Mirroring
/
yt-dlp
mirror of https://github.com/yt-dlp/yt-dlp
1
0
Fork 0
Code Issues Releases Wiki Activity

[web.archive:youtube] Improve metadata extraction (#1785)

Browse Source 
Authored-by: coletdjnz
pull/1962/head
coletdjnz 3 years ago committed by GitHub
parent 487c5b3389
commit aa4b054512
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 360 additions and 104 deletions
Show Stats Download Patch File Download Diff File

3
README.md
Unescape Escape View File

@ -1580,6 +1580,9 @@ The following extractors use this feature:
#### vikichannel #### vikichannel
* `video_types`: Types of videos to download - one or more of `episodes`, `movies`, `clips`, `trailers` * `video_types`: Types of videos to download - one or more of `episodes`, `movies`, `clips`, `trailers`
#### youtubewebarchive
* `check_all`: Try to check more at the cost of more requests. One or more of `thumbnails`, `captures`
NOTE: These options may be changed/removed in the future without concern for backward compatibility NOTE: These options may be changed/removed in the future without concern for backward compatibility

433
yt_dlp/extractor/archiveorg.py
Unescape Escape View File

@ -3,33 +3,36 @@ from __future__ import unicode_literals
import re import re
import json import json
from .common import InfoExtractor from .common import InfoExtractor
from .youtube import YoutubeIE from .youtube import YoutubeIE, YoutubeBaseInfoExtractor
from ..compat import ( from ..compat import (
compat_urllib_parse_unquote, compat_urllib_parse_unquote,
compat_urllib_parse_unquote_plus, compat_urllib_parse_unquote_plus,
compat_HTTPError compat_HTTPError
) )
from ..utils import ( from ..utils import (
bug_reports_message,
clean_html, clean_html,
determine_ext,
dict_get, dict_get,
extract_attributes, extract_attributes,
ExtractorError, ExtractorError,
get_element_by_id,
HEADRequest, HEADRequest,
int_or_none, int_or_none,
KNOWN_EXTENSIONS, KNOWN_EXTENSIONS,
merge_dicts, merge_dicts,
mimetype2ext, mimetype2ext,
orderedSet,
parse_duration, parse_duration,
parse_qs, parse_qs,
RegexNotFoundError,
str_to_int, str_to_int,
str_or_none, str_or_none,
traverse_obj,
try_get, try_get,
unified_strdate, unified_strdate,
unified_timestamp, unified_timestamp,
urlhandle_detect_ext,
url_or_none
) )
@ -262,12 +265,12 @@ class YoutubeWebArchiveIE(InfoExtractor):
_VALID_URL = r"""(?x)^ _VALID_URL = r"""(?x)^
(?:https?://)?web\.archive\.org/ (?:https?://)?web\.archive\.org/
(?:web/)? (?:web/)?
(?:[0-9A-Za-z_*]+/)? # /web and the version index is optional (?:(?P<date>[0-9]{14})?[0-9A-Za-z_*]*/)? # /web and the version index is optional
(?:https?(?::|%3[Aa])//)? (?:https?(?::|%3[Aa])//)?
