X-Git-Url: https://git.rapsys.eu/youtubedl/blobdiff_plain/ced7488f6d3a519b2c1b1cbd31048743fb8285bd..6d2b0da328fef8e5ddefe29086a6692d7ab6e68b:/youtube_dl/extractor/common.py
diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py
index 9427ff4..4252d68 100644
--- a/youtube_dl/extractor/common.py
+++ b/youtube_dl/extractor/common.py
@@ -6,6 +6,7 @@ import hashlib
import json
import netrc
import os
+import random
import re
import socket
import sys
@@ -21,6 +22,7 @@ from ..compat import (
compat_os_name,
compat_str,
compat_urllib_error,
+ compat_urllib_parse_unquote,
compat_urllib_parse_urlencode,
compat_urllib_request,
compat_urlparse,
@@ -29,6 +31,7 @@ from ..downloader.f4m import remove_encrypted_media
from ..utils import (
NO_DEFAULT,
age_restricted,
+ base_url,
bug_reports_message,
clean_html,
compiled_regex_type,
@@ -37,7 +40,10 @@ from ..utils import (
ExtractorError,
fix_xml_ampersands,
float_or_none,
+ GeoRestrictedError,
+ GeoUtils,
int_or_none,
+ js_to_json,
parse_iso8601,
RegexNotFoundError,
sanitize_filename,
@@ -57,6 +63,7 @@ from ..utils import (
parse_m3u8_attributes,
extract_attributes,
parse_codecs,
+ urljoin,
)
@@ -87,6 +94,9 @@ class InfoExtractor(object):
Potential fields:
* url Mandatory. The URL of the video file
+ * manifest_url
+ The URL of the manifest file in case of
+ fragmented media (DASH, hls, hds)
* ext Will be calculated from URL if missing
* format A human-readable description of the format
("mp4 container with h264/opus").
@@ -115,6 +125,21 @@ class InfoExtractor(object):
download, lower-case.
"http", "https", "rtsp", "rtmp", "rtmpe",
"m3u8", "m3u8_native" or "http_dash_segments".
+ * fragment_base_url
+ Base URL for fragments. Each fragment's path
+ value (if present) will be relative to
+ this URL.
+ * fragments A list of fragments of a fragmented media.
+ Each fragment entry must contain either an url
+ or a path. If an url is present it should be
+ considered by a client. Otherwise both path and
+ fragment_base_url must be present. Here is
+ the list of all potential fields:
+ * "url" - fragment's URL
+ * "path" - fragment's path relative to
+ fragment_base_url
+ * "duration" (optional, int or float)
+ * "filesize" (optional, int)
* preference Order number of this format. If this field is
present and not None, the formats get sorted
by this field, regardless of all other values.
@@ -178,9 +203,10 @@ class InfoExtractor(object):
uploader_url: Full URL to a personal webpage of the video uploader.
location: Physical location where the video was filmed.
subtitles: The available subtitles as a dictionary in the format
- {language: subformats}. "subformats" is a list sorted from
- lower to higher preference, each element is a dictionary
- with the "ext" entry and one of:
+ {tag: subformats}. "tag" is usually a language code, and
+ "subformats" is a list sorted from lower to higher
+ preference, each element is a dictionary with the "ext"
+ entry and one of:
* "data": The subtitles file contents
* "url": A URL pointing to the subtitles file
"ext" will be calculated from URL if missing
@@ -226,7 +252,7 @@ class InfoExtractor(object):
chapter_id: Id of the chapter the video belongs to, as a unicode string.
The following fields should only be used when the video is an episode of some
- series or programme:
+ series, programme or podcast:
series: Title of the series or programme the video episode belongs to.
season: Title of the season the video episode belongs to.
@@ -297,17 +323,34 @@ class InfoExtractor(object):
_real_extract() methods and define a _VALID_URL regexp.
Probably, they should also be added to the list of extractors.
+ _GEO_BYPASS attribute may be set to False in order to disable
+ geo restriction bypass mechanisms for a particular extractor.
+ Though it won't disable explicit geo restriction bypass based on
+ country code provided with geo_bypass_country. (experimental)
+
+ _GEO_COUNTRIES attribute may contain a list of presumably geo unrestricted
+ countries for this extractor. One of these countries will be used by
+ geo restriction bypass mechanism right away in order to bypass
+ geo restriction, of course, if the mechanism is not disabled. (experimental)
+
+ NB: both these geo attributes are experimental and may change in future
+ or be completely removed.
+
Finally, the _WORKING attribute should be set to False for broken IEs
in order to warn the users and skip the tests.
"""
_ready = False
_downloader = None
+ _x_forwarded_for_ip = None
+ _GEO_BYPASS = True
+ _GEO_COUNTRIES = None
_WORKING = True
def __init__(self, downloader=None):
"""Constructor. Receives an optional downloader."""
self._ready = False
+ self._x_forwarded_for_ip = None
self.set_downloader(downloader)
@classmethod
@@ -336,15 +379,59 @@ class InfoExtractor(object):
def initialize(self):
"""Initializes an instance (authentication, etc)."""
