
if kind in [0x00, 0x06]: # Slot or Const
- _, pos = u30(pos=pos) # Slot id
- type_name_idx, pos = u30(pos=pos)
- vindex, pos = u30(pos=pos)
+ u30() # Slot id
+ u30() # type_name_idx
+ vindex = u30()
if vindex != 0:
- _, pos = read_byte(pos=pos) # vkind
+ read_byte() # vkind
elif kind in [0x01, 0x02, 0x03]: # Method / Getter / Setter
- _, pos = u30(pos=pos) # disp_id
- method_idx, pos = u30(pos=pos)
+ u30() # disp_id
+ method_idx = u30()
methods[multinames[trait_name_idx]] = method_idx
elif kind == 0x04: # Class
- _, pos = u30(pos=pos) # slot_id
- _, pos = u30(pos=pos) # classi
+ u30() # slot_id
+ u30() # classi
elif kind == 0x05: # Function
- _, pos = u30(pos=pos) # slot_id
- function_idx, pos = u30(pos=pos)
+ u30() # slot_id
+ function_idx = u30()
methods[function_idx] = multinames[trait_name_idx]
else:
raise ExtractorError(u'Unsupported trait kind %d' % kind)
if attrs & 0x4 != 0: # Metadata present
- metadata_count, pos = u30(pos=pos)
+ metadata_count = u30()
for _c3 in range(metadata_count):
- _, pos = u30(pos=pos)
+ u30() # metadata index
- return (methods, pos)
+ return methods
# Classes
TARGET_CLASSNAME = u'SignatureDecipher'
searched_idx = multinames.index(TARGET_CLASSNAME)
searched_class_id = None
- class_count, p = u30()
+ class_count = u30()
for class_id in range(class_count):
- name_idx, p = u30()
+ name_idx = u30()
if name_idx == searched_idx:
# We found the class we're looking for!
searched_class_id = class_id
- _, p = u30() # super_name idx
- flags, p = read_byte()
+ u30() # super_name idx
+ flags = read_byte()
if flags & 0x08 != 0: # Protected namespace is present
- protected_ns_idx, p = u30()
- intrf_count, p = u30()
+ u30() # protected_ns_idx
+ intrf_count = u30()
for _c2 in range(intrf_count):
- _, p = u30()
- _, p = u30() # iinit
- trait_count, p = u30()
+ u30()
+ u30() # iinit
+ trait_count = u30()
for _c2 in range(trait_count):
- _, p = parse_traits_info()
+ parse_traits_info()
if searched_class_id is None:
raise ExtractorError(u'Target class %r not found' %
method_names = {}
method_idxs = {}
for class_id in range(class_count):
- _, p = u30() # cinit
- trait_count, p = u30()
+ u30() # cinit
+ trait_count = u30()
for _c2 in range(trait_count):
- trait_methods, p = parse_traits_info()
+ trait_methods = parse_traits_info()
if class_id == searched_class_id:
method_names.update(trait_methods.items())
method_idxs.update(dict(
for name, idx in trait_methods.items()))
# Scripts
- script_count, p = u30()
+ script_count = u30()
for _c in range(script_count):
- _, p = u30() # init
- trait_count, p = u30()
+ u30() # init
+ trait_count = u30()
for _c2 in range(trait_count):
- _, p = parse_traits_info()
+ parse_traits_info()
# Method bodies
- method_body_count, p = u30()
+ method_body_count = u30()
Method = collections.namedtuple('Method', ['code', 'local_count'])
methods = {}
for _c in range(method_body_count):
- method_idx, p = u30()
- max_stack, p = u30()
- local_count, p = u30()