(?: (?:
(?:\w+\.)?youtube\.com/watch(?:\?|%3[fF])(?:[^\#]+(?:&|%26))?v(?:=|%3[dD]) # Youtube URL (?:\w+\.)?youtube\.com(?::(?:80|443))?/watch(?:\.php)?(?:\?|%3[fF])(?:[^\#]+(?:&|%26))?v(?:=|%3[dD]) # Youtube URL
|(wayback-fakeurl\.archive\.org/yt/) # Or the internal fake url |(?:wayback-fakeurl\.archive\.org/yt/) # Or the internal fake url
) )
(?P<id>[0-9A-Za-z_-]{11})(?:%26|\#|&|$) (?P<id>[0-9A-Za-z_-]{11})(?:%26|\#|&|$)
""" """
@ -278,141 +281,391 @@ class YoutubeWebArchiveIE(InfoExtractor):
'info_dict': { 'info_dict': {
'id': 'aYAGB11YrSs', 'id': 'aYAGB11YrSs',
'ext': 'webm', 'ext': 'webm',
'title': 'Team Fortress 2 - Sandviches!' 'title': 'Team Fortress 2 - Sandviches!',
'description': 'md5:4984c0f9a07f349fc5d8e82ab7af4eaf',
'upload_date': '20110926',
'uploader': 'Zeurel',
'channel_id': 'UCukCyHaD-bK3in_pKpfH9Eg',
'duration': 32,
'uploader_id': 'Zeurel',
'uploader_url': 'http://www.youtube.com/user/Zeurel'
} }
}, }, {
{
# Internal link # Internal link
'url': 'https://web.archive.org/web/2oe/http://wayback-fakeurl.archive.org/yt/97t7Xj_iBv0', 'url': 'https://web.archive.org/web/2oe/http://wayback-fakeurl.archive.org/yt/97t7Xj_iBv0',
'info_dict': { 'info_dict': {
'id': '97t7Xj_iBv0', 'id': '97t7Xj_iBv0',
'ext': 'mp4', 'ext': 'mp4',
'title': 'How Flexible Machines Could Save The World' 'title': 'Why Machines That Bend Are Better',
'description': 'md5:00404df2c632d16a674ff8df1ecfbb6c',
'upload_date': '20190312',
'uploader': 'Veritasium',
'channel_id': 'UCHnyfMqiRRG1u-2MsSQLbXA',
'duration': 771,
'uploader_id': '1veritasium',
'uploader_url': 'http://www.youtube.com/user/1veritasium'
} }
}, }, {
{ # Video from 2012, webm format itag 45. Newest capture is deleted video, with an invalid description.
# Video from 2012, webm format itag 45. # Should use the date in the link. Title ends with '- Youtube'. Capture has description in eow-description
'url': 'https://web.archive.org/web/20120712231619/http://www.youtube.com/watch?v=AkhihxRKcrs&gl=US&hl=en', 'url': 'https://web.archive.org/web/20120712231619/http://www.youtube.com/watch?v=AkhihxRKcrs&gl=US&hl=en',
'info_dict': { 'info_dict': {
'id': 'AkhihxRKcrs', 'id': 'AkhihxRKcrs',
'ext': 'webm', 'ext': 'webm',
'title': 'Limited Run: Mondo\'s Modern Classic 1 of 3 (SDCC 2012)' 'title': 'Limited Run: Mondo\'s Modern Classic 1 of 3 (SDCC 2012)',
'upload_date': '20120712',
'duration': 398,
'description': 'md5:ff4de6a7980cb65d951c2f6966a4f2f3',
'uploader_id': 'machinima',
'uploader_url': 'http://www.youtube.com/user/machinima'
} }
}, }, {
{ # FLV video. Video file URL does not provide itag information
# Old flash-only video. Webpage title starts with "YouTube - ".
'url': 'https://web.archive.org/web/20081211103536/http://www.youtube.com/watch?v=jNQXAC9IVRw', 'url': 'https://web.archive.org/web/20081211103536/http://www.youtube.com/watch?v=jNQXAC9IVRw',
'info_dict': { 'info_dict': {
'id': 'jNQXAC9IVRw', 'id': 'jNQXAC9IVRw',
'ext': 'unknown_video', 'ext': 'flv',
'title': 'Me at the zoo' 'title': 'Me at the zoo',
'upload_date': '20050423',
'channel_id': 'UC4QobU6STFB0P71PMvOGN5A',
'duration': 19,
'description': 'md5:10436b12e07ac43ff8df65287a56efb4',
'uploader_id': 'jawed',
'uploader_url': 'http://www.youtube.com/user/jawed'
