5 from urllib
.parse
import urlparse
6 #from flask_login import current_user, login_required
7 from flask_login
import LoginManager
, UserMixin
, current_user
, login_user
, logout_user
, login_required
8 from flask
import Blueprint
, render_template
, request
, redirect
, flash
, url_for
, jsonify
, g
, current_app
9 from werkzeug
.exceptions
import BadRequest
, NotFound
, BadGateway
11 from ..common
.common
import *
12 from ..common
.anticaptcha
import submit_captcha
15 frontend
= Blueprint('youtube', __name__
,
16 template_folder
='templates',
17 static_folder
='static',
18 static_url_path
='/static/yt')
22 return redirect(url_for('.feed'), code
=302)
24 @frontend.route('/feed/subscriptions')
25 # disabled for guest user: @login_required
27 if current_user
.is_anonymous
:
30 if 'welcome_message' in cf
['frontend']:
31 flash(cf
['frontend']['welcome_message'], "welcome")
33 token
= current_user
.token
34 settings
= current_user
.get_settings()
35 page
= request
.args
.get('page', 0, type=int)
36 with sqlite3
.connect(cf
['global']['database']) as conn
:
40 SELECT videos.id, channel_id, name, title, length, livestream, premiere, shorts, published > datetime('now') as upcoming, published, playlist_videos.playlist_id, display
42 JOIN channels ON videos.channel_id = channels.id
43 LEFT JOIN playlist_videos ON (videos.id = playlist_videos.video_id)
44 LEFT JOIN flags ON (videos.id = flags.video_id) AND (flags.user = ?)
45 WHERE (channel_id IN (SELECT channel_id FROM subscriptions WHERE user=? AND type = 'channel')
46 OR playlist_videos.playlist_id IN (SELECT channel_id FROM subscriptions WHERE user=? AND type = 'playlist')
47 OR flags.display = 'pinned')
48 AND flags.display IS NOT 'hidden'
49 AND (flags.display = 'pinned' OR not ? or shorts is null or not shorts)
50 ORDER BY (display = 'pinned') DESC, crawled DESC
52 OFFSET 36*?""", (token
, token
, token
, settings
.get('noshorts', False), page
))
55 'channel_id': channel_id
,
59 'livestream': livestream
,
63 'published': published
,
65 'pinned': display
== 'pinned',
66 } for (video_id
, channel_id
, author
, title
, length
, livestream
, premiere
, shorts
, upcoming
, published
, playlist
, display
) in c
.fetchall()]
67 return render_template('index.html.j2', rows
=rows
, page
=page
)
69 @frontend.route('/watch')
71 if current_user
.is_anonymous
:
74 token
= current_user
.token
76 if not 'v' in request
.args
:
77 return "missing video id", 400
78 if len(request
.args
.get('v')) != 11:
79 return "malformed video id", 400
82 'content-type': 'text/plain',
83 'Link': "<data:text/css,body%7Bcolor:%23eee;background:%23333%7D>; rel=stylesheet;"
86 video_id
= request
.args
.get('v')
87 playlist
= request
.args
.get('list')
88 show
= request
.args
.get("show")
89 itag
= request
.args
.get("itag")
90 video_url
, stream_map
, metadata
, error
, errdetails
= get_video_info(video_id
, metaOnly
=(show
== 'meta'))
92 if itag
: # if itag is specified and available, force its use even if not highest resolution
93 video_url
= next((e
['url'] for e
in stream_map
['muxed'] if 'url' in e
and e
['itag'] == int(itag
)), video_url
)
95 proxy_on
= getattr(g
, 'proxy_on', None)
97 False: "Log in to use proxy.", # proxy for logged-in users only
98 None: "", # proxy disabled globally
99 True: "", # proxy is available to this user
102 'banned': "Instance is being rate limited.",
103 'malformed': "Video ID is invalid.",
104 'geolocked': f
"This video is geolocked. {proxy_msg}",
105 'livestream': f
"Livestreams disabled. {proxy_msg}",
106 'throttled': "Stream is severely throttled.",
107 'agegated': "Unable to bypass age-restriction.",
108 'scrambled': "VEVO style cyphering not supported.",
109 'no-url': "No muxed stream available.",
110 'exhausted': errdetails
or "Couldn't extract video URLs.",
111 'player': errdetails
,
114 # if the video is geolocked, and the proxy is enabled, we can still play
115 # it, if the video is available in the instance server's region:
116 if error
== 'geolocked' and video_url
and proxy_on
:
117 videoplayback
= url_for('proxy.videoplayback')
118 query
= urlparse(video_url
).query
119 video_url
= f
"{videoplayback}?{query}"
120 for t
in ('adaptive_video', 'adaptive_audio', 'muxed'):
121 for s
in stream_map
[t
]:
122 query
= urlparse(s
['url']).query
123 s
['url'] = f
"{videoplayback}?{query}"
126 # if the proxy is enabled, we can also play livestreams:
127 if error
== 'livestream' and proxy_on
:
128 # Note: hlsManifestUrl's hostname will be replaced client-side
129 video_url
= stream_map
['hlsManifestUrl']
132 # if the instance is blocked, try submitting a job to the anti captcha service:
133 if error
== 'banned' and cf
['captcha']['api_key']:
134 r2
= requests
.get(f
'https://www.youtube.com/watch?v={video_id}&hl=en&gl=US')
135 status
= submit_captcha(r2
)
137 raise Exception("we are banned, but captcha wasn't triggered!")