+ self._initialize_geo_bypass(self._GEO_COUNTRIES)
if not self._ready:
self._real_initialize()
self._ready = True
+ def _initialize_geo_bypass(self, countries):
+ """
+ Initialize geo restriction bypass mechanism.
+
+ This method is used to initialize geo bypass mechanism based on faking
+ X-Forwarded-For HTTP header. A random country from provided country list
+ is selected and a random IP belonging to this country is generated. This
+ IP will be passed as X-Forwarded-For HTTP header in all subsequent
+ HTTP requests.
+
+ This method will be used for initial geo bypass mechanism initialization
+ during the instance initialization with _GEO_COUNTRIES.
+
+ You may also manually call it from extractor's code if geo countries
+ information is not available beforehand (e.g. obtained during
+ extraction) or due to some another reason.
+ """
+ if not self._x_forwarded_for_ip:
+ country_code = self._downloader.params.get('geo_bypass_country', None)
+ # If there is no explicit country for geo bypass specified and
+ # the extractor is known to be geo restricted let's fake IP
+ # as X-Forwarded-For right away.
+ if (not country_code and
+ self._GEO_BYPASS and
+ self._downloader.params.get('geo_bypass', True) and
+ countries):
+ country_code = random.choice(countries)
+ if country_code:
+ self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
+ if self._downloader.params.get('verbose', False):
+ self._downloader.to_stdout(
+ '[debug] Using fake IP %s (%s) as X-Forwarded-For.'
+ % (self._x_forwarded_for_ip, country_code.upper()))
+
def extract(self, url):
"""Extracts URL information and returns it in list of dicts."""
try:
- self.initialize()
- return self._real_extract(url)
+ for _ in range(2):
+ try:
+ self.initialize()
+ ie_result = self._real_extract(url)
+ if self._x_forwarded_for_ip:
+ ie_result['__x_forwarded_for_ip'] = self._x_forwarded_for_ip
+ return ie_result
+ except GeoRestrictedError as e:
+ if self.__maybe_fake_ip_and_retry(e.countries):
+ continue
+ raise
except ExtractorError:
raise
except compat_http_client.IncompleteRead as e:
@@ -352,6 +439,21 @@ class InfoExtractor(object):
except (KeyError, StopIteration) as e:
raise ExtractorError('An extractor error has occurred.', cause=e)
+ def __maybe_fake_ip_and_retry(self, countries):
+ if (not self._downloader.params.get('geo_bypass_country', None) and
+ self._GEO_BYPASS and
+ self._downloader.params.get('geo_bypass', True) and
+ not self._x_forwarded_for_ip and
+ countries):
+ country_code = random.choice(countries)
+ self._x_forwarded_for_ip = GeoUtils.random_ipv4(country_code)
+ if self._x_forwarded_for_ip:
+ self.report_warning(
+ 'Video is geo restricted. Retrying extraction with fake IP %s (%s) as X-Forwarded-For.'
+ % (self._x_forwarded_for_ip, country_code.upper()))
+ return True
+ return False
+
def set_downloader(self, downloader):
"""Sets the downloader for this IE."""
self._downloader = downloader
@@ -411,6 +513,15 @@ class InfoExtractor(object):
if isinstance(url_or_request, (compat_str, str)):
url_or_request = url_or_request.partition('#')[0]
+ # Some sites check X-Forwarded-For HTTP header in order to figure out
+ # the origin of the client behind proxy. This allows bypassing geo
+ # restriction by faking this header's value to IP that belongs to some
+ # geo unrestricted country. We will do so once we encounter any
+ # geo restriction error.
+ if self._x_forwarded_for_ip:
+ if 'X-Forwarded-For' not in headers:
+ headers['X-Forwarded-For'] = self._x_forwarded_for_ip
+
urlh = self._request_webpage(url_or_request, video_id, note, errnote, fatal, data=data, headers=headers, query=query)
if urlh is False:
assert not fatal
@@ -586,10 +697,8 @@ class InfoExtractor(object):
expected=True)
@staticmethod
- def raise_geo_restricted(msg='This video is not available from your location due to geo restriction'):
- raise ExtractorError(
- '%s. You might want to use --proxy to workaround.' % msg,
- expected=True)
+ def raise_geo_restricted(msg='This video is not available from your location due to geo restriction', countries=None):
+ raise GeoRestrictedError(msg, countries=countries)
# Methods for following #608
@staticmethod
@@ -674,33 +783,36 @@ class InfoExtractor(object):
username = info[0]
password = info[2]
else:
- raise netrc.NetrcParseError('No authenticators for %s' % netrc_machine)
+ raise netrc.NetrcParseError(
+ 'No authenticators for %s' % netrc_machine)
except (IOError, netrc.NetrcParseError) as err:
- self._downloader.report_warning('parsing .netrc: %s' % error_to_compat_str(err))
+ self._downloader.report_warning(
+ 'parsing .netrc: %s' % error_to_compat_str(err))
- return (username, password)
+ return username, password
- def _get_login_info(self):
+ def _get_login_info(self, username_option='username', password_option='password', netrc_machine=None):
"""
Get the login info as (username, password)
- It will look in the netrc file using the _NETRC_MACHINE value
+ First look for the manually specified credentials using username_option
+ and password_option as keys in params dictionary. If no such credentials
+ available look in the netrc file using the netrc_machine or _NETRC_MACHINE
+ value.