} }
}, }, {
{
# Flash video with .flv extension (itag 34). Title has prefix "YouTube -"
# Title has some weird unicode characters too.
'url': 'https://web.archive.org/web/20110712231407/http://www.youtube.com/watch?v=lTx3G6h2xyA', 'url': 'https://web.archive.org/web/20110712231407/http://www.youtube.com/watch?v=lTx3G6h2xyA',
'info_dict': { 'info_dict': {
'id': 'lTx3G6h2xyA', 'id': 'lTx3G6h2xyA',
'ext': 'flv', 'ext': 'flv',
'title': '‪Madeon - Pop Culture (live mashup)‬‏' 'title': 'Madeon - Pop Culture (live mashup)',
'upload_date': '20110711',
'uploader': 'Madeon',
'channel_id': 'UCqMDNf3Pn5L7pcNkuSEeO3w',
'duration': 204,
'description': 'md5:f7535343b6eda34a314eff8b85444680',
'uploader_id': 'itsmadeon',
'uploader_url': 'http://www.youtube.com/user/itsmadeon'
} }
}, }, {
{ # Some versions of Youtube have have "YouTube" as page title in html (and later rewritten by js). # First capture is of dead video, second is the oldest from CDX response.
'url': 'https://web.archive.org/web/http://www.youtube.com/watch?v=kH-G_aIBlFw', 'url': 'https://web.archive.org/https://www.youtube.com/watch?v=1JYutPM8O6E',
'info_dict': { 'info_dict': {
'id': 'kH-G_aIBlFw', 'id': '1JYutPM8O6E',
'ext': 'mp4', 'ext': 'mp4',
'title': 'kH-G_aIBlFw' 'title': 'Fake Teen Doctor Strikes AGAIN! - Weekly Weird News',
}, 'upload_date': '20160218',
'expected_warnings': [ 'channel_id': 'UCdIaNUarhzLSXGoItz7BHVA',
'unable to extract title', 'duration': 1236,
] 'description': 'md5:21032bae736421e89c2edf36d1936947',
}, 'uploader_id': 'MachinimaETC',
{ 'uploader_url': 'http://www.youtube.com/user/MachinimaETC'
# First capture is a 302 redirect intermediary page. }
'url': 'https://web.archive.org/web/20050214000000/http://www.youtube.com/watch?v=0altSZ96U4M', }, {
# First capture of dead video, capture date in link links to dead capture.
'url': 'https://web.archive.org/web/20180803221945/https://www.youtube.com/watch?v=6FPhZJGvf4E',
'info_dict': { 'info_dict': {
'id': '0altSZ96U4M', 'id': '6FPhZJGvf4E',
'ext': 'mp4', 'ext': 'mp4',
'title': '0altSZ96U4M' 'title': 'WTF: Video Games Still Launch BROKEN?! - T.U.G.S.',
'upload_date': '20160219',
'channel_id': 'UCdIaNUarhzLSXGoItz7BHVA',
'duration': 798,
'description': 'md5:a1dbf12d9a3bd7cb4c5e33b27d77ffe7',
'uploader_id': 'MachinimaETC',
'uploader_url': 'http://www.youtube.com/user/MachinimaETC'
}, },
'expected_warnings': [ 'expected_warnings': [
'unable to extract title', r'unable to download capture webpage \(it may not be archived\)'
] ]
}, }, { # Very old YouTube page, has - YouTube in title.
{ 'url': 'http://web.archive.org/web/20070302011044/http://youtube.com/watch?v=-06-KB9XTzg',
'info_dict': {
'id': '-06-KB9XTzg',
'ext': 'flv',
'title': 'New Coin Hack!! 100% Safe!!'