139 message
= "right now" if status
is True else f
"{int(status)} seconds ago"
140 raise BadGateway(f
"""
141 {errdetails} An attempt at getting unblocked has been made {message}.
142 Please try again in 30 seconds.
147 msg
= errdetails
if error
=='player' else f
"{error.upper()}: {errdetails}"
148 return msg
, 400, plaintextheaders
# TODO: nicer
149 return redirect(video_url
, code
=307)
151 if error
and not metadata
:
152 return {'error': True, error
: errdetails
}, 400 # TODO: better (test _CpR4o81XQc)
153 return jsonify(metadata
)
154 elif show
== "audio":
155 # sorting: we want to prioritize mp4a over opus, and sort by highest quality first
156 # todo: geolocking; prefer open format?
157 if error
and not stream_map
:
158 msg
= errdetails
if error
=='player' else f
"{error.upper()}: {errdetails}"
159 return msg
, 400, plaintextheaders
# TODO: nicer
160 stream
= next(iter(sorted(
161 stream_map
['adaptive_audio'],
162 key
=lambda e
: ('opus' not in e
['mimeType'], e
['bitrate']),
165 return redirect(stream
)
167 # this is the subset of (useful) keys that are not present in the
168 # Android API response. the special key '_' contains ready-to-use
169 # parsed versions of that data.
170 if error
and not metadata
:
171 return {'error': True, error
: errdetails
}, 400 # TODO: better (test _CpR4o81XQc)
172 parsed
= microformat_parser(metadata
)
173 return {'microformat': metadata
.get('microformat'),'cards':metadata
.get('cards'), '_':parsed
}
175 extra
= {'geolocked':'local=1', 'livestream':'raw=0'}.get(error
,'')
176 invidious_url
= f
"https://redirect.invidious.io/watch?v={video_id}&{extra}"
177 if error
and not metadata
: # e.g. malformed, private/deleted video, ...
178 return render_template('video-error.html.j2', video_id
=video_id
,
179 video_error
=error
, errdetails
=errdetails
, invidious_url
=invidious_url
)
180 meta
= prepare_metadata(metadata
)
181 with sqlite3
.connect(cf
['global']['database']) as conn
:
185 SELECT 1 FROM subscriptions WHERE channel_id = ? AND user = ?
187 SELECT 1 FROM flags WHERE video_id = ? AND display = 'pinned' AND user = ?
188 ))""", (meta
['channel_id'], token
, video_id
, token
))
189 (is_subscribed
, is_pinned
) = c
.fetchone()
190 return render_template('watch.html.j2',
191 video_id
=video_id
, video_url
=video_url
, stream_map
=stream_map
,
192 video_error
=error
, errdetails
=errdetails
, invidious_url
=invidious_url
,
193 playlist
=playlist
, is_pinned
=is_pinned
, is_subscribed
=is_subscribed
,
196 @frontend.route('/embed/videoseries')
197 def embed_videoseries():
198 return redirect(url_for('.playlist', list=request
.args
.get('list')))
199 @frontend.route('/embed/<video_id>', strict_slashes
=False)
201 if video_id
== "videoseries":
202 return redirect(url_for('.playlist', list=request
.args
.get('list')))
204 return redirect(url_for('.watch', v
=video_id
, t
=request
.args
.get('start')))
206 @frontend.route('/v/<video_id>', strict_slashes
=False)
207 @frontend.route('/live/<video_id>', strict_slashes
=False)
208 @frontend.route('/shorts/<video_id>', strict_slashes
=False)
209 def shorts_or_live(video_id
):
210 return redirect(url_for('.watch', v
=video_id
, t
=request
.args
.get('start')))
212 @frontend.route('/<something>', strict_slashes
=False)
213 def plain_user_or_video(something
):
214 # yt.com interprets this as a username, but we also want to catch youtu.be
215 # short-urls. so we check if it's a channel by querying the RSS feed (this
216 # shoudn't be rate-limited); if that fails, check if it looks like a video
217 # id; or finally give up.
218 if '.' not in something
and channel_exists(something
):
219 # periods are not valid in usernames, vanity urls or ucids, but common
220 # in urls that get crawled by bots (e.g. index.php). failing early
221 # reduces the amount of invalid channel names getting looked up.
222 return redirect(url_for('.channel', channel_id
=something
))
223 elif re
.match(r
"^[-_0-9A-Za-z]{11}$", something
): # looks like a video id
224 return redirect(url_for('.watch', v
=something
, t
=request
.args
.get('t')))
225 else: # ¯\_(ツ)_/¯
226 # XXX: something == 'thethoughtemporium' -> 404s
227 raise NotFound("Note: some usernames not recognized; try searching it")
229 @frontend.route('/attribution_link', strict_slashes
=False)
230 def attribution_link():
231 # /attribution_link?a=anything&u=/channel/UCZYTClx2T1of7BRZ86-8fow
232 # /attribution_link?a=JdfC0C9V6ZI&u=%2Fwatch%3Fv%3DEhxJLojIE_o%26feature%3Dshare
233 return redirect(request
.args
.get('u') or '/')
235 @frontend.route('/c/<username>/<subpage>')
236 @frontend.route('/c/<username>/')
237 @frontend.route('/user/<username>/<subpage>')
238 @frontend.route('/user/<username>/')
239 def channel_redirect(username
, subpage
=None):
240 # Note: we can't check /c/, so we have to assume it is the same as /user/,
241 # which is sometimes wrong.
242 xmlfeed
= fetch_xml("user", username
)
245 raise NotFound("unknown channel name")
247 _
, _
, _
, channel_id
, _
= parse_xml(xmlfeed
)
249 return redirect(url_for('.channel', channel_id
=channel_id
))
251 @frontend.route('/channel/<channel_id>/<subpage>')
252 @frontend.route('/channel/<channel_id>/')
253 def channel(channel_id
, _
=None):
254 token
= getattr(current_user
, 'token', 'guest')
255 sort
= request
.args
.get("sort", "newest")
257 if not re
.match(r
"(UC[A-Za-z0-9_-]{22})", channel_id
):
258 # canonicalize channel id, otherwise popular won't work
259 return redirect(url_for('.channel_redirect', channel_id
=channel_id
))
261 if sort
== "popular":
262 xmlfeed
= fetch_xml("playlist_id", f
"PU{channel_id[2:]}")
264 xmlfeed
= fetch_xml("channel_id", channel_id
)
265 #^note: could also use playlist_id=UU...
268 raise NotFound("unknown channel id")
270 title
, author
, videos
, _
, _
= parse_xml(xmlfeed
)
272 with sqlite3
.connect(cf
['global']['database']) as conn
:
277 WHERE channel_id = ? AND user = ?
278 """, (channel_id
, token
))
279 (is_subscribed
,) = c
.fetchone()
281 return render_template('xmlfeed.html.j2', title
=author
, rows
=videos
,
282 is_subscribed
=is_subscribed
, channel_id
=channel_id
)
284 @frontend.route('/playlist')
286 playlist_id
= request
.args
.get('list')
288 return "bad list id", 400 # todo
290 xmlfeed
= fetch_xml("playlist_id", playlist_id
)
292 return "not found or something", 404 # XXX
293 title
, author
, videos
, _
, _
= parse_xml(xmlfeed
)
294 return render_template('xmlfeed.html.j2', title
=f
"{title} by {author}", rows
=videos
)
296 @frontend.route('/api/timedtext')
298 r
= requests
.get("https://www.youtube.com/api/timedtext", request
.args
.to_dict())
299 # Note: in srv1 format, xmlentities are double-encoded m( a smart quote is
300 # even worse: it's '&39;<smartquote>' wtf!? (at least vvt seems ok)
302 return "error: {r.text}", 400 # TODO: better
304 if request
.args
.get('fmt') == 'vtt' and request
.args
.get('kind') == 'asr':
305 # autocaptions are extremely confusing, and stuck in the lower-left corner. fix it up a bit
306 retval
= re
.sub(r
"<.+?>", "", retval
) # remove inline html-like markup that times each word/adds styles
307 retval
= retval
.replace("align:start position:0%", "") # let browser position the text itself
308 # each subtitle-line is repeated twice (first on the lower line, then
309 # on the next "frame" on the upper line). we want to remove the
310 # repetition, as that's confusing without word and line animations:
311 lines
= retval
.split('\n')
312 retval
= '\n'.join([line
for line
, prev
in zip(lines
, ['']+lines
) if not " --> " in prev
])
313 return retval
, {'Content-Type': r
.headers
.get("Content-Type")}
315 @frontend.route('/feeds/videos.xml')
317 valid
= ("channel_id", "playlist_id", "user")
318 key
= next(iter(request
.args
), None)
319 if key
not in valid
: raise BadRequest
320 data
= fetch_xml(key
, request
.args
[key
])
321 if not data
: raise NotFound
322 return data
, {'content-type': 'text/xml; charset=UTF-8'}
324 @frontend.route('/manage/subscriptions')
325 # disabled for guest user: @login_required
326 def subscription_manager():
327 if current_user
.is_anonymous
:
330 token
= current_user
.token
331 with sqlite3
.connect(cf
['global']['database']) as conn
:
332 #with conn.cursor() as c:
335 SELECT subscriptions.channel_id, name, type,
336 (subscribed_until < datetime('now')) AS obsolete
338 LEFT JOIN (SELECT name, id FROM channels
340 SELECT name, id FROM playlists
341 ) AS channels ON channels.id = subscriptions.channel_id
342 left JOIN websub ON channels.id = websub.channel_id
344 AND subscriptions.type IN ('channel', 'playlist')
345 ORDER BY obsolete=0, name COLLATE NOCASE ASC""", (token
,))
347 'channel_id': channel_id
,
348 'author': author
or channel_id
,
350 'subscribed_until': subscribed_until
351 } for (channel_id
, author
, type, subscribed_until
) in c
.fetchall()]
352 return render_template('subscription_manager.html.j2', rows
=rows
)
354 @frontend.route('/feed/subscriptions', methods
=['POST'])
357 token
= current_user
.token
358 action
= next(iter(list(request
.form
.keys())), None)
359 if action
in ['pin', 'unpin', 'hide', 'unhide']:
360 video_id
= request
.form
.get(action
)
367 with sqlite3
.connect(cf
['global']['database']) as conn
:
369 store_video_metadata(video_id
) # only needed for pinning
371 INSERT OR REPLACE INTO flags (user, video_id, display)
373 """, (token
, video_id
, display
))
374 undo_flash(video_id
, action
)
376 flash("unsupported action", "error")
377 return redirect(request
.url
, code
=303)
379 @frontend.route('/manage/subscriptions', methods
=['POST'])
381 def manage_subscriptions():
382 token
= current_user
.token
383 if 'subscribe' in request
.form
:
384 some_id
= request
.form
.get("subscribe")
385 match
= re
.search(r
"(UC[A-Za-z0-9_-]{22})", some_id
)
387 some_id
= match
.group(1)
390 match
= re
.search(r
"((?:PL|LL|EC|UU|FL|UL|OL)[A-Za-z0-9_-]{10,})", some_id
)
391 if match
: # NOTE: PL-playlists are 32chars, others differ in length.