If there's no info available, return (None, None)
"""
if self._downloader is None:
return (None, None)
- username = None
- password = None
downloader_params = self._downloader.params
# Attempt to use provided username and password or .netrc data
- if downloader_params.get('username') is not None:
- username = downloader_params['username']
- password = downloader_params['password']
+ if downloader_params.get(username_option) is not None:
+ username = downloader_params[username_option]
+ password = downloader_params[password_option]
else:
- username, password = self._get_netrc_login_info()
+ username, password = self._get_netrc_login_info(netrc_machine)
- return (username, password)
+ return username, password
def _get_tfa_info(self, note='two-factor verification code'):
"""
@@ -873,7 +985,7 @@ class InfoExtractor(object):
'url': e.get('contentUrl'),
'title': unescapeHTML(e.get('name')),
'description': unescapeHTML(e.get('description')),
- 'thumbnail': e.get('thumbnailUrl'),
+ 'thumbnail': e.get('thumbnailUrl') or e.get('thumbnailURL'),
'duration': parse_duration(e.get('duration')),
'timestamp': unified_timestamp(e.get('uploadDate')),
'filesize': float_or_none(e.get('contentSize')),
@@ -888,16 +1000,16 @@ class InfoExtractor(object):
def _hidden_inputs(html):
html = re.sub(r'', '', html)
hidden_inputs = {}
- for input in re.findall(r'(?i)]+)>', html):
- if not re.search(r'type=(["\'])(?:hidden|submit)\1', input):
- continue
- name = re.search(r'(?:name|id)=(["\'])(?P.+?)\1', input)
- if not name:
+ for input in re.findall(r'(?i)(]+>)', html):
+ attrs = extract_attributes(input)
+ if not input:
continue
- value = re.search(r'value=(["\'])(?P.*?)\1', input)
- if not value:
+ if attrs.get('type') not in ('hidden', 'submit'):
continue
- hidden_inputs[name.group('value')] = value.group('value')
+ name = attrs.get('name') or attrs.get('id')
+ value = attrs.get('value')
+ if name and value is not None:
+ hidden_inputs[name] = value
return hidden_inputs
def _form_hidden_inputs(self, form_id, html):
@@ -1000,13 +1112,13 @@ class InfoExtractor(object):
unique_formats.append(f)
formats[:] = unique_formats
- def _is_valid_url(self, url, video_id, item='video'):
+ def _is_valid_url(self, url, video_id, item='video', headers={}):
url = self._proto_relative_url(url, scheme='http:')
# For now assume non HTTP(S) URLs always valid
if not (url.startswith('http://') or url.startswith('https://')):
return True
try:
- self._request_webpage(url, video_id, 'Checking %s URL' % item)
+ self._request_webpage(url, video_id, 'Checking %s URL' % item, headers=headers)
return True
except ExtractorError as e:
if isinstance(e.cause, compat_urllib_error.URLError):
@@ -1088,6 +1200,13 @@ class InfoExtractor(object):
manifest, ['{http://ns.adobe.com/f4m/1.0}bootstrapInfo', '{http://ns.adobe.com/f4m/2.0}bootstrapInfo'],
'bootstrap info', default=None)
+ vcodec = None
+ mime_type = xpath_text(
+ manifest, ['{http://ns.adobe.com/f4m/1.0}mimeType', '{http://ns.adobe.com/f4m/2.0}mimeType'],
+ 'base URL', default=None)
+ if mime_type and mime_type.startswith('audio/'):
+ vcodec = 'none'
+
for i, media_el in enumerate(media_nodes):
tbr = int_or_none(media_el.attrib.get('bitrate'))
width = int_or_none(media_el.attrib.get('width'))
@@ -1128,6 +1247,7 @@ class InfoExtractor(object):
'width': f.get('width') or width,
'height': f.get('height') or height,
'format_id': f.get('format_id') if not tbr else format_id,
+ 'vcodec': vcodec,
})
formats.extend(f4m_formats)
continue
@@ -1139,10 +1259,12 @@ class InfoExtractor(object):
formats.append({
'format_id': format_id,
'url': manifest_url,
+ 'manifest_url': manifest_url,
'ext': 'flv' if bootstrap_info is not None else None,
'tbr': tbr,
'width': width,
'height': height,
+ 'vcodec': vcodec,
'preference': preference,
})
return formats
@@ -1163,13 +1285,6 @@ class InfoExtractor(object):
m3u8_id=None, note=None, errnote=None,
fatal=True, live=False):
- formats = [self._m3u8_meta_format(m3u8_url, ext, preference, m3u8_id)]
-
- format_url = lambda u: (
- u
- if re.match(r'^https?://', u)
- else compat_urlparse.urljoin(m3u8_url, u))
-
res = self._download_webpage_handle(
m3u8_url, video_id,
note=note or 'Downloading m3u8 information',
@@ -1180,6 +1295,16 @@ class InfoExtractor(object):
m3u8_doc, urlh = res
m3u8_url = urlh.geturl()
+ if '#EXT-X-FAXS-CM:' in m3u8_doc: # Adobe Flash Access
+ return []
+
+ formats = [self._m3u8_meta_format(m3u8_url, ext, preference, m3u8_id)]
+
+ format_url = lambda u: (
+ u
+ if re.match(r'^https?://', u)
+ else compat_urlparse.urljoin(m3u8_url, u))
