[tumblr] Fix extractor (#2883)

Authored by: foghawk
This commit is contained in:
foghawk 2022-03-04 21:24:49 -06:00 committed by GitHub
parent 27231526ae
commit bed30106f5
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23

View File

@ -6,6 +6,7 @@
from ..utils import ( from ..utils import (
ExtractorError, ExtractorError,
int_or_none, int_or_none,
traverse_obj,
urlencode_postdata urlencode_postdata
) )
@ -14,31 +15,130 @@ class TumblrIE(InfoExtractor):
_VALID_URL = r'https?://(?P<blog_name>[^/?#&]+)\.tumblr\.com/(?:post|video)/(?P<id>[0-9]+)(?:$|[/?#])' _VALID_URL = r'https?://(?P<blog_name>[^/?#&]+)\.tumblr\.com/(?:post|video)/(?P<id>[0-9]+)(?:$|[/?#])'
_NETRC_MACHINE = 'tumblr' _NETRC_MACHINE = 'tumblr'
_LOGIN_URL = 'https://www.tumblr.com/login' _LOGIN_URL = 'https://www.tumblr.com/login'
_OAUTH_URL = 'https://www.tumblr.com/api/v2/oauth2/token'
_TESTS = [{ _TESTS = [{
'url': 'http://tatianamaslanydaily.tumblr.com/post/54196191430/orphan-black-dvd-extra-behind-the-scenes', 'url': 'http://tatianamaslanydaily.tumblr.com/post/54196191430/orphan-black-dvd-extra-behind-the-scenes',
'md5': '479bb068e5b16462f5176a6828829767', 'md5': '479bb068e5b16462f5176a6828829767',
'info_dict': { 'info_dict': {
'id': '54196191430', 'id': '54196191430',
'ext': 'mp4', 'ext': 'mp4',
'title': 'tatiana maslany news, Orphan Black || DVD extra - behind the scenes ↳...', 'title': 'md5:dfac39636969fe6bf1caa2d50405f069',
'description': 'md5:390ab77358960235b6937ab3b8528956', 'description': 'md5:390ab77358960235b6937ab3b8528956',
'uploader_id': 'tatianamaslanydaily',
'uploader_url': 'https://tatianamaslanydaily.tumblr.com/',
'thumbnail': r're:^https?://.*\.jpg', 'thumbnail': r're:^https?://.*\.jpg',
'duration': 127, 'duration': 127,
'like_count': int,
'repost_count': int,
'age_limit': 0,
'tags': ['Orphan Black', 'Tatiana Maslany', 'Interview', 'Video', 'OB S1 DVD Extras'],
} }
}, { }, {
'note': 'multiple formats',
'url': 'https://maskofthedragon.tumblr.com/post/626907179849564160/mona-talking-in-english', 'url': 'https://maskofthedragon.tumblr.com/post/626907179849564160/mona-talking-in-english',
'md5': 'f43ff8a8861712b6cf0e0c2bd84cfc68', 'md5': 'f43ff8a8861712b6cf0e0c2bd84cfc68',
'info_dict': { 'info_dict': {
'id': '626907179849564160', 'id': '626907179849564160',
'ext': 'mp4', 'ext': 'mp4',
'title': 'Me roast is buggered!, Mona\xa0“talking” in\xa0“english”', 'title': 'Mona\xa0“talking” in\xa0“english”',
'description': 'md5:082a3a621530cb786ad2b7592a6d9e2c', 'description': 'md5:082a3a621530cb786ad2b7592a6d9e2c',
'uploader_id': 'maskofthedragon',
'uploader_url': 'https://maskofthedragon.tumblr.com/',
'thumbnail': r're:^https?://.*\.jpg', 'thumbnail': r're:^https?://.*\.jpg',
'duration': 7, 'duration': 7,
'like_count': int,
'repost_count': int,
'age_limit': 0,
'tags': 'count:19',
}, },
'params': { 'params': {
'format': 'hd', 'format': 'hd',
}, },
}, {
'note': 'non-iframe video (with related posts)',
'url': 'https://shieldfoss.tumblr.com/post/675519763813908480',
'md5': '12bdb75661ef443bffe5a4dac1dbf118',
'info_dict': {
'id': '675519763813908480',
'ext': 'mp4',
'title': 'Shieldfoss',
'uploader_id': 'nerviovago',
'uploader_url': 'https://nerviovago.tumblr.com/',
'thumbnail': r're:^https?://.*\.jpg',
'like_count': int,
'repost_count': int,
'age_limit': 0,
'tags': [],
}
}, {