392 some_id
= match
.group(1)
395 flash("not a valid/subscribable URI", "error")
396 return redirect(request
.url
, code
=303)
397 with sqlite3
.connect(cf
['global']['database']) as conn
:
398 #with conn.cursor() as c:
401 INSERT OR IGNORE INTO subscriptions (user, channel_id, type)
403 """, (token
, some_id
, id_type
))
404 # TODO: sql-error-handling, asynchronically calling update-subs.pl
405 undo_flash(some_id
, 'subscribe')
407 elif 'unsubscribe' in request
.form
:
408 some_id
= request
.form
.get("unsubscribe")
409 with sqlite3
.connect(cf
['global']['database']) as conn
:
410 #with conn.cursor() as c:
413 DELETE FROM subscriptions
414 WHERE user = ? AND channel_id = ?
415 """, (token
, some_id
))
416 # TODO: sql-error-handling, report success
417 undo_flash(some_id
, 'unsubscribe')
420 flash("unsupported action", "error")
422 return redirect(request
.url
, code
=303)
424 @frontend.route('/vi/<vid>/<res>.jpg')
425 def redirect_thumbnails(vid
, res
):
426 return redirect(f
"https://i.ytimg.com{request.path}", code
=301)
429 def redirect_youtube_dot_com(state
):
431 This is executed when the blueprint is loaded dynamically builds a number
432 of routes so that URLs like
433 https://subscriptions.gir.st/https://www.youtube.com/watch?v=dQw4w9WgXcQ
434 redirect to the /watch page. Works with /watch, /embed/ and youtu.be short
435 links, with or without protocl and/or 'www'.
437 def real_redirect_youtube_dot_com(video_id
=None):
438 if not re
.match(r
"^[-_0-9A-Za-z]{11}$", video_id
or ''): video_id
= None
439 if not video_id
: video_id
= request
.args
.get('v')
440 return redirect(url_for('.watch', v
=video_id
))
442 for protocol
in ("", "http://", "https://"):
443 for prefix
in ("", "www.", "m."):
444 for domain
in ("youtube.com", "youtu.be", "youtube-nocookie.com"):
445 for urlpath
in ("/watch", "/embed/<video_id>", "/live/<video_id>", "/shorts/<video_id>", "/<video_id>"):
446 if domain
!= "youtu.be" and urlpath
== "/<video_id>":
447 continue # that's a channel, not a video
448 frontend
.add_url_rule(
449 f
"/{protocol}{prefix}{domain}{urlpath}",
450 view_func
=real_redirect_youtube_dot_com
,
454 def undo_flash(thing_id
, action
):
455 undo_action
, past_action
= {
456 'pin': ('unpin', 'pinned'),
457 'unpin': ('pin', 'unpinned'),
458 'hide': ('unhide', 'hidden'),
459 'unhide': ('hide', 'unhidden'),
460 'subscribe': ('unsubscribe', 'subscribed'),
461 'unsubscribe': ('subscribe', 'unsubscribed'),
463 if 'subscribe' in action
and thing_id
.startswith('UC'):
465 thing_url
= url_for('.channel', channel_id
=thing_id
)
466 elif 'subscribe' in action
:
468 thing_url
= url_for('.playlist', playlist_id
=thing_id
)
471 thing_url
= url_for('.watch', v
=thing_id
)
473 <form method=post><input type=hidden name="{undo_action}" value="{thing_id}">
474 <a href="{thing_url}">{thing}</a> {past_action}.
475 <label><input type="submit" hidden>
476 <span style="text-decoration:underline;cursor:pointer">undo</span>.
477 </label></form>''', "info")
479 @frontend.app_template_filter('format_date')
481 import datetime
# can't import at top level, because it is inherited from common
482 (y
,m
,d
) = (int(n
) for n
in s
.split('T')[0].split(' ')[0].split('-')) # iso-dates can seperate date from time with space or 'T'
483 M
= '_ Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec'.split()
484 if y
== datetime
.datetime
.now().year
:
487 return f
"{M[m]} '{y%100}"
489 @frontend.app_template_filter('format_time')
496 return '%d:%02d:%02d' % (h
,m
,s
) if h
else '%02d:%02d' % (m
,s
)
498 @frontend.app_template_filter('timeoffset')
502 match
= re
.match(r
"^(\d+)s?$", s
) # e.g. 2040s
504 return match
.group(1)
505 match
= re
.match(r
"^(?:(\d+)h)?(?:(\d+)m)?(?:(\d+)s)?$", s
) # e.g. 34m, 1h23s
507 return ":".join([n
.zfill(2) for n
in match
.groups('0')])