+
# We should try extracting formats only from master playlists [1], i.e.
# playlists that describe available qualities. On the other hand media
# playlists [2] should be returned as is since they contain just the media
@@ -1201,35 +1326,62 @@ class InfoExtractor(object):
'protocol': entry_protocol,
'preference': preference,
}]
- last_info = None
- last_media = None
+ audio_in_video_stream = {}
+ last_info = {}
+ last_media = {}
for line in m3u8_doc.splitlines():
if line.startswith('#EXT-X-STREAM-INF:'):
last_info = parse_m3u8_attributes(line)
elif line.startswith('#EXT-X-MEDIA:'):
- last_media = parse_m3u8_attributes(line)
+ media = parse_m3u8_attributes(line)
+ media_type = media.get('TYPE')
+ if media_type in ('VIDEO', 'AUDIO'):
+ group_id = media.get('GROUP-ID')
+ media_url = media.get('URI')
+ if media_url:
+ format_id = []
+ for v in (group_id, media.get('NAME')):
+ if v:
+ format_id.append(v)
+ f = {
+ 'format_id': '-'.join(format_id),
+ 'url': format_url(media_url),
+ 'language': media.get('LANGUAGE'),
+ 'ext': ext,
+ 'protocol': entry_protocol,
+ 'preference': preference,
+ }
+ if media_type == 'AUDIO':
+ f['vcodec'] = 'none'
+ if group_id and not audio_in_video_stream.get(group_id):
+ audio_in_video_stream[group_id] = False
+ formats.append(f)
+ else:
+ # When there is no URI in EXT-X-MEDIA let this tag's
+ # data be used by regular URI lines below
+ last_media = media
+ if media_type == 'AUDIO' and group_id:
+ audio_in_video_stream[group_id] = True
elif line.startswith('#') or not line.strip():
continue
else:
- if last_info is None:
- formats.append({'url': format_url(line)})
- continue
- tbr = int_or_none(last_info.get('BANDWIDTH'), scale=1000)
+ tbr = int_or_none(last_info.get('AVERAGE-BANDWIDTH') or last_info.get('BANDWIDTH'), scale=1000)
format_id = []
if m3u8_id:
format_id.append(m3u8_id)
- last_media_name = last_media.get('NAME') if last_media and last_media.get('TYPE') not in ('SUBTITLES', 'CLOSED-CAPTIONS') else None
# Despite specification does not mention NAME attribute for
# EXT-X-STREAM-INF it still sometimes may be present
- stream_name = last_info.get('NAME') or last_media_name
+ stream_name = last_info.get('NAME') or last_media.get('NAME')