'note': 'dashboard only (original post)',
'url': 'https://jujanon.tumblr.com/post/159704441298/my-baby-eating',
'md5': '029f7c91ab386701b211e3d494d2d95e',
'info_dict': {
'id': '159704441298',
'ext': 'mp4',
'title': 'md5:ba79365861101f4911452728d2950561',
'description': 'md5:773738196cea76b6996ec71e285bdabc',
'uploader_id': 'jujanon',
'uploader_url': 'https://jujanon.tumblr.com/',
'thumbnail': r're:^https?://.*\.jpg',
'like_count': int,
'repost_count': int,
'age_limit': 0,
'tags': ['crabs', 'my video', 'my pets'],
}
}, {
'note': 'dashboard only (reblog)',
'url': 'https://bartlebyshop.tumblr.com/post/180294460076/duality-of-bird',
'md5': '04334e7cadb1af680d162912559f51a5',
'info_dict': {
'id': '180294460076',
'ext': 'mp4',
'title': 'duality of bird',
'description': 'duality of bird',
'uploader_id': 'todaysbird',
'uploader_url': 'https://todaysbird.tumblr.com/',
'thumbnail': r're:^https?://.*\.jpg',
'like_count': int,
'repost_count': int,
'age_limit': 0,
'tags': [],
}
}, {
'note': 'dashboard only (external)',
'url': 'https://afloweroutofstone.tumblr.com/post/675661759168823296/the-blues-remembers-everything-the-country-forgot',
'info_dict': {
'id': 'q67_fd7b8SU',
'ext': 'mp4',
'title': 'The Blues Remembers Everything the Country Forgot',
'alt_title': 'The Blues Remembers Everything the Country Forgot',
'description': 'md5:1a6b4097e451216835a24c1023707c79',
'release_date': '20201224',
'creator': 'md5:c2239ba15430e87c3b971ba450773272',
'uploader': 'Moor Mother - Topic',
'upload_date': '20201223',
'uploader_id': 'UCxrMtFBRkFvQJ_vVM4il08w',
'uploader_url': 'http://www.youtube.com/channel/UCxrMtFBRkFvQJ_vVM4il08w',
'thumbnail': r're:^https?://i.ytimg.com/.*',
'channel': 'Moor Mother - Topic',
'channel_id': 'UCxrMtFBRkFvQJ_vVM4il08w',
'channel_url': 'https://www.youtube.com/channel/UCxrMtFBRkFvQJ_vVM4il08w',
'channel_follower_count': int,
'duration': 181,
'view_count': int,
'like_count': int,
'age_limit': 0,
'categories': ['Music'],
'tags': 'count:7',
'live_status': 'not_live',
'playable_in_embed': True,
'availability': 'public',
'track': 'The Blues Remembers Everything the Country Forgot',
'artist': 'md5:c2239ba15430e87c3b971ba450773272',
'album': 'Brass',
'release_year': 2020,
},
'add_ie': ['Youtube'],
}, { }, {
'url': 'http://naked-yogi.tumblr.com/post/118312946248/naked-smoking-stretching', 'url': 'http://naked-yogi.tumblr.com/post/118312946248/naked-smoking-stretching',
'md5': 'de07e5211d60d4f3a2c3df757ea9f6ab', 'md5': 'de07e5211d60d4f3a2c3df757ea9f6ab',
@ -54,17 +154,48 @@ class TumblrIE(InfoExtractor):
}, },
# 'add_ie': ['Vidme'], # 'add_ie': ['Vidme'],
'skip': 'dead embedded video host' 'skip': 'dead embedded video host'
}, {
'url': 'https://prozdvoices.tumblr.com/post/673201091169681408/what-recording-voice-acting-sounds-like',
'md5': 'a0063fc8110e6c9afe44065b4ea68177',
'info_dict': {
'id': 'eomhW5MLGWA',
'ext': 'mp4',
'title': 'what recording voice acting sounds like',
'description': 'md5:1da3faa22d0e0b1d8b50216c284ee798',
'uploader': 'ProZD',
'upload_date': '20220112',
'uploader_id': 'ProZD',
'uploader_url': 'http://www.youtube.com/user/ProZD',
'thumbnail': r're:^https?://i.ytimg.com/.*',
'channel': 'ProZD',
'channel_id': 'UC6MFZAOHXlKK1FI7V0XQVeA',
'channel_url': 'https://www.youtube.com/channel/UC6MFZAOHXlKK1FI7V0XQVeA',
'channel_follower_count': int,
'duration': 20,