- init_scope_depth, p = u30()
- max_scope_depth, p = u30()
- code_length, p = u30()
+ method_idx = u30()
+ u30() # max_stack
+ local_count = u30()
+ u30() # init_scope_depth
+ u30() # max_scope_depth
+ code_length = u30()
+ code = read_bytes(code_length)
if method_idx in method_idxs:
- m = Method(code_tag[p:p+code_length], local_count)
+ m = Method(code, local_count)
methods[method_idxs[method_idx]] = m
- p += code_length
- exception_count, p = u30()
+ exception_count = u30()
for _c2 in range(exception_count):
- _, p = u30() # from
- _, p = u30() # to
- _, p = u30() # target
- _, p = u30() # exc_type
- _, p = u30() # var_name
- trait_count, p = u30()
+ u30() # from
+ u30() # to
+ u30() # target
+ u30() # exc_type
+ u30() # var_name
+ trait_count = u30()
for _c2 in range(trait_count):
- _, p = parse_traits_info()
+ parse_traits_info()
- assert p == len(code_tag)
+ assert p + code_reader.tell() == len(code_tag)
assert len(methods) == len(method_idxs)
method_pyfunctions = {}
m = methods[func_name]
def resfunc(args):
- print('Entering function %s(%r)' % (func_name, args))
registers = ['(this)'] + list(args) + [None] * m.local_count
stack = []
coder = io.BytesIO(m.code)
assert isinstance(obj, list)
stack.append(obj[idx])
elif opcode == 128: # coerce
- _ = u30(coder)
+ u30(coder)
elif opcode == 133: # coerce_s
assert isinstance(stack[-1], (type(None), compat_str))
elif opcode == 164: # modulo
initial_function = extract_function(u'decipher')
return lambda s: initial_function([s])
- def _decrypt_signature(self, s, video_id, jsplayer_url, age_gate=False):
+ def _decrypt_signature(self, s, video_id, player_url, age_gate=False):
"""Turn the encrypted s field into a working signature"""
- if jsplayer_url is not None:
+ if player_url is not None:
+ if player_url.startswith(u'//'):
+ player_url = u'https:' + player_url
try:
- if jsplayer_url not in self._jsplayer_cache:
- self._jsplayer_cache[jsplayer_url] = self._extract_signature_function(
- video_id, jsplayer_url
+ player_id = (player_url, len(s))
+ if player_id not in self._player_cache:
+ func = self._extract_signature_function(
+ video_id, player_url, len(s)
)
- return self._jsplayer_cache[jsplayer_url]([s])
- except Exception as e:
+ self._player_cache[player_id] = func
+ func = self._player_cache[player_id]
+ if self._downloader.params.get('youtube_print_sig_code'):
+ self._print_sig_code(func, len(s))
+ return func(s)
+ except Exception:
tb = traceback.format_exc()
- self._downloader.report_warning(u'Automatic signature extraction failed: ' + tb)
+ self._downloader.report_warning(
+ u'Automatic signature extraction failed: ' + tb)
+
+ self._downloader.report_warning(
+ u'Warning: Falling back to static signature algorithm')
- self._downloader.report_warning(u'Warning: Falling back to static signature algorithm')
+ return self._static_decrypt_signature(
+ s, video_id, player_url, age_gate)
+ def _static_decrypt_signature(self, s, video_id, player_url, age_gate):
if age_gate:
