app/common/common.py

   1 import os
   2 import re
   3 import json
   4 import base64
   5 import sqlite3
   6 import requests
   7 import hmac, hashlib
   8 import requests_cache
   9 import dateutil.parser
  10 from xml.etree import ElementTree
  11 from configparser import ConfigParser
  12 from datetime import datetime, timezone
  13 from urllib.parse import parse_qs, urlparse
  14
  15 from .innertube import parse_infocard, parse_endcard
  16
  17 cf = ConfigParser()
  18 config_filename = os.environ.get('YT_CONFIG', '/etc/yt/config.ini')
  19 cf.read(config_filename)
  20 if not 'global' in cf: # todo: full config check
  21     raise Exception("Configuration file not found or empty")
  22
  23 # Note: currently expiring after 10 minutes. googlevideo-urls are valid for 5h59m, but this makes reddit very stale and premiere videos won't start.  TODO: exipre when video is livestream/premiere/etc
  24 requests_cache.install_cache(backend='memory', expire_after=10*60, allowable_codes=(200,))
  25
  26 # Note: this should only be required for the 'memory' backed cache.
  27 # TODO: only run for long-running processes, i.e. the frontend
  28 from threading import Timer
  29 def purge_cache(sec):
  30     requests_cache.remove_expired_responses()
  31     t = Timer(sec, purge_cache, args=(sec,))
  32     t.setDaemon(True)
  33     t.start()
  34 purge_cache(10*60)
  35
  36 # for debugging purposes, monkey patch requests session to store each requests-request in a flask-request's g object (url and response). we can then use a flask error_handler to include the request data in the error log.
  37 # since we also call config from outside the flask appcontext, it is wrapped in a try-catch block.
  38 from flask import g
  39 import requests
  40 from requests import Session as OriginalSession
  41 class _NSASession(OriginalSession):
  42     def request(self, method, url, params=None, data=None, **kwargs):
  43         response = super(_NSASession, self).request(
  44             method, url, params, data, **kwargs
  45             )
  46         try:
  47             if 'api_requests' not in g:
  48                 g.api_requests = []
  49             g.api_requests.append((url, params, response.text))
  50         except RuntimeError: pass # not within flask (e.g. utils.py)
  51         return response
  52 requests.Session = requests.sessions.Session = _NSASession
  53
  54 def fetch_xml(feed_type, feed_id):
  55     # TODO: handle requests.exceptions.ConnectionError
  56     r = requests.get("https://www.youtube.com/feeds/videos.xml", {
  57         feed_type: feed_id,
  58     })
  59     if not r.ok:
  60         return None
  61
  62     return r.content
  63
  64 def parse_xml(xmldata):
  65     ns = {
  66         'atom':"http://www.w3.org/2005/Atom",
  67         'yt': "http://www.youtube.com/xml/schemas/2015",
  68         'media':"http://search.yahoo.com/mrss/",
  69         'at': "http://purl.org/atompub/tombstones/1.0",
  70     }
  71
  72     feed = ElementTree.fromstring(xmldata)
  73
  74     if feed.find('at:deleted-entry',ns):
  75         (_,_,vid) = feed.find('at:deleted-entry',ns).get('ref').rpartition(':')
  76         return None, None, [{'deleted': True, 'video_id': vid}], None, None
  77
  78     title = feed.find('atom:title',ns).text
  79     author = feed.find('atom:author/atom:name',ns).text \
  80         if feed.find('atom:author',ns) else None
  81     # for /user/<> endpoint: find out UC-id:
  82     # for playlists: this is who created the playlist:
  83     try:   channel_id = feed.find('yt:channelId',ns).text
  84     except:channel_id=None # XXX: why does ternary not work!?
  85     # for pullsub: if this exists, we're looking at a playlist:
  86     try:   playlist_id = feed.find('yt:playlistId',ns).text
  87     except:playlist_id=None # XXX: why does ternary not work!?