'view_count': int,
'like_count': int,
'age_limit': 0,
'categories': ['Film & Animation'],
'tags': [],
'live_status': 'not_live',
'playable_in_embed': True,
'availability': 'public',
},
'add_ie': ['Youtube'],
}, { }, {
'url': 'https://dominustempori.tumblr.com/post/673572712813297664/youtubes-all-right-for-some-pretty-cool', 'url': 'https://dominustempori.tumblr.com/post/673572712813297664/youtubes-all-right-for-some-pretty-cool',
'md5': '5e45724c70b748f64f5a1731ac72c84a', 'md5': '203e9eb8077e3f45bfaeb4c86c1467b8',
'info_dict': { 'info_dict': {
'id': '87816359', 'id': '87816359',
'ext': 'mp4', 'ext': 'mov',
'title': 'Harold Ramis', 'title': 'Harold Ramis',
'description': 'md5:be8e68cbf56ce0785c77f0c6c6dfaf2c',
'uploader': 'Resolution Productions Group', 'uploader': 'Resolution Productions Group',
'uploader_id': 'resolutionproductions', 'uploader_id': 'resolutionproductions',
'uploader_url': 'https://vimeo.com/resolutionproductions', 'uploader_url': 'https://vimeo.com/resolutionproductions',
'upload_date': '20140227',
'thumbnail': r're:^https?://i.vimeocdn.com/video/.*', 'thumbnail': r're:^https?://i.vimeocdn.com/video/.*',
'timestamp': 1393523719,
'duration': 291, 'duration': 291,
}, },
'add_ie': ['Vimeo'], 'add_ie': ['Vimeo'],
@ -107,116 +238,163 @@ class TumblrIE(InfoExtractor):
'add_ie': ['Instagram'], 'add_ie': ['Instagram'],
}] }]
_providers = {
'instagram': 'Instagram',
'vimeo': 'Vimeo',
'vine': 'Vine',
'youtube': 'Youtube',
}
_ACCESS_TOKEN = None
def _real_initialize(self): def _real_initialize(self):
self.get_access_token()
self._login() self._login()
def get_access_token(self):
login_page = self._download_webpage(
self._LOGIN_URL, None, 'Downloading login page', fatal=False)
if login_page:
self._ACCESS_TOKEN = self._search_regex(
r'"API_TOKEN":\s*"(\w+)"', login_page, 'API access token', fatal=False)
if not self._ACCESS_TOKEN:
self.report_warning('Failed to get access token; metadata will be missing and some videos may not work')
def _login(self): def _login(self):
username, password = self._get_login_info() username, password = self._get_login_info()
if username is None: if not username:
return return
login_page = self._download_webpage( if not self._ACCESS_TOKEN:
self._LOGIN_URL, None, 'Downloading login page') return
login_form = self._hidden_inputs(login_page) self._download_json(
login_form.update({ self._OAUTH_URL, None, 'Logging in',
'user[email]': username, data=urlencode_postdata({
'user[password]': password 'password': password,
}) 'grant_type': 'password',
'username': username,
response, urlh = self._download_webpage_handle( }), headers={
self._LOGIN_URL, None, 'Logging in',
data=urlencode_postdata(login_form), headers={
'Content-Type': 'application/x-www-form-urlencoded', 'Content-Type': 'application/x-www-form-urlencoded',
'Referer': self._LOGIN_URL, 'Authorization': f'Bearer {self._ACCESS_TOKEN}',
}) },
errnote='Login failed', fatal=False)
# Successful login
if '/dashboard' in urlh.geturl():
return
login_errors = self._parse_json(
self._search_regex(
r'RegistrationForm\.errors\s*=\s*(\[.+?\])\s*;', response,
'login errors', default='[]'),
None, fatal=False)
if login_errors:
raise ExtractorError(
'Unable to login: %s' % login_errors[0], expected=True)
self.report_warning('Login has probably failed')
def _real_extract(self, url): def _real_extract(self, url):