# Bandwidth of live streams may differ over time thus making
# format_id unpredictable. So it's better to keep provided
# format_id intact.
if not live:
format_id.append(stream_name if stream_name else '%d' % (tbr if tbr else len(formats)))
+ manifest_url = format_url(line.strip())
f = {
'format_id': '-'.join(format_id),
- 'url': format_url(line.strip()),
+ 'url': manifest_url,
+ 'manifest_url': manifest_url,
'tbr': tbr,
'ext': ext,
'fps': float_or_none(last_info.get('FRAME-RATE')),
@@ -1238,9 +1390,10 @@ class InfoExtractor(object):
}
resolution = last_info.get('RESOLUTION')
if resolution:
- width_str, height_str = resolution.split('x')
- f['width'] = int(width_str)
- f['height'] = int(height_str)
+ mobj = re.search(r'(?P\d+)[xX](?P\d+)', resolution)
+ if mobj:
+ f['width'] = int(mobj.group('width'))
+ f['height'] = int(mobj.group('height'))
# Unified Streaming Platform
mobj = re.search(
r'audio.*?(?:%3D|=)(\d+)(?:-video.*?(?:%3D|=)(\d+))?', f['url'])
@@ -1252,11 +1405,12 @@ class InfoExtractor(object):
'abr': abr,
})
f.update(parse_codecs(last_info.get('CODECS')))
- if last_media is not None:
- f['m3u8_media'] = last_media
- last_media = None
+ if audio_in_video_stream.get(last_info.get('AUDIO')) is False and f['vcodec'] != 'none':
+ # TODO: update acodec for audio only formats with the same GROUP-ID
+ f['acodec'] = 'none'
formats.append(f)
last_info = {}
+ last_media = {}
return formats
@staticmethod
@@ -1500,12 +1654,13 @@ class InfoExtractor(object):
if res is False:
return []
mpd, urlh = res
- mpd_base_url = re.match(r'https?://.+/', urlh.geturl()).group()
+ mpd_base_url = base_url(urlh.geturl())
return self._parse_mpd_formats(
- compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url, formats_dict=formats_dict)
+ compat_etree_fromstring(mpd.encode('utf-8')), mpd_id, mpd_base_url,
+ formats_dict=formats_dict, mpd_url=mpd_url)
- def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}):
+ def _parse_mpd_formats(self, mpd_doc, mpd_id=None, mpd_base_url='', formats_dict={}, mpd_url=None):
"""
Parse formats from MPD manifest.
References:
@@ -1526,52 +1681,60 @@ class InfoExtractor(object):
def extract_multisegment_info(element, ms_parent_info):
ms_info = ms_parent_info.copy()
+
+ # As per [1, 5.3.9.2.2] SegmentList and SegmentTemplate share some
+ # common attributes and elements. We will only extract relevant
+ # for us.
+ def extract_common(source):
+ segment_timeline = source.find(_add_ns('SegmentTimeline'))
+ if segment_timeline is not None:
+ s_e = segment_timeline.findall(_add_ns('S'))
+ if s_e:
+ ms_info['total_number'] = 0
+ ms_info['s'] = []
+ for s in s_e:
+ r = int(s.get('r', 0))
+ ms_info['total_number'] += 1 + r
+ ms_info['s'].append({
+ 't': int(s.get('t', 0)),
+ # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
+ 'd': int(s.attrib['d']),
+ 'r': r,
+ })
+ start_number = source.get('startNumber')
+ if start_number:
+ ms_info['start_number'] = int(start_number)
+ timescale = source.get('timescale')
+ if timescale:
+ ms_info['timescale'] = int(timescale)
+ segment_duration = source.get('duration')
+ if segment_duration:
+ ms_info['segment_duration'] = int(segment_duration)
+
+ def extract_Initialization(source):
+ initialization = source.find(_add_ns('Initialization'))
+ if initialization is not None:
+ ms_info['initialization_url'] = initialization.attrib['sourceURL']
+
segment_list = element.find(_add_ns('SegmentList'))
if segment_list is not None:
+ extract_common(segment_list)
+ extract_Initialization(segment_list)
segment_urls_e = segment_list.findall(_add_ns('SegmentURL'))
if segment_urls_e:
ms_info['segment_urls'] = [segment.attrib['media'] for segment in segment_urls_e]
- initialization = segment_list.find(_add_ns('Initialization'))
- if initialization is not None:
- ms_info['initialization_url'] = initialization.attrib['sourceURL']
else:
segment_template = element.find(_add_ns('SegmentTemplate'))
if segment_template is not None:
- start_number = segment_template.get('startNumber')
- if start_number:
- ms_info['start_number'] = int(start_number)
- segment_timeline = segment_template.find(_add_ns('SegmentTimeline'))
- if segment_timeline is not None:
- s_e = segment_timeline.findall(_add_ns('S'))
- if s_e:
- ms_info['total_number'] = 0
- ms_info['s'] = []
- for s in s_e:
- r = int(s.get('r', 0))
- ms_info['total_number'] += 1 + r
- ms_info['s'].append({
- 't': int(s.get('t', 0)),
- # @d is mandatory (see [1, 5.3.9.6.2, Table 17, page 60])
- 'd': int(s.attrib['d']),
- 'r': r,
- })
- else:
- timescale = segment_template.get('timescale')
- if timescale:
- ms_info['timescale'] = int(timescale)
- segment_duration = segment_template.get('duration')
- if segment_duration:
- ms_info['segment_duration'] = int(segment_duration)
- media_template = segment_template.get('media')
- if media_template:
- ms_info['media_template'] = media_template
+ extract_common(segment_template)
+ media = segment_template.get('media')
+ if media:
+ ms_info['media'] = media
initialization = segment_template.get('initialization')
if initialization:
- ms_info['initialization_url'] = initialization
+ ms_info['initialization'] = initialization
else:
- initialization = segment_template.find(_add_ns('Initialization'))
- if initialization is not None:
- ms_info['initialization_url'] = initialization.attrib['sourceURL']
+ extract_Initialization(segment_template)
return ms_info
mpd_duration = parse_duration(mpd_doc.get('mediaPresentationDuration'))
@@ -1613,74 +1776,127 @@ class InfoExtractor(object):
lang = representation_attrib.get('lang')
url_el = representation.find(_add_ns('BaseURL'))
filesize = int_or_none(url_el.attrib.get('{http://youtube.com/yt/2012/10/10}contentLength') if url_el is not None else None)
+ bandwidth = int_or_none(representation_attrib.get('bandwidth'))
f = {
'format_id': '%s-%s' % (mpd_id, representation_id) if mpd_id else representation_id,
'url': base_url,
+ 'manifest_url': mpd_url,
'ext': mimetype2ext(mime_type),
'width': int_or_none(representation_attrib.get('width')),
'height': int_or_none(representation_attrib.get('height')),
- 'tbr': int_or_none(representation_attrib.get('bandwidth'), 1000),
+ 'tbr': int_or_none(bandwidth, 1000),
'asr': int_or_none(representation_attrib.get('audioSamplingRate')),
'fps': int_or_none(representation_attrib.get('frameRate')),
- 'vcodec': 'none' if content_type == 'audio' else representation_attrib.get('codecs'),
- 'acodec': 'none' if content_type == 'video' else representation_attrib.get('codecs'),
'language': lang if lang not in ('mul', 'und', 'zxx', 'mis') else None,
'format_note': 'DASH %s' % content_type,
'filesize': filesize,
}
+ f.update(parse_codecs(representation_attrib.get('codecs')))
representation_ms_info = extract_multisegment_info(representation, adaption_set_ms_info)
- if 'segment_urls' not in representation_ms_info and 'media_template' in representation_ms_info:
- if 'total_number' not in representation_ms_info and 'segment_duration':
- segment_duration = float(representation_ms_info['segment_duration']) / float(representation_ms_info['timescale'])
- representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
- media_template = representation_ms_info['media_template']
- media_template = media_template.replace('$RepresentationID$', representation_id)
- media_template = re.sub(r'\$(Number|Bandwidth|Time)\$', r'%(\1)d', media_template)
- media_template = re.sub(r'\$(Number|Bandwidth|Time)%([^$]+)\$', r'%(\1)\2', media_template)
- media_template.replace('$$', '$')
+
+ def prepare_template(template_name, identifiers):
+ t = representation_ms_info[template_name]
+ t = t.replace('$RepresentationID$', representation_id)
+ t = re.sub(r'\$(%s)\$' % '|'.join(identifiers), r'%(\1)d', t)
+ t = re.sub(r'\$(%s)%%([^$]+)\$' % '|'.join(identifiers), r'%(\1)\2', t)
+ t.replace('$$', '$')
+ return t
+
+ # @initialization is a regular template like @media one
+ # so it should be handled just the same way (see
+ # https://github.com/rg3/youtube-dl/issues/11605)
+ if 'initialization' in representation_ms_info:
+ initialization_template = prepare_template(
+ 'initialization',
+ # As per [1, 5.3.9.4.2, Table 15, page 54] $Number$ and
+ # $Time$ shall not be included for @initialization thus
+ # only $Bandwidth$ remains
+ ('Bandwidth', ))
+ representation_ms_info['initialization_url'] = initialization_template % {
+ 'Bandwidth': bandwidth,
+ }
+
+ if 'segment_urls' not in representation_ms_info and 'media' in representation_ms_info:
+
+ media_template = prepare_template('media', ('Number', 'Bandwidth', 'Time'))
# As per [1, 5.3.9.4.4, Table 16, page 55] $Number$ and $Time$
# can't be used at the same time
- if '%(Number' in media_template:
- representation_ms_info['segment_urls'] = [
- media_template % {
+ if '%(Number' in media_template and 's' not in representation_ms_info:
+ segment_duration = None
+ if 'total_number' not in representation_ms_info and 'segment_duration':
+ segment_duration = float_or_none(representation_ms_info['segment_duration'], representation_ms_info['timescale'])