}
}, {
'url': 'web.archive.org/https://www.youtube.com/watch?v=dWW7qP423y8',
'info_dict': {
'id': 'dWW7qP423y8',
'ext': 'mp4',
'title': 'It\'s Bootleg AirPods Time.',
'upload_date': '20211021',
'channel_id': 'UC7Jwj9fkrf1adN4fMmTkpug',
'channel_url': 'http://www.youtube.com/channel/UC7Jwj9fkrf1adN4fMmTkpug',
'duration': 810,
'description': 'md5:7b567f898d8237b256f36c1a07d6d7bc',
'uploader': 'DankPods',
'uploader_id': 'UC7Jwj9fkrf1adN4fMmTkpug',
'uploader_url': 'http://www.youtube.com/channel/UC7Jwj9fkrf1adN4fMmTkpug'
}
}, {
# player response contains '};' See: https://github.com/ytdl-org/youtube-dl/issues/27093
'url': 'https://web.archive.org/web/20200827003909if_/http://www.youtube.com/watch?v=6Dh-RL__uN4',
'info_dict': {
'id': '6Dh-RL__uN4',
'ext': 'mp4',
'title': 'bitch lasagna',
'upload_date': '20181005',
'channel_id': 'UC-lHJZR3Gqxm24_Vd_AJ5Yw',
'channel_url': 'http://www.youtube.com/channel/UC-lHJZR3Gqxm24_Vd_AJ5Yw',
'duration': 135,
'description': 'md5:2dbe4051feeff2dab5f41f82bb6d11d0',
'uploader': 'PewDiePie',
'uploader_id': 'PewDiePie',
'uploader_url': 'http://www.youtube.com/user/PewDiePie'
}
}, {
'url': 'https://web.archive.org/web/http://www.youtube.com/watch?v=kH-G_aIBlFw',
'only_matching': True
}, {
'url': 'https://web.archive.org/web/20050214000000_if/http://www.youtube.com/watch?v=0altSZ96U4M',
'only_matching': True
}, {
# Video not archived, only capture is unavailable video page # Video not archived, only capture is unavailable video page
'url': 'https://web.archive.org/web/20210530071008/https://www.youtube.com/watch?v=lHJTf93HL1s&spfreload=10', 'url': 'https://web.archive.org/web/20210530071008/https://www.youtube.com/watch?v=lHJTf93HL1s&spfreload=10',
'only_matching': True, 'only_matching': True
}, }, { # Encoded url
{ # Encoded url
'url': 'https://web.archive.org/web/20120712231619/http%3A//www.youtube.com/watch%3Fgl%3DUS%26v%3DAkhihxRKcrs%26hl%3Den', 'url': 'https://web.archive.org/web/20120712231619/http%3A//www.youtube.com/watch%3Fgl%3DUS%26v%3DAkhihxRKcrs%26hl%3Den',
'only_matching': True, 'only_matching': True
}, }, {
{
'url': 'https://web.archive.org/web/20120712231619/http%3A//www.youtube.com/watch%3Fv%3DAkhihxRKcrs%26gl%3DUS%26hl%3Den', 'url': 'https://web.archive.org/web/20120712231619/http%3A//www.youtube.com/watch%3Fv%3DAkhihxRKcrs%26gl%3DUS%26hl%3Den',
'only_matching': True, 'only_matching': True
}, {
'url': 'https://web.archive.org/web/20060527081937/http://www.youtube.com:80/watch.php?v=ELTFsLT73fA&amp;search=soccer',
'only_matching': True
}, {
'url': 'https://web.archive.org/http://www.youtube.com:80/watch?v=-05VVye-ffg',
'only_matching': True
} }
] ]
_YT_INITIAL_DATA_RE = r'(?:(?:(?:window\s*\[\s*["\']ytInitialData["\']\s*\]|ytInitialData)\s*=\s*({.+?})\s*;)|%s)' % YoutubeBaseInfoExtractor._YT_INITIAL_DATA_RE
_YT_INITIAL_PLAYER_RESPONSE_RE = r'(?:(?:(?:window\s*\[\s*["\']ytInitialPlayerResponse["\']\s*\]|ytInitialPlayerResponse)\s*=[(\s]*({.+?})[)\s]*;)|%s)' % YoutubeBaseInfoExtractor._YT_INITIAL_PLAYER_RESPONSE_RE
_YT_INITIAL_BOUNDARY_RE = r'(?:(?:var\s+meta|</script|\n)|%s)' % YoutubeBaseInfoExtractor._YT_INITIAL_BOUNDARY_RE
def _real_extract(self, url): _YT_DEFAULT_THUMB_SERVERS = ['i.ytimg.com'] # thumbnails most likely archived on these servers
video_id = self._match_id(url) _YT_ALL_THUMB_SERVERS = orderedSet(
title = video_id # if we are not able get a title _YT_DEFAULT_THUMB_SERVERS + ['img.youtube.com', *[f'{c}{n or ""}.ytimg.com' for c in ('i', 's') for n in (*range(0, 5), 9)]])
def _extract_title(webpage): _WAYBACK_BASE_URL = 'https://web.archive.org/web/%sif_/'
page_title = self._html_search_regex( _OLDEST_CAPTURE_DATE = 20050214000000
r'<title>([^<]*)</title>', webpage, 'title', fatal=False) or '' _NEWEST_CAPTURE_DATE = 20500101000000