m_url = self._match_valid_url(url) blog, video_id = self._match_valid_url(url).groups()
video_id = m_url.group('id')
blog = m_url.group('blog_name')
url = 'http://%s.tumblr.com/post/%s/' % (blog, video_id) url = f'http://{blog}.tumblr.com/post/{video_id}/'
webpage, urlh = self._download_webpage_handle(url, video_id) webpage, urlh = self._download_webpage_handle(url, video_id)
redirect_url = urlh.geturl() redirect_url = urlh.geturl()
if 'tumblr.com/safe-mode' in redirect_url or redirect_url.startswith('/safe-mode'):
raise ExtractorError(
'This Tumblr may contain sensitive media. '
'Disable safe mode in your account settings '
'at https://www.tumblr.com/settings/account#safe_mode',
expected=True)
api_only = bool(self._search_regex(
r'(tumblr.com|^)/(safe-mode|login_required|blog/view)',
redirect_url, 'redirect', default=None))
if api_only and not self._ACCESS_TOKEN:
raise ExtractorError('Cannot get data for dashboard-only post without access token')
post_json = {}
if self._ACCESS_TOKEN:
post_json = traverse_obj(
self._download_json(
f'https://www.tumblr.com/api/v2/blog/{blog}/posts/{video_id}/permalink',
video_id, headers={'Authorization': f'Bearer {self._ACCESS_TOKEN}'}, fatal=False),
('response', 'timeline', 'elements', 0)) or {}
content_json = traverse_obj(post_json, ('trail', 0, 'content'), ('content')) or []
video_json = next(
(item for item in content_json if item.get('type') == 'video'), {})
media_json = video_json.get('media') or {}
if api_only and not media_json.get('url') and not video_json.get('url'):
raise ExtractorError('Failed to find video data for dashboard-only post')
if not media_json.get('url') and video_json.get('url'):
# external video host
return self.url_result(
video_json['url'],
self._providers.get(video_json.get('provider'), 'Generic'))
video_url = self._og_search_video_url(webpage, default=None)
duration = None
formats = []
# iframes can supply duration and sometimes additional formats, so check for one
iframe_url = self._search_regex( iframe_url = self._search_regex(
r'src=\'(https?://www\.tumblr\.com/video/[^\']+)\'', fr'src=\'(https?://www\.tumblr\.com/video/{blog}/{video_id}/[^\']+)\'',
webpage, 'iframe url', default=None) webpage, 'iframe url', default=None)
if iframe_url is None: if iframe_url:
iframe = self._download_webpage(
iframe_url, video_id, 'Downloading iframe page',
headers={'Referer': redirect_url})
options = self._parse_json(
self._search_regex(
r'data-crt-options=(["\'])(?P<options>.+?)\1', iframe,
'hd video url', default='', group='options'),
video_id, fatal=False)
if options:
duration = int_or_none(options.get('duration'))
hd_url = options.get('hdUrl')
if hd_url:
# there are multiple formats; extract them
# ignore other sources of width/height data as they may be wrong
sources = []
sd_url = self._search_regex(
r'<source[^>]+src=(["\'])(?P<url>.+?)\1', iframe,
'sd video url', default=None, group='url')
if sd_url:
sources.append((sd_url, 'sd'))
sources.append((hd_url, 'hd'))
formats = [{
'url': video_url,
'format_id': format_id,
'height': int_or_none(self._search_regex(
r'_(\d+)\.\w+$', video_url, 'height', default=None)),
'quality': quality,
} for quality, (video_url, format_id) in enumerate(sources)]
if not media_json.get('url') and not video_url and not iframe_url:
# external video host (but we weren't able to figure it out from the api)
iframe_url = self._search_regex( iframe_url = self._search_regex(
r'src=["\'](https?://safe\.txmblr\.com/svc/embed/inline/[^"\']+)["\']', r'src=["\'](https?://safe\.txmblr\.com/svc/embed/inline/[^"\']+)["\']',
webpage, 'embed iframe url', default=None) webpage, 'embed iframe url', default=None)
return self.url_result(iframe_url or redirect_url, 'Generic') return self.url_result(iframe_url or redirect_url, 'Generic')
iframe = self._download_webpage( formats = formats or [{
iframe_url, video_id, 'Downloading iframe page', 'url': media_json.get('url') or video_url,
headers={'Referer': redirect_url}) 'width': int_or_none(
media_json.get('width') or self._og_search_property('video:width', webpage, default=None)),
duration = None 'height': int_or_none(
sources = [] media_json.get('height') or self._og_search_property('video:height', webpage, default=None)),
}]
sd_url = self._search_regex(
r'<source[^>]+src=(["\'])(?P<url>.+?)\1', iframe,
'sd video url', default=None, group='url')
if sd_url:
sources.append((sd_url, 'sd'))
options = self._parse_json(
self._search_regex(
r'data-crt-options=(["\'])(?P<options>.+?)\1', iframe,
'hd video url', default='', group='options'),
video_id, fatal=False)
if options:
duration = int_or_none(options.get('duration'))
hd_url = options.get('hdUrl')
if hd_url:
sources.append((hd_url, 'hd'))
formats = [{
'url': video_url,
'ext': 'mp4',
'format_id': format_id,
'height': int_or_none(self._search_regex(
r'/(\d{3,4})$', video_url, 'height', default=None)),
'quality': quality,
} for quality, (video_url, format_id) in enumerate(sources)]
self._sort_formats(formats) self._sort_formats(formats)
# The only place where you can get a title, it's not complete, # the url we're extracting from might be an original post or it might be a reblog.
# but searching in other places doesn't work for all videos # if it's a reblog, og:description will be the reblogger's comment, not the uploader's.
video_title = self._html_search_regex( # content_json is always the op, so if it exists but has no text, there's no description
r'(?s)<title>(?P<title>.*?)(?: \| Tumblr)?</title>', if content_json:
webpage, 'title') description = '\n\n'.join((
item.get('text') for item in content_json if item.get('type') == 'text')) or None
else:
description = self._og_search_description(webpage, default=None)
uploader_id = traverse_obj(post_json, 'reblogged_root_name', 'blog_name')
return { return {
'id': video_id, 'id': video_id,
'title': video_title, 'title': post_json.get('summary') or (blog if api_only else self._html_search_regex(
'description': self._og_search_description(webpage, default=None), r'(?s)<title>(?P<title>.*?)(?: \| Tumblr)?</title>', webpage, 'title')),
'thumbnail': self._og_search_thumbnail(webpage, default=None), 'description': description,
'thumbnail': (traverse_obj(video_json, ('poster', 0, 'url'))
or self._og_search_thumbnail(webpage, default=None)),
'uploader_id': uploader_id,
'uploader_url': f'https://{uploader_id}.tumblr.com/' if uploader_id else None,
'duration': duration, 'duration': duration,
'like_count': post_json.get('like_count'),
'repost_count': post_json.get('reblog_count'),
'age_limit': {True: 18, False: 0}.get(post_json.get('is_nsfw')),
'tags': post_json.get('tags'),
'formats': formats, 'formats': formats,
} }