# The videos with age protection use another player, so the
# algorithms can be different.
if len(s) == 86:
return s[2:63] + s[82] + s[64:82] + s[63]
- if len(s) == 92:
+ if len(s) == 93:
+ return s[86:29:-1] + s[88] + s[28:5:-1]
+ elif len(s) == 92:
return s[25] + s[3:25] + s[0] + s[26:42] + s[79] + s[43:79] + s[91] + s[80:83]
+ elif len(s) == 91:
+ return s[84:27:-1] + s[86] + s[26:5:-1]
elif len(s) == 90:
return s[25] + s[3:25] + s[2] + s[26:40] + s[77] + s[41:77] + s[89] + s[78:81]
elif len(s) == 89:
elif len(s) == 87:
return s[6:27] + s[4] + s[28:39] + s[27] + s[40:59] + s[2] + s[60:]
elif len(s) == 86:
- return s[5:34] + s[0] + s[35:38] + s[3] + s[39:45] + s[38] + s[46:53] + s[73] + s[54:73] + s[85] + s[74:85] + s[53]
+ return s[80:72:-1] + s[16] + s[71:39:-1] + s[72] + s[38:16:-1] + s[82] + s[15::-1]
elif len(s) == 85:
return s[3:11] + s[0] + s[12:55] + s[84] + s[56:84]
elif len(s) == 84:
- return s[81:36:-1] + s[0] + s[35:2:-1]
+ return s[78:70:-1] + s[14] + s[69:37:-1] + s[70] + s[36:14:-1] + s[80] + s[:14][::-1]
elif len(s) == 83:
- return s[81:64:-1] + s[82] + s[63:52:-1] + s[45] + s[51:45:-1] + s[1] + s[44:1:-1] + s[0]
+ return s[80:63:-1] + s[0] + s[62:0:-1] + s[63]
elif len(s) == 82:
- return s[80:73:-1] + s[81] + s[72:54:-1] + s[2] + s[53:43:-1] + s[0] + s[42:2:-1] + s[43] + s[1] + s[54]
+ return s[80:37:-1] + s[7] + s[36:7:-1] + s[0] + s[6:0:-1] + s[37]
elif len(s) == 81:
return s[56] + s[79:56:-1] + s[41] + s[55:41:-1] + s[80] + s[40:34:-1] + s[0] + s[33:29:-1] + s[34] + s[28:9:-1] + s[29] + s[8:0:-1] + s[9]
elif len(s) == 80:
else:
raise ExtractorError(u'Unable to decrypt signature, key length %d not supported; retrying might work' % (len(s)))
- def _decrypt_signature_age_gate(self, s):
- # The videos with age protection use another player, so the algorithms
- # can be different.
- if len(s) == 86:
- return s[2:63] + s[82] + s[64:82] + s[63]
- else:
- # Fallback to the other algortihms
- return self._decrypt_signature(s)
-
- def _get_available_subtitles(self, video_id):
+ def _get_available_subtitles(self, video_id, webpage):
try:
sub_list = self._download_webpage(
'http://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id,
params = compat_urllib_parse.urlencode({
'lang': lang,
'v': video_id,
- 'fmt': self._downloader.params.get('subtitlesformat'),
+ 'fmt': self._downloader.params.get('subtitlesformat', 'srt'),
+ 'name': l[0].encode('utf-8'),
})
url = u'http://www.youtube.com/api/timedtext?' + params
sub_lang_list[lang] = url
def _get_available_automatic_caption(self, video_id, webpage):
"""We need the webpage for getting the captions url, pass it as an
argument to speed up the process."""
- sub_format = self._downloader.params.get('subtitlesformat')
+ sub_format = self._downloader.params.get('subtitlesformat', 'srt')
self.to_screen(u'%s: Looking for automatic captions' % video_id)
mobj = re.search(r';ytplayer.config = ({.*?});', webpage)
err_msg = u'Couldn\'t find automatic captions for %s' % video_id
list_page = self._download_webpage(list_url, video_id)
caption_list = xml.etree.ElementTree.fromstring(list_page.encode('utf-8'))
original_lang_node = caption_list.find('track')
- if original_lang_node.attrib.get('kind') != 'asr' :
+ if original_lang_node is None or original_lang_node.attrib.get('kind') != 'asr' :
self._downloader.report_warning(u'Video doesn\'t have automatic captions')
return {}
original_lang = original_lang_node.attrib['lang_code']
url_map[itag] = format_url
return url_map
- def _real_extract(self, url):
- if re.match(r'(?:https?://)?[^/]+/watch\?feature=[a-z_]+$', url):
- self._downloader.report_warning(u'Did you forget to quote the URL? Remember that & is a meta-character in most shells, so you want to put the URL in quotes, like youtube-dl \'http://www.youtube.com/watch?feature=foo&v=BaW_jenozKc\' (or simply youtube-dl BaW_jenozKc ).')
+ def _extract_annotations(self, video_id):
+ url = 'https://www.youtube.com/annotations_invideo?features=1&legacy=1&video_id=%s' % video_id
+ return self._download_webpage(url, video_id, note=u'Searching for annotations.', errnote=u'Unable to download video annotations.')