+ representation_ms_info['total_number'] = int(math.ceil(float(period_duration) / segment_duration))
+ representation_ms_info['fragments'] = [{
+ 'url': media_template % {
'Number': segment_number,
- 'Bandwidth': representation_attrib.get('bandwidth'),
- }
- for segment_number in range(
- representation_ms_info['start_number'],
- representation_ms_info['total_number'] + representation_ms_info['start_number'])]
+ 'Bandwidth': bandwidth,
+ },
+ 'duration': segment_duration,
+ } for segment_number in range(
+ representation_ms_info['start_number'],
+ representation_ms_info['total_number'] + representation_ms_info['start_number'])]
else:
- representation_ms_info['segment_urls'] = []
+ # $Number*$ or $Time$ in media template with S list available
+ # Example $Number*$: http://www.svtplay.se/klipp/9023742/stopptid-om-bjorn-borg
+ # Example $Time$: https://play.arkena.com/embed/avp/v2/player/media/b41dda37-d8e7-4d3f-b1b5-9a9db578bdfe/1/129411
+ representation_ms_info['fragments'] = []
segment_time = 0
+ segment_d = None
+ segment_number = representation_ms_info['start_number']
def add_segment_url():
- representation_ms_info['segment_urls'].append(
- media_template % {
- 'Time': segment_time,
- 'Bandwidth': representation_attrib.get('bandwidth'),
- }
- )
+ segment_url = media_template % {
+ 'Time': segment_time,
+ 'Bandwidth': bandwidth,
+ 'Number': segment_number,
+ }
+ representation_ms_info['fragments'].append({
+ 'url': segment_url,
+ 'duration': float_or_none(segment_d, representation_ms_info['timescale']),
+ })
for num, s in enumerate(representation_ms_info['s']):
segment_time = s.get('t') or segment_time
+ segment_d = s['d']
add_segment_url()
+ segment_number += 1
for r in range(s.get('r', 0)):
- segment_time += s['d']
+ segment_time += segment_d
add_segment_url()
- segment_time += s['d']
- if 'segment_urls' in representation_ms_info:
+ segment_number += 1
+ segment_time += segment_d
+ elif 'segment_urls' in representation_ms_info and 's' in representation_ms_info:
+ # No media template
+ # Example: https://www.youtube.com/watch?v=iXZV5uAYMJI
+ # or any YouTube dashsegments video
+ fragments = []
+ segment_index = 0
+ timescale = representation_ms_info['timescale']
+ for s in representation_ms_info['s']:
+ duration = float_or_none(s['d'], timescale)
+ for r in range(s.get('r', 0) + 1):
+ fragments.append({
+ 'url': representation_ms_info['segment_urls'][segment_index],
+ 'duration': duration,
+ })
+ segment_index += 1
+ representation_ms_info['fragments'] = fragments
+ # NB: MPD manifest may contain direct URLs to unfragmented media.
+ # No fragments key is present in this case.
+ if 'fragments' in representation_ms_info:
f.update({
- 'segment_urls': representation_ms_info['segment_urls'],
+ 'fragments': [],
'protocol': 'http_dash_segments',
})
if 'initialization_url' in representation_ms_info:
- initialization_url = representation_ms_info['initialization_url'].replace('$RepresentationID$', representation_id)
- f.update({
- 'initialization_url': initialization_url,
- })
+ initialization_url = representation_ms_info['initialization_url']
if not f.get('url'):
f['url'] = initialization_url
+ f['fragments'].append({'url': initialization_url})
+ f['fragments'].extend(representation_ms_info['fragments'])
+ for fragment in f['fragments']:
+ fragment['url'] = urljoin(base_url, fragment['url'])
try:
existing_format = next(
fo for fo in formats
@@ -1695,7 +1911,106 @@ class InfoExtractor(object):
self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
return formats
- def _parse_html5_media_entries(self, base_url, webpage):
+ def _extract_ism_formats(self, ism_url, video_id, ism_id=None, note=None, errnote=None, fatal=True):
+ res = self._download_webpage_handle(
+ ism_url, video_id,
+ note=note or 'Downloading ISM manifest',
+ errnote=errnote or 'Failed to download ISM manifest',
+ fatal=fatal)
+ if res is False:
+ return []
+ ism, urlh = res
+
+ return self._parse_ism_formats(
+ compat_etree_fromstring(ism.encode('utf-8')), urlh.geturl(), ism_id)
+
+ def _parse_ism_formats(self, ism_doc, ism_url, ism_id=None):
+ if ism_doc.get('IsLive') == 'TRUE' or ism_doc.find('Protection') is not None:
+ return []
+
+ duration = int(ism_doc.attrib['Duration'])
+ timescale = int_or_none(ism_doc.get('TimeScale')) or 10000000
+
+ formats = []
+ for stream in ism_doc.findall('StreamIndex'):
+ stream_type = stream.get('Type')
+ if stream_type not in ('video', 'audio'):
+ continue
+ url_pattern = stream.attrib['Url']
+ stream_timescale = int_or_none(stream.get('TimeScale')) or timescale
+ stream_name = stream.get('Name')