  88     videos = []
  89     for entry in feed.findall('atom:entry',ns):
  90         videos.append({
  91             'video_id': entry.find('yt:videoId',ns).text,
  92             'title': entry.find('atom:title',ns).text,
  93             'published': entry.find('atom:published',ns).text,
  94             'channel_id': entry.find('yt:channelId',ns).text,
  95             'author': entry.find('atom:author',ns).find('atom:name',ns).text,
  96             # extra fields for pull_subs/webhook:
  97             'updated': entry.find('atom:updated',ns).text,
  98         })
  99
 100     return title, author, videos, channel_id, playlist_id
 101
 102 def update_channel(db, xmldata, from_webhook=False):
 103     if not xmldata: return False
 104
 105     # Note: websub does not return global author, hence taking from first video
 106     title, author, videos, channel, playlist = parse_xml(xmldata)
 107
 108     c = db.cursor()
 109     from flask import current_app # XXX: remove
 110     for i, video in enumerate(videos):
 111         if video.get('deleted'):
 112             if from_webhook: current_app.logger.warning(f"ignoring deleted video {video['video_id']}") # XXX: remove
 113             # TODO: enable once we enforce hmac validation:
 114             #c.execute("DELETE FROM videos WHERE id = ?", (video['video_id'],))
 115             break
 116
 117         now = datetime.now(timezone.utc)
 118         updated = dateutil.parser.parse(video['updated'])
 119         published = dateutil.parser.parse(video['published'])
 120         # if update and published time are near-identical, we assume it's new.
 121         # checking if it was posted this week is necessary during xmlfeed pulling.
 122         if (updated - published).seconds < 60 and (now - published).days < 7:
 123             timestamp = now
 124             if from_webhook: current_app.logger.warning(f"fresh video {video['video_id']}") # XXX: remove
 125         else:#, it might just an update to an older video, or a previously unlisted one.
 126             # first, assume it's an older video (correct when pulling xmlfeeds)
 127             timestamp = published
 128             # then, check if we don't know about it and if so, look up the real date.
 129
 130             # The 'published' timestamp sent in websub POSTs are often wrong (e.g.:
 131             # video gets uploaded as unlisted on day A and set to public on day B;
 132             # the webhook is sent on day B, but 'published' says A. The video
 133             # therefore looks like it's just an update to an older video). If
 134             # that's the case, we fetch get_video_info and double-check.
 135             # We only need to do this to not-yet-in-the-database videos.
 136             c.execute("SELECT 1 from videos where id = ?", (video['video_id'],))
 137             new_video = len(c.fetchall()) < 1
 138             if from_webhook: current_app.logger.warning(f"video {video['video_id']}") # XXX: remove
 139             if from_webhook and new_video:
 140                 if from_webhook: current_app.logger.warning(f"  is webhook and new") # XXX: remove
 141                 _, _, meta, _, _ = get_video_info(video['video_id'])
 142                 if meta:
 143                     meta = prepare_metadata(meta)
 144                     published = dateutil.parser.parse(meta['published'])
 145                     if from_webhook: current_app.logger.warning(f"  uploaded {published}") # XXX: remove
 146                     if (now - published).days < 7:
 147                         timestamp = now
 148                     else:#, it's just an update to an older video.
 149                         timestamp = published
 150
 151         c.execute("""
 152             INSERT OR IGNORE INTO videos (id, channel_id, title, published, crawled)
 153                            VALUES (?, ?, ?, datetime(?), datetime(?))
 154         """, (
 155             video['video_id'],
 156             video['channel_id'],
 157             video['title'],
 158             video['published'],
 159             timestamp
 160         ))
 161
 162         # for channels, this is obviously always the same, but playlists can
 163         # consist of videos from different channels:
 164         if i == 0 or playlist:
 165             c.execute("""
 166                 INSERT OR REPLACE INTO channels (id, name)
 167                                 VALUES (?, ?)
 168             """, (video['channel_id'], video['author']))
 169
 170         # keep track of which videos are in a playlist, so we can show the user
 171         # why a video is in their feed:
 172         if playlist:
 173             c.execute("""
 174                 INSERT OR IGNORE INTO playlist_videos (video_id, playlist_id)
 175                                VALUES (?, ?)
 176             """, (video['video_id'], playlist))
 177
 178     if playlist and not from_webhook: # Note: playlists can't get updated via websub
 179         c.execute("""