+ def _real_extract(self, url):
# Extract original video URL from URL with redirection, like age verification, using next_url parameter
mobj = re.search(self._NEXT_URL_RE, url)
if mobj:
# We simulate the access to the video from www.youtube.com/v/{video_id}
# this can be viewed without login into Youtube
data = compat_urllib_parse.urlencode({'video_id': video_id,
- 'el': 'embedded',
+ 'el': 'player_embedded',
'gl': 'US',
'hl': 'en',
'eurl': 'https://youtube.googleapis.com/v/' + video_id,
else:
raise ExtractorError(u'"token" parameter not in video info for unknown reason')
+ if 'view_count' in video_info:
+ view_count = int(video_info['view_count'][0])
+ else:
+ view_count = None
+
# Check for "rental" videos
if 'ypc_video_rental_bar_text' in video_info and 'author' not in video_info:
raise ExtractorError(u'"rental" videos not supported')
self._downloader.report_warning(u'unable to extract uploader nickname')
# title
- if 'title' not in video_info:
- raise ExtractorError(u'Unable to extract video title')
- video_title = compat_urllib_parse.unquote_plus(video_info['title'][0])
+ if 'title' in video_info:
+ video_title = compat_urllib_parse.unquote_plus(video_info['title'][0])
+ else:
+ self._downloader.report_warning(u'Unable to extract video title')
+ video_title = u'_'
# thumbnail image
# We try first to get a high quality image:
video_thumbnail = m_thumb.group(1)
elif 'thumbnail_url' not in video_info:
self._downloader.report_warning(u'unable to extract video thumbnail')
- video_thumbnail = ''
+ video_thumbnail = None
else: # don't panic if we can't find it
video_thumbnail = compat_urllib_parse.unquote_plus(video_info['thumbnail_url'][0])
else:
video_duration = compat_urllib_parse.unquote_plus(video_info['length_seconds'][0])
+ # annotations
+ video_annotations = None
+ if self._downloader.params.get('writeannotations', False):
+ video_annotations = self._extract_annotations(video_id)
+
# Decide which formats to download
try:
args = info['args']
# Easy way to know if the 's' value is in url_encoded_fmt_stream_map
# this signatures are encrypted
- m_s = re.search(r'[&,]s=', args['url_encoded_fmt_stream_map'])
+ if 'url_encoded_fmt_stream_map' not in args:
+ raise ValueError(u'No stream_map present') # caught below
+ re_signature = re.compile(r'[&,]s=')
+ m_s = re_signature.search(args['url_encoded_fmt_stream_map'])
if m_s is not None:
self.to_screen(u'%s: Encrypted signatures detected.' % video_id)
video_info['url_encoded_fmt_stream_map'] = [args['url_encoded_fmt_stream_map']]
- m_s = re.search(r'[&,]s=', args.get('adaptive_fmts', u''))
+ m_s = re_signature.search(args.get('adaptive_fmts', u''))
if m_s is not None:
- if 'url_encoded_fmt_stream_map' in video_info:
- video_info['url_encoded_fmt_stream_map'][0] += ',' + args['adaptive_fmts']
- else:
- video_info['url_encoded_fmt_stream_map'] = [args['adaptive_fmts']]
- elif 'adaptive_fmts' in video_info:
- if 'url_encoded_fmt_stream_map' in video_info:
- video_info['url_encoded_fmt_stream_map'][0] += ',' + video_info['adaptive_fmts'][0]
+ if 'adaptive_fmts' in video_info:
+ video_info['adaptive_fmts'][0] += ',' + args['adaptive_fmts']
else:
- video_info['url_encoded_fmt_stream_map'] = video_info['adaptive_fmts']
+ video_info['adaptive_fmts'] = [args['adaptive_fmts']]
except ValueError:
pass
if 'conn' in video_info and video_info['conn'][0].startswith('rtmp'):
self.report_rtmp_download()
video_url_list = [(None, video_info['conn'][0])]
- elif 'url_encoded_fmt_stream_map' in video_info and len(video_info['url_encoded_fmt_stream_map']) >= 1:
- if 'rtmpe%3Dyes' in video_info['url_encoded_fmt_stream_map'][0]:
+ elif len(video_info.get('url_encoded_fmt_stream_map', [])) >= 1 or len(video_info.get('adaptive_fmts', [])) >= 1:
+ encoded_url_map = video_info.get('url_encoded_fmt_stream_map', [''])[0] + ',' + video_info.get('adaptive_fmts',[''])[0]
+ if 'rtmpe%3Dyes' in encoded_url_map:
raise ExtractorError('rtmpe downloads are not supported, see https://github.com/rg3/youtube-dl/issues/343 for more information.', expected=True)
url_map = {}
- for url_data_str in video_info['url_encoded_fmt_stream_map'][0].split(','):
+ for url_data_str in encoded_url_map.split(','):
url_data = compat_parse_qs(url_data_str)
if 'itag' in url_data and 'url' in url_data:
url = url_data['url'][0]
encrypted_sig = url_data['s'][0]
if self._downloader.params.get('verbose'):
if age_gate:
- player_version = self._search_regex(r'-(.+)\.swf$',
- player_url if player_url else 'NOT FOUND',
- 'flash player', fatal=False)
+ if player_url is None:
+ player_version = 'unknown'
+ else:
+ player_version = self._search_regex(
+ r'-(.+)\.swf$', player_url,
+ u'flash player', fatal=False)
player_desc = 'flash player %s' % player_version
else:
- player_version = self._search_regex(r'html5player-(.+?)\.js', video_webpage,
+ player_version = self._search_regex(
+ r'html5player-(.+?)\.js', video_webpage,
'html5 player', fatal=False)
player_desc = u'html5 player %s' % player_version
self.to_screen(u'encrypted signature length %d (%s), itag %s, %s' %
(len(encrypted_sig), parts_sizes, url_data['itag'][0], player_desc))
- if age_gate:
- jsplayer_url = None
- else:
+ if not age_gate:
jsplayer_url_json = self._search_regex(
r'"assets":.+?"js":\s*("[^"]+")',
video_webpage, u'JS player URL')
- jsplayer_url = json.loads(jsplayer_url_json)
+ player_url = json.loads(jsplayer_url_json)
- signature = self._decrypt_signature(encrypted_sig, video_id, jsplayer_url, age_gate)
+ signature = self._decrypt_signature(
+ encrypted_sig, video_id, player_url, age_gate)
url += '&signature=' + signature
if 'ratebypass' not in url:
url += '&ratebypass=yes'
return
else:
- raise ExtractorError(u'no conn or url_encoded_fmt_stream_map information found in video info')
+ raise ExtractorError(u'no conn, hlsvp or url_encoded_fmt_stream_map information found in video info')
results = []
- for format_param, video_real_url in video_url_list:
+ for itag, video_real_url in video_url_list:
# Extension
- video_extension = self._video_extensions.get(format_param, 'flv')
+ video_extension = self._video_extensions.get(itag, 'flv')
- video_format = '{0} - {1}{2}'.format(format_param if format_param else video_extension,
- self._video_dimensions.get(format_param, '???'),
- ' ('+self._special_itags[format_param]+')' if format_param in self._special_itags else '')
+ video_format = '{0} - {1}{2}'.format(itag if itag else video_extension,
+ self._video_dimensions.get(itag, '???'),
+ ' ('+self._special_itags[itag]+')' if itag in self._special_itags else '')
results.append({
'id': video_id,
'title': video_title,
'ext': video_extension,
'format': video_format,
+ 'format_id': itag,
'thumbnail': video_thumbnail,
'description': video_description,
'player_url': player_url,
'subtitles': video_subtitles,
- 'duration': video_duration
+ 'duration': video_duration,
+ 'age_limit': 18 if age_gate else 0,
+ 'annotations': video_annotations,
+ 'webpage_url': 'https://www.youtube.com/watch?v=%s' % video_id,
+ 'view_count': view_count,
})
return results
-class YoutubePlaylistIE(InfoExtractor):
+class YoutubePlaylistIE(YoutubeBaseInfoExtractor):
IE_DESC = u'YouTube.com playlists'
_VALID_URL = r"""(?:
(?:https?://)?