# YouTube video pages appear to always have either 'YouTube -' as suffix or '- YouTube' as prefix.
try: def _call_cdx_api(self, item_id, url, filters: list = None, collapse: list = None, query: dict = None, note='Downloading CDX API JSON'):
# CDX docs: https://github.com/internetarchive/wayback/blob/master/wayback-cdx-server/README.md
query = {
'url': url,
'output': 'json',
'fl': 'original,mimetype,length,timestamp',
'limit': 500,
'filter': ['statuscode:200'] + (filters or []),
'collapse': collapse or [],
**(query or {})
}
res = self._download_json('https://web.archive.org/cdx/search/cdx', item_id, note, query=query)
if isinstance(res, list) and len(res) >= 2:
# format response to make it easier to use
return list(dict(zip(res[0], v)) for v in res[1:])
elif not isinstance(res, list) or len(res) != 0:
self.report_warning('Error while parsing CDX API response' + bug_reports_message())
def _extract_yt_initial_variable(self, webpage, regex, video_id, name):
return self._parse_json(self._search_regex(
(r'%s\s*%s' % (regex, self._YT_INITIAL_BOUNDARY_RE),
regex), webpage, name, default='{}'), video_id, fatal=False)
def _extract_webpage_title(self, webpage):
page_title = self._html_search_regex( page_title = self._html_search_regex(
r'<title>([^<]*)</title>', webpage, 'title', default='')
# YouTube video pages appear to always have either 'YouTube -' as prefix or '- YouTube' as suffix.
return self._html_search_regex(
r'(?:YouTube\s*-\s*(.*)$)|(?:(.*)\s*-\s*YouTube$)', r'(?:YouTube\s*-\s*(.*)$)|(?:(.*)\s*-\s*YouTube$)',
page_title, 'title', default='') page_title, 'title', default='')
except RegexNotFoundError:
page_title = None
if not page_title: def _extract_metadata(self, video_id, webpage):
self.report_warning('unable to extract title', video_id=video_id)
return
return page_title
# If the video is no longer available, the oldest capture may be one before it was removed. search_meta = ((lambda x: self._html_search_meta(x, webpage, default=None)) if webpage else (lambda x: None))
# Setting the capture date in url to early date seems to redirect to earliest capture. player_response = self._extract_yt_initial_variable(
webpage = self._download_webpage( webpage, self._YT_INITIAL_PLAYER_RESPONSE_RE, video_id, 'initial player response') or {}
'https://web.archive.org/web/20050214000000/http://www.youtube.com/watch?v=%s' % video_id, initial_data = self._extract_yt_initial_variable(
video_id=video_id, fatal=False, errnote='unable to download video webpage (probably not archived).') webpage, self._YT_INITIAL_DATA_RE, video_id, 'initial player response') or {}
if webpage:
title = _extract_title(webpage) or title initial_data_video = traverse_obj(
initial_data, ('contents', 'twoColumnWatchNextResults', 'results', 'results', 'contents', ..., 'videoPrimaryInfoRenderer'),
expected_type=dict, get_all=False, default={})
video_details = traverse_obj(
player_response, 'videoDetails', expected_type=dict, get_all=False, default={})
microformats = traverse_obj(
player_response, ('microformat', 'playerMicroformatRenderer'), expected_type=dict, get_all=False, default={})
video_title = (
video_details.get('title')
or YoutubeBaseInfoExtractor._get_text(microformats, 'title')
or YoutubeBaseInfoExtractor._get_text(initial_data_video, 'title')