 180             INSERT OR REPLACE INTO playlists (id, name, author)
 181                             VALUES (?, ?, ?)
 182             """, (playlist, title, channel))
 183         c.execute("""
 184             INSERT OR REPLACE INTO channels (id, name)
 185                             VALUES (?, ?)
 186         """, (channel, author))
 187
 188     db.commit()
 189
 190     return True
 191
 192 def get_video_info(video_id, sts=0, algo=""):
 193     """
 194     returns: best-quality muxed video stream, stream map, player_response, error-type/mesage
 195     error types: player, malformed, livestream, geolocked, exhausted
 196     """
 197     player_error = None # for 'exhausted'
 198     for el in ['embedded', 'detailpage']:#sometimes, only one or the other works
 199         r = requests.get("https://www.youtube.com/get_video_info", {
 200             "video_id": video_id,
 201             "eurl": f"https://youtube.googleapis.com/v/{video_id}",
 202             "el": el,
 203             "sts": sts,
 204             "hl": "en_US",
 205         })
 206         params = parse_qs(r.text)
 207         if 'errorcode' in params: # status=fail
 208             return None, None, None, 'malformed', params['reason'][0]
 209
 210         metadata = json.loads(params.get('player_response')[0])
 211         playabilityStatus = metadata['playabilityStatus']['status']
 212         if playabilityStatus != "OK":
 213             playabilityReason = metadata['playabilityStatus'].get('reason',
 214                     '//'.join(metadata['playabilityStatus'].get('messages',[])))
 215             player_error = f"{playabilityStatus}: {playabilityReason}"
 216             if playabilityStatus == "UNPLAYABLE":
 217                 continue  # try again with next el value (or fail as exhausted)
 218             # without videoDetails, there's only the error message
 219             maybe_metadata = metadata if 'videoDetails' in metadata else None
 220             return None, None, maybe_metadata, 'player', player_error
 221         if metadata['videoDetails'].get('isLive', False):
 222             return None, None, metadata, 'livestream', None
 223
 224         if not 'formats' in metadata['streamingData']:
 225             continue # no urls
 226
 227         formats = metadata['streamingData']['formats']
 228         for (i,v) in enumerate(formats):
 229             if not ('cipher' in v or 'signatureCipher' in v): continue
 230             cipher = parse_qs(v.get('cipher') or v.get('signatureCipher'))
 231             formats[i]['url'] = unscramble(cipher, algo)
 232
 233         adaptive = metadata['streamingData']['adaptiveFormats']
 234         for (i,v) in enumerate(adaptive):
 235             if not ('cipher' in v or 'signatureCipher' in v): continue
 236             cipher = parse_qs(v.get('cipher') or v.get('signatureCipher'))
 237             adaptive[i]['url'] = unscramble(cipher, algo)
 238
 239         stream_map = {'adaptive': adaptive, 'muxed': formats}
 240
 241         # todo: check if we have urls or try again
 242         url = sorted(formats, key=lambda k: k['height'], reverse=True)[0]['url']
 243
 244         # ip-locked videos can be recovered if the proxy module is loaded:
 245         is_geolocked = 'geolocked' if 'gcr' in parse_qs(urlparse(url).query) else None
 246
 247         return url, stream_map, metadata, is_geolocked, None
 248     else:
 249         return None, None, metadata, 'exhausted', player_error
 250
 251 def unscramble(cipher, algo):  # test video id: UxxajLWwzqY
 252     signature = list(cipher['s'][0])
 253     for c in algo.split():
 254         op, ix = re.match(r"([rsw])(\d+)?", c).groups()
 255         ix = int(ix) % len(signature) if ix else 0
 256         if not op: continue
 257         if op == 'r': signature = list(reversed(signature))
 258         if op == 's': signature = signature[ix:]
 259         if op == 'w': signature[0], signature[ix] = signature[ix], signature[0]
 260     sp = cipher.get('sp', ['signature'])[0]
 261     sig = cipher.get('sig', [''.join(signature)])[0]
 262     return f"{cipher['url'][0]}&{sp}={sig}"
 263
 264 def prepare_metadata(metadata):
 265     meta1 = metadata['videoDetails']
 266     meta2 = metadata['microformat']['playerMicroformatRenderer']
 267     cards = metadata['cards']['cardCollectionRenderer']['cards'] \
 268         if 'cards' in metadata else []
 269     endsc = metadata['endscreen']['endscreenRenderer']['elements'] \
 270         if 'endscreen' in metadata else []
 271
 272     # the actual video streams have exact information:
 273     try:
 274         sd = metadata['streamingData']
 275         some_stream = (sd.get('adaptiveFormats',[]) + sd.get('formats',[]))[0]
 276         aspect_ratio = some_stream['width'] / some_stream['height']
 277     # if that's unavailable (e.g. on livestreams), fall back to
 278     # thumbnails (only either 4:3 or 16:9).