|
((?:PL|EC|UU|FL)[0-9A-Za-z-_]{10,})
)"""
- _TEMPLATE_URL = 'https://gdata.youtube.com/feeds/api/playlists/%s?max-results=%i&start-index=%i&v=2&alt=json&safeSearch=none'
- _MAX_RESULTS = 50
+ _TEMPLATE_URL = 'https://www.youtube.com/playlist?list=%s&page=%s'
+ _MORE_PAGES_INDICATOR = r'data-link-type="next"'
+ _VIDEO_RE = r'href="/watch\?v=([0-9A-Za-z_-]{11})&'
IE_NAME = u'youtube:playlist'
@classmethod
"""Receives a URL and returns True if suitable for this IE."""
return re.match(cls._VALID_URL, url, re.VERBOSE) is not None
+ def _real_initialize(self):
+ self._login()
+
def _real_extract(self, url):
# Extract playlist id
mobj = re.match(self._VALID_URL, url, re.VERBOSE)
if mobj is None:
raise ExtractorError(u'Invalid URL: %s' % url)
-
- # Download playlist videos from API
playlist_id = mobj.group(1) or mobj.group(2)
- videos = []
+
+ # Check if it's a video-specific URL
+ query_dict = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
+ if 'v' in query_dict:
+ video_id = query_dict['v'][0]
+ if self._downloader.params.get('noplaylist'):
+ self.to_screen(u'Downloading just video %s because of --no-playlist' % video_id)
+ return self.url_result('https://www.youtube.com/watch?v=' + video_id, 'Youtube')
+ else:
+ self.to_screen(u'Downloading playlist PL%s - add --no-playlist to just download video %s' % (playlist_id, video_id))
+
+ # Extract the video ids from the playlist pages
+ ids = []
for page_num in itertools.count(1):
- start_index = self._MAX_RESULTS * (page_num - 1) + 1
- if start_index >= 1000:
- self._downloader.report_warning(u'Max number of results reached')
- break
- url = self._TEMPLATE_URL % (playlist_id, self._MAX_RESULTS, start_index)
+ url = self._TEMPLATE_URL % (playlist_id, page_num)
page = self._download_webpage(url, playlist_id, u'Downloading page #%s' % page_num)
+ # The ids are duplicated
+ new_ids = orderedSet(re.findall(self._VIDEO_RE, page))
+ ids.extend(new_ids)
- try:
- response = json.loads(page)
- except ValueError as err:
- raise ExtractorError(u'Invalid JSON in API response: ' + compat_str(err))
-
- if 'feed' not in response:
- raise ExtractorError(u'Got a malformed response from YouTube API')
- playlist_title = response['feed']['title']['$t']
- if 'entry' not in response['feed']:
- # Number of videos is a multiple of self._MAX_RESULTS
+ if re.search(self._MORE_PAGES_INDICATOR, page) is None:
break
- for entry in response['feed']['entry']:
- index = entry['yt$position']['$t']
- if 'media$group' in entry and 'yt$videoid' in entry['media$group']:
- videos.append((
- index,
- 'https://www.youtube.com/watch?v=' + entry['media$group']['yt$videoid']['$t']
- ))
+ playlist_title = self._og_search_title(page)
- videos = [v[1] for v in sorted(videos)]
-
- url_results = [self.url_result(vurl, 'Youtube') for vurl in videos]
- return [self.playlist_result(url_results, playlist_id, playlist_title)]
+ url_results = [self.url_result(vid, 'Youtube') for vid in ids]
+ return self.playlist_result(url_results, playlist_id, playlist_title)
class YoutubeChannelIE(InfoExtractor):
IE_DESC = u'YouTube.com channels'
_VALID_URL = r"^(?:https?://)?(?:youtu\.be|(?:\w+\.)?youtube(?:-nocookie)?\.com)/channel/([0-9A-Za-z_-]+)"
- _TEMPLATE_URL = 'http://www.youtube.com/channel/%s/videos?sort=da&flow=list&view=0&page=%s&gl=US&hl=en'
_MORE_PAGES_INDICATOR = 'yt-uix-load-more'
_MORE_PAGES_URL = 'http://www.youtube.com/c4_browse_ajax?action_load_more_videos=1&flow=list&paging=%s&view=0&sort=da&channel_id=%s'