or self._extract_webpage_title(webpage)
or search_meta(['og:title', 'twitter:title', 'title']))
channel_id = str_or_none(
video_details.get('channelId')
or microformats.get('externalChannelId')
or search_meta('channelId')
or self._search_regex(
r'data-channel-external-id=(["\'])(?P<id>(?:(?!\1).)+)\1', # @b45a9e6
webpage, 'channel id', default=None, group='id'))
channel_url = f'http://www.youtube.com/channel/{channel_id}' if channel_id else None
duration = int_or_none(
video_details.get('lengthSeconds')
or microformats.get('lengthSeconds')
or parse_duration(search_meta('duration')))
description = (
video_details.get('shortDescription')
or YoutubeBaseInfoExtractor._get_text(microformats, 'description')
or clean_html(get_element_by_id('eow-description', webpage)) # @9e6dd23
or search_meta(['description', 'og:description', 'twitter:description']))
uploader = video_details.get('author')
# Uploader ID and URL
uploader_mobj = re.search(
r'<link itemprop="url" href="(?P<uploader_url>https?://www\.youtube\.com/(?:user|channel)/(?P<uploader_id>[^"]+))">', # @fd05024
webpage)
if uploader_mobj is not None:
uploader_id, uploader_url = uploader_mobj.group('uploader_id'), uploader_mobj.group('uploader_url')
else:
# @a6211d2
uploader_url = url_or_none(microformats.get('ownerProfileUrl'))
uploader_id = self._search_regex(
r'(?:user|channel)/([^/]+)', uploader_url or '', 'uploader id', default=None)
upload_date = unified_strdate(
dict_get(microformats, ('uploadDate', 'publishDate'))
or search_meta(['uploadDate', 'datePublished'])
or self._search_regex(
[r'(?s)id="eow-date.*?>(.*?)</span>',
r'(?:id="watch-uploader-info".*?>.*?|["\']simpleText["\']\s*:\s*["\'])(?:Published|Uploaded|Streamed live|Started) on (.+?)[<"\']'], # @7998520
webpage, 'upload date', default=None))
return {
'title': video_title,
'description': description,
'upload_date': upload_date,
'uploader': uploader,
'channel_id': channel_id,
'channel_url': channel_url,
'duration': duration,
'uploader_url': uploader_url,
'uploader_id': uploader_id,
}
def _extract_thumbnails(self, video_id):
try_all = 'thumbnails' in self._configuration_arg('check_all')
thumbnail_base_urls = ['http://{server}/vi{webp}/{video_id}'.format(
webp='_webp' if ext == 'webp' else '', video_id=video_id, server=server)
for server in (self._YT_ALL_THUMB_SERVERS if try_all else self._YT_DEFAULT_THUMB_SERVERS) for ext in (('jpg', 'webp') if try_all else ('jpg',))]
thumbnails = []
for url in thumbnail_base_urls:
response = self._call_cdx_api(
video_id, url, filters=['mimetype:image/(?:webp|jpeg)'],
collapse=['urlkey'], query={'matchType': 'prefix'})
if not response:
continue
thumbnails.extend(
{
'url': (self._WAYBACK_BASE_URL % (int_or_none(thumbnail_dict.get('timestamp')) or self._OLDEST_CAPTURE_DATE)) + thumbnail_dict.get('original'),
'filesize': int_or_none(thumbnail_dict.get('length')),
'preference': int_or_none(thumbnail_dict.get('length'))
} for thumbnail_dict in response)
if not try_all:
break
# Use link translator mentioned in https://github.com/ytdl-org/youtube-dl/issues/13655 self._remove_duplicate_formats(thumbnails)
internal_fake_url = 'https://web.archive.org/web/2oe_/http://wayback-fakeurl.archive.org/yt/%s' % video_id return thumbnails
def _get_capture_dates(self, video_id, url_date):
capture_dates = []
