# 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):
'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"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"uploader_id": u"justintimberlakeVEVO"
}
},
- {
- 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,
- },
- },
]
"""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>[a-zA-Z0-9]+)\.(?P<ext>[a-z]+)$',
+ def _extract_signature_function(self, video_id, player_url, slen):
+ id_m = re.match(r'.*-(?P<id>[a-zA-Z0-9_-]+)\.(?P<ext>[a-z]+)$',
player_url)
player_type = id_m.group('ext')
player_id = id_m.group('id')
- # TODO read from filesystem cache
+ # 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(
note=u'Downloading %s player %s' % (player_type, player_id),
errnote=u'Download of %s failed' % player_url)
res = self._parse_sig_js(code)
- elif player_tpye == 'swf':
+ elif player_type == 'swf':
urlh = self._request_webpage(
player_url, video_id,
note=u'Downloading %s player %s' % (player_type, player_id),
else:
assert False, 'Invalid player type %r' % player_type
- # TODO write cache
+ 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 '):]
v = - ((v ^ 0xffffffff) + 1)
return v
- def string(reader=None):
+ def read_string(reader=None):
if reader is None:
reader = code_reader
slen = u30(reader)
return res
# minor_version + major_version
- _ = read_bytes(2 + 2)
+ read_bytes(2 + 2)
# Constant pool
int_count = u30()
for _c in range(1, int_count):
- _ = s32()
+ s32()
uint_count = u30()
for _c in range(1, uint_count):
- _ = u32()
+ u32()
double_count = u30()
- _ = read_bytes((double_count-1) * 8)
+ read_bytes((double_count-1) * 8)
string_count = u30()
constant_strings = [u'']
for _c in range(1, string_count):
- s = string()
+ s = read_string()
constant_strings.append(s)
namespace_count = u30()
for _c in range(1, namespace_count):
- _ = read_bytes(1) # kind
- _ = u30() # name
+ read_bytes(1) # kind
+ u30() # name
ns_set_count = u30()
for _c in range(1, ns_set_count):
count = u30()
for _c2 in range(count):
- _ = u30()
+ u30()
multiname_count = u30()
MULTINAME_SIZES = {
0x07: 2, # QName
kind = u30()
assert kind in MULTINAME_SIZES, u'Invalid multiname kind %r' % kind
if kind == 0x07:
- namespace_idx = 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]):
- _ = u30()
+ u30()
# Methods
method_count = u30()
method_infos = []
for method_id in range(method_count):
param_count = u30()
- _ = u30() # return type
+ u30() # return type
for _ in range(param_count):
- _ = u30() # param type
- _ = u30() # name index (always 0 for youtube)
+ u30() # param type
+ u30() # name index (always 0 for youtube)
flags = read_byte()
if flags & 0x08 != 0:
# Options present
option_count = u30()
for c in range(option_count):
- _ = u30() # val
- _ = read_bytes(1) # kind
+ u30() # val
+ read_bytes(1) # kind
if flags & 0x80 != 0:
# Param names present
for _ in range(param_count):
- _ = u30() # param name
+ u30() # param name
mi = MethodInfo(flags & 0x01 != 0, flags & 0x04 != 0)
method_infos.append(mi)
# Metadata
metadata_count = u30()
for _c in range(metadata_count):
- _ = u30() # name
+ u30() # name
item_count = u30()
for _c2 in range(item_count):
- _ = u30() # key
- _ = u30() # value
+ u30() # key
+ u30() # value
def parse_traits_info():
trait_name_idx = u30()
attrs = kind_full >> 4
methods = {}
if kind in [0x00, 0x06]: # Slot or Const
- _ = u30() # Slot id
- type_name_idx = u30()
+ u30() # Slot id
+ u30() # type_name_idx
vindex = u30()
if vindex != 0:
- _ = read_byte() # vkind
+ read_byte() # vkind
elif kind in [0x01, 0x02, 0x03]: # Method / Getter / Setter
- _ = u30() # disp_id
+ u30() # disp_id
method_idx = u30()
methods[multinames[trait_name_idx]] = method_idx
elif kind == 0x04: # Class
- _ = u30() # slot_id
- _ = u30() # classi
+ u30() # slot_id
+ u30() # classi
elif kind == 0x05: # Function
- _ = u30() # slot_id
+ u30() # slot_id
function_idx = u30()
methods[function_idx] = multinames[trait_name_idx]
else:
if attrs & 0x4 != 0: # Metadata present
metadata_count = u30()
for _c3 in range(metadata_count):
- _ = u30()
+ u30() # metadata index
return methods
if name_idx == searched_idx:
# We found the class we're looking for!
searched_class_id = class_id
- _ = u30() # super_name idx
+ u30() # super_name idx
flags = read_byte()
if flags & 0x08 != 0: # Protected namespace is present
- protected_ns_idx = u30()
+ u30() # protected_ns_idx
intrf_count = u30()
for _c2 in range(intrf_count):
- _ = u30()
- _ = u30() # iinit
+ u30()
+ u30() # iinit
trait_count = u30()
for _c2 in range(trait_count):
- _ = 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):
- _ = u30() # cinit
+ u30() # cinit
trait_count = u30()
for _c2 in range(trait_count):
trait_methods = parse_traits_info()
# Scripts
script_count = u30()
for _c in range(script_count):
- _ = u30() # init
+ u30() # init
trait_count = u30()
for _c2 in range(trait_count):
- _ = parse_traits_info()
+ parse_traits_info()
# Method bodies
method_body_count = u30()
methods = {}
for _c in range(method_body_count):
method_idx = u30()
- max_stack = u30()
+ u30() # max_stack
local_count = u30()
- init_scope_depth = u30()
- max_scope_depth = u30()
+ u30() # init_scope_depth
+ u30() # max_scope_depth
code_length = u30()
code = read_bytes(code_length)
if method_idx in method_idxs:
methods[method_idxs[method_idx]] = m
exception_count = u30()
for _c2 in range(exception_count):
- _ = u30() # from
- _ = u30() # to
- _ = u30() # target
- _ = u30() # exc_type
- _ = u30() # var_name
+ u30() # from
+ u30() # to
+ u30() # target
+ u30() # exc_type
+ u30() # var_name
trait_count = u30()
for _c2 in range(trait_count):
- _ = parse_traits_info()
+ parse_traits_info()
assert p + code_reader.tell() == len(code_tag)
assert len(methods) == len(method_idxs)
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
if player_url is not None:
try:
- if player_url not in self._player_cache:
+ player_id = (player_url, len(s))
+ if player_id not in self._player_cache:
func = self._extract_signature_function(
- video_id, player_url
+ video_id, player_url, len(s)
)
- self._player_cache[player_url] = func
- return self._player_cache[player_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'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)
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):
try:
sub_list = self._download_webpage(
'lang': lang,
'v': video_id,
'fmt': self._downloader.params.get('subtitlesformat'),
+ 'name': l[0],
})
url = u'http://www.youtube.com/api/timedtext?' + params
sub_lang_list[lang] = url
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:
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
+ if 'url_encoded_fmt_stream_map' not in args:
+ raise ValueError(u'No stream_map present') # caught below
m_s = re.search(r'[&,]s=', args['url_encoded_fmt_stream_map'])
if m_s is not None:
self.to_screen(u'%s: Encrypted signatures detected.' % video_id)
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 None,
- 'flash player', fatal=False)
+ if player_url is None:
+ player_version = 'unknown'
+ else:
+ player_version = self._search_regex(
+ r'-(.+)\.swf$', player_url,
+ u'flash player', fatal=False)
player_desc = 'flash player %s' % player_version
else:
player_version = self._search_regex(
return
else:
- raise ExtractorError(u'no conn or url_encoded_fmt_stream_map information found in video info')
+ raise ExtractorError(u'no conn, hlsvp or url_encoded_fmt_stream_map information found in video info')
results = []
for format_param, video_real_url in video_url_list:
'description': video_description,
'player_url': player_url,
'subtitles': video_subtitles,
- 'duration': video_duration
+ 'duration': video_duration,
+ 'age_limit': 18 if age_gate else 0,
+ 'annotations': video_annotations
})
return results
mobj = re.match(self._VALID_URL, url, re.VERBOSE)
if mobj is None:
raise ExtractorError(u'Invalid URL: %s' % url)
+ playlist_id = mobj.group(1) or mobj.group(2)
+
+ # Check if it's a video-specific URL
+ query_dict = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
+ if 'v' in query_dict:
+ video_id = query_dict['v'][0]
+ if self._downloader.params.get('noplaylist'):
+ self.to_screen(u'Downloading just video %s because of --no-playlist' % video_id)
+ return self.url_result('https://www.youtube.com/watch?v=' + video_id, 'Youtube')
+ else:
+ self.to_screen(u'Downloading playlist PL%s - add --no-playlist to just download video %s' % (playlist_id, video_id))
# Download playlist videos from API
- playlist_id = mobj.group(1) or mobj.group(2)
videos = []
for page_num in itertools.count(1):
class YoutubeUserIE(InfoExtractor):
IE_DESC = u'YouTube.com user videos (URL or "ytuser" keyword)'
- _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?)|ytuser:)(?!feed/)([A-Za-z0-9_-]+)'
+ _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?(?!(?:attribution_link|watch)(?:$|[^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'
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