+ for track in stream.findall('QualityLevel'):
+ fourcc = track.get('FourCC')
+ # TODO: add support for WVC1 and WMAP
+ if fourcc not in ('H264', 'AVC1', 'AACL'):
+ self.report_warning('%s is not a supported codec' % fourcc)
+ continue
+ tbr = int(track.attrib['Bitrate']) // 1000
+ width = int_or_none(track.get('MaxWidth'))
+ height = int_or_none(track.get('MaxHeight'))
+ sampling_rate = int_or_none(track.get('SamplingRate'))
+
+ track_url_pattern = re.sub(r'{[Bb]itrate}', track.attrib['Bitrate'], url_pattern)
+ track_url_pattern = compat_urlparse.urljoin(ism_url, track_url_pattern)
+
+ fragments = []
+ fragment_ctx = {
+ 'time': 0,
+ }
+ stream_fragments = stream.findall('c')
+ for stream_fragment_index, stream_fragment in enumerate(stream_fragments):
+ fragment_ctx['time'] = int_or_none(stream_fragment.get('t')) or fragment_ctx['time']
+ fragment_repeat = int_or_none(stream_fragment.get('r')) or 1
+ fragment_ctx['duration'] = int_or_none(stream_fragment.get('d'))
+ if not fragment_ctx['duration']:
+ try:
+ next_fragment_time = int(stream_fragment[stream_fragment_index + 1].attrib['t'])
+ except IndexError:
+ next_fragment_time = duration
+ fragment_ctx['duration'] = (next_fragment_time - fragment_ctx['time']) / fragment_repeat
+ for _ in range(fragment_repeat):
+ fragments.append({
+ 'url': re.sub(r'{start[ _]time}', compat_str(fragment_ctx['time']), track_url_pattern),
+ 'duration': fragment_ctx['duration'] / stream_timescale,
+ })
+ fragment_ctx['time'] += fragment_ctx['duration']
+
+ format_id = []
+ if ism_id:
+ format_id.append(ism_id)
+ if stream_name:
+ format_id.append(stream_name)
+ format_id.append(compat_str(tbr))
+
+ formats.append({
+ 'format_id': '-'.join(format_id),
+ 'url': ism_url,
+ 'manifest_url': ism_url,
+ 'ext': 'ismv' if stream_type == 'video' else 'isma',
+ 'width': width,
+ 'height': height,
+ 'tbr': tbr,
+ 'asr': sampling_rate,
+ 'vcodec': 'none' if stream_type == 'audio' else fourcc,
+ 'acodec': 'none' if stream_type == 'video' else fourcc,
+ 'protocol': 'ism',
+ 'fragments': fragments,
+ '_download_params': {
+ 'duration': duration,
+ 'timescale': stream_timescale,
+ 'width': width or 0,
+ 'height': height or 0,
+ 'fourcc': fourcc,
+ 'codec_private_data': track.get('CodecPrivateData'),
+ 'sampling_rate': sampling_rate,
+ 'channels': int_or_none(track.get('Channels', 2)),
+ 'bits_per_sample': int_or_none(track.get('BitsPerSample', 16)),
+ 'nal_unit_length_field': int_or_none(track.get('NALUnitLengthField', 4)),
+ },
+ })
+ return formats
+
+ def _parse_html5_media_entries(self, base_url, webpage, video_id, m3u8_id=None, m3u8_entry_protocol='m3u8', mpd_id=None):
def absolute_url(video_url):
return compat_urlparse.urljoin(base_url, video_url)
@@ -1710,8 +2025,37 @@ class InfoExtractor(object):
return f
return {}
+ def _media_formats(src, cur_media_type):
+ full_url = absolute_url(src)
+ ext = determine_ext(full_url)
+ if ext == 'm3u8':
+ is_plain_url = False
+ formats = self._extract_m3u8_formats(
+ full_url, video_id, ext='mp4',
+ entry_protocol=m3u8_entry_protocol, m3u8_id=m3u8_id)
+ elif ext == 'mpd':
+ is_plain_url = False
+ formats = self._extract_mpd_formats(
+ full_url, video_id, mpd_id=mpd_id)
+ else:
+ is_plain_url = True
+ formats = [{
+ 'url': full_url,
+ 'vcodec': 'none' if cur_media_type == 'audio' else None,
+ }]
+ return is_plain_url, formats
+
entries = []
- for media_tag, media_type, media_content in re.findall(r'(?s)(<(?Pvideo|audio)[^>]*>)(.*?)(?P=tag)>', webpage):
+ media_tags = [(media_tag, media_type, '')
+ for media_tag, media_type
+ in re.findall(r'(?s)(<(video|audio)[^>]*/>)', webpage)]
+ media_tags.extend(re.findall(
+ # We only allow video|audio followed by a whitespace or '>'.
+ # Allowing more characters may end up in significant slow down (see
+ # https://github.com/rg3/youtube-dl/issues/11979, example URL:
+ # http://www.porntrex.com/maps/videositemap.xml).
+ r'(?s)(<(?Pvideo|audio)(?:\s+[^>]*)?>)(.*?)(?P=tag)>', webpage))
+ for media_tag, media_type, media_content in media_tags:
media_info = {
'formats': [],
'subtitles': {},
@@ -1719,10 +2063,8 @@ class InfoExtractor(object):
media_attributes = extract_attributes(media_tag)
src = media_attributes.get('src')
if src:
- media_info['formats'].append({
- 'url': absolute_url(src),
- 'vcodec': 'none' if media_type == 'audio' else None,
- })
+ _, formats = _media_formats(src, media_type)
+ media_info['formats'].extend(formats)
media_info['thumbnail'] = media_attributes.get('poster')
if media_content:
for source_tag in re.findall(r'