# Note: CDX API will not find watch pages with extra params in the url.
response = self._call_cdx_api(
video_id, f'https://www.youtube.com/watch?v={video_id}',
filters=['mimetype:text/html'], collapse=['timestamp:6', 'digest'], query={'matchType': 'prefix'}) or []
all_captures = sorted([int_or_none(r['timestamp']) for r in response if int_or_none(r['timestamp']) is not None])
# Prefer the new polymer UI captures as we support extracting more metadata from them
# WBM captures seem to all switch to this layout ~July 2020
modern_captures = list(filter(lambda x: x >= 20200701000000, all_captures))
if modern_captures:
capture_dates.append(modern_captures[0])
capture_dates.append(url_date)
if all_captures:
capture_dates.append(all_captures[0])
if 'captures' in self._configuration_arg('check_all'):
capture_dates.extend(modern_captures + all_captures)
# Fallbacks if any of the above fail
capture_dates.extend([self._OLDEST_CAPTURE_DATE, self._NEWEST_CAPTURE_DATE])
return orderedSet(capture_dates)
def _real_extract(self, url):
url_date, video_id = self._match_valid_url(url).groups()
urlh = None
try: try:
video_file_webpage = self._request_webpage( urlh = self._request_webpage(
HEADRequest(internal_fake_url), video_id, HEADRequest('https://web.archive.org/web/2oe_/http://wayback-fakeurl.archive.org/yt/%s' % video_id),
note='Fetching video file url', expected_status=True) video_id, note='Fetching archived video file url', expected_status=True)
except ExtractorError as e: except ExtractorError as e:
# HTTP Error 404 is expected if the video is not saved. # HTTP Error 404 is expected if the video is not saved.
if isinstance(e.cause, compat_HTTPError) and e.cause.code == 404: if isinstance(e.cause, compat_HTTPError) and e.cause.code == 404:
raise ExtractorError( self.raise_no_formats(
'HTTP Error %s. Most likely the video is not archived or issue with web.archive.org.' % e.cause.code, 'The requested video is not archived, indexed, or there is an issue with web.archive.org',
expected=True) expected=True)
else:
raise raise
video_file_url = compat_urllib_parse_unquote(video_file_webpage.url)
video_file_url_qs = parse_qs(video_file_url)
# Attempt to recover any ext & format info from playback url capture_dates = self._get_capture_dates(video_id, int_or_none(url_date))
format = {'url': video_file_url} self.write_debug('Captures to try: ' + ', '.join(str(i) for i in capture_dates if i is not None))
info = {'id': video_id}
for capture in capture_dates:
if not capture:
continue
webpage = self._download_webpage(
(self._WAYBACK_BASE_URL + 'http://www.youtube.com/watch?v=%s') % (capture, video_id),
video_id=video_id, fatal=False, errnote='unable to download capture webpage (it may not be archived)',
note='Downloading capture webpage')
current_info = self._extract_metadata(video_id, webpage or '')
# Try avoid getting deleted video metadata
if current_info.get('title'):
info = merge_dicts(info, current_info)
if 'captures' not in self._configuration_arg('check_all'):
break
info['thumbnails'] = self._extract_thumbnails(video_id)
if urlh:
url = compat_urllib_parse_unquote(urlh.url)
video_file_url_qs = parse_qs(url)
# Attempt to recover any ext & format info from playback url & response headers
format = {'url': url, 'filesize': int_or_none(urlh.headers.get('x-archive-orig-content-length'))}
itag = try_get(video_file_url_qs, lambda x: x['itag'][0]) itag = try_get(video_file_url_qs, lambda x: x['itag'][0])
if itag and itag in YoutubeIE._formats: # Naughty access but it works if itag and itag in YoutubeIE._formats:
format.update(YoutubeIE._formats[itag]) format.update(YoutubeIE._formats[itag])
format.update({'format_id': itag}) format.update({'format_id': itag})
else: else:
mime = try_get(video_file_url_qs, lambda x: x['mime'][0]) mime = try_get(video_file_url_qs, lambda x: x['mime'][0])
ext = mimetype2ext(mime) or determine_ext(video_file_url) ext = (mimetype2ext(mime)
or urlhandle_detect_ext(urlh)
or mimetype2ext(urlh.headers.get('x-archive-guessed-content-type')))
format.update({'ext': ext}) format.update({'ext': ext})
return { info['formats'] = [format]
'id': video_id, if not info.get('duration'):
'title': title, info['duration'] = str_to_int(try_get(video_file_url_qs, lambda x: x['dur'][0]))
'formats': [format],
'duration': str_to_int(try_get(video_file_url_qs, lambda x: x['dur'][0])) if not info.get('title'):
} info['title'] = video_id
return info

Write Preview
Loading…
Cancel
Save