 279     except:
 280         some_img = meta2['thumbnail']['thumbnails'][0]
 281         aspect_ratio = some_img['width'] / some_img['height']
 282
 283     # Note: we could get subtitles in multiple formats directly by querying
 284     # https://video.google.com/timedtext?hl=en&type=list&v=<VIDEO_ID> followed by
 285     # https://www.youtube.com/api/timedtext?lang=<LANG_CODE>&v=<VIDEO_ID>&fmt={srv1|srv2|srv3|ttml|vtt},
 286     # but that won't give us autogenerated subtitles (and is an extra request).
 287     # we can still add &fmt= to the extracted URLs below (first one takes precedence).
 288     try: # find the native language captions (assuming there is only 1 audioTrack) (any level might not exist):
 289         default_track = metadata.get('captions',{}).get('playerCaptionsTracklistRenderer',{}).get('defaultAudioTrackIndex', 0)
 290         main_subtitle = metadata['captions']['playerCaptionsTracklistRenderer']['audioTracks'][default_track]['defaultCaptionTrackIndex']
 291     except:
 292         main_subtitle = -1
 293     subtitles = sorted([
 294         {'url':cc['baseUrl'],
 295          'code':cc['languageCode'],
 296          'autogenerated':cc.get('kind')=="asr",
 297          'name':cc['name']['simpleText'],
 298          'default':i==main_subtitle,
 299          'query':"fmt=vtt&"+urlparse(cc['baseUrl']).query} # for our internal proxy
 300         for i,cc in enumerate(metadata.get('captions',{})
 301             .get('playerCaptionsTracklistRenderer',{})
 302             .get('captionTracks',[]))
 303     # sort order: default lang gets weight 0 (first), other manually translated weight 1, autogenerated weight 2:
 304     ], key=lambda cc: (not cc['default']) + cc['autogenerated'])
 305
 306     infocards = [parse_infocard(card) for card in cards]
 307     endcards = [parse_endcard(card) for card in endsc]
 308     # combine cards to weed out duplicates. for videos and playlists prefer
 309     # infocards, for channels and websites prefer endcards, as those have more
 310     # information than the other.
 311     # if the card type is not in ident, we use the whole card for comparison
 312     # (otherwise they'd all replace each other)
 313     ident = { # ctype -> ident
 314         'VIDEO': 'video_id',
 315         'PLAYLIST': 'playlist_id',
 316         'CHANNEL': 'channel_id',
 317         'WEBSITE': 'url',
 318         'POLL': 'question',
 319     }
 320     getident = lambda c: c['content'].get(ident.get(c['type']), c)
 321     mkexclude = lambda cards, types: [getident(c) for c in cards if c['type'] in types]
 322     exclude = lambda cards, without: [c for c in cards if getident(c) not in without]
 323
 324     allcards = exclude(infocards, mkexclude(endcards,  ['CHANNEL','WEBSITE'])) + \
 325                exclude(endcards,  mkexclude(infocards, ['VIDEO','PLAYLIST']))
 326
 327     all_countries = """AD AE AF AG AI AL AM AO AQ AR AS AT AU AW AX AZ BA BB BD
 328         BE BF BG BH BI BJ BL BM BN BO BQ BR BS BT BV BW BY BZ CA CC CD CF CG CH
 329         CI CK CL CM CN CO CR CU CV CW CX CY CZ DE DJ DK DM DO DZ EC EE EG EH ER
 330         ES ET FI FJ FK FM FO FR GA GB GD GE GF GG GH GI GL GM GN GP GQ GR GS GT
 331         GU GW GY HK HM HN HR HT HU