IE_NAME = u'youtube:channel'
# Download channel page
channel_id = mobj.group(1)
video_ids = []
- pagenum = 1
-
- url = self._TEMPLATE_URL % (channel_id, pagenum)
- page = self._download_webpage(url, channel_id,
- u'Downloading page #%s' % pagenum)
-
- # Extract video identifiers
- ids_in_page = self.extract_videos_from_page(page)
- video_ids.extend(ids_in_page)
+ url = 'https://www.youtube.com/channel/%s/videos' % channel_id
+ channel_page = self._download_webpage(url, channel_id)
+ if re.search(r'channel-header-autogenerated-label', channel_page) is not None:
+ autogenerated = True
+ else:
+ autogenerated = False
- # Download any subsequent channel pages using the json-based channel_ajax query
- if self._MORE_PAGES_INDICATOR in page:
+ if autogenerated:
+ # The videos are contained in a single page
+ # the ajax pages can't be used, they are empty
+ video_ids = self.extract_videos_from_page(channel_page)
+ else:
+ # Download all channel pages using the json-based channel_ajax query
for pagenum in itertools.count(1):
url = self._MORE_PAGES_URL % (pagenum, channel_id)
page = self._download_webpage(url, channel_id,
u'Downloading page #%s' % pagenum)
-
+
page = json.loads(page)
-
+
ids_in_page = self.extract_videos_from_page(page['content_html'])
video_ids.extend(ids_in_page)
-
- if self._MORE_PAGES_INDICATOR not in page['load_more_widget_html']:
+
+ if self._MORE_PAGES_INDICATOR not in page['load_more_widget_html']:
break
self._downloader.to_screen(u'[youtube] Channel %s: Found %i videos' % (channel_id, len(video_ids)))
class YoutubeUserIE(InfoExtractor):
IE_DESC = u'YouTube.com user videos (URL or "ytuser" keyword)'
- _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?)|ytuser:)(?!feed/)([A-Za-z0-9_-]+)'
+ _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?(?!(?:attribution_link|watch)(?:$|[^a-z_A-Z0-9-])))|ytuser:)(?!feed/)([A-Za-z0-9_-]+)'
_TEMPLATE_URL = 'http://gdata.youtube.com/feeds/api/users/%s'
_GDATA_PAGE_SIZE = 50
_GDATA_URL = 'http://gdata.youtube.com/feeds/api/users/%s/uploads?max-results=%d&start-index=%d&alt=json'
videos = [self.url_result('http://www.youtube.com/watch?v=%s' % id, 'Youtube') for id in video_ids]
return self.playlist_result(videos, query)
+class YoutubeSearchDateIE(YoutubeSearchIE):
+ _API_URL = 'https://gdata.youtube.com/feeds/api/videos?q=%s&start-index=%i&max-results=50&v=2&alt=jsonc&orderby=published'
+ _SEARCH_KEY = 'ytsearchdate'
+ IE_DESC = u'YouTube.com searches, newest videos first'
class YoutubeShowIE(InfoExtractor):
IE_DESC = u'YouTube.com (multi-season) shows'
webpage = self._download_webpage('https://www.youtube.com/my_favorites', 'Youtube Favourites videos')
playlist_id = self._search_regex(r'list=(.+?)["&]', webpage, u'favourites playlist id')
return self.url_result(playlist_id, 'YoutubePlaylist')
+
+
+class YoutubeTruncatedURLIE(InfoExtractor):
+ IE_NAME = 'youtube:truncated_url'
+ IE_DESC = False # Do not list
+ _VALID_URL = r'(?:https?://)?[^/]+/watch\?feature=[a-z_]+$'
+
+ def _real_extract(self, url):
+ raise ExtractorError(
+ u'Did you forget to quote the URL? Remember that & is a meta '
+ u'character in most shells, so you want to put the URL in quotes, '
+ u'like youtube-dl '
+ u'\'http://www.youtube.com/watch?feature=foo&v=BaW_jenozKc\''
+ u' (or simply youtube-dl BaW_jenozKc ).',
+ expected=True)
projects / youtube-dl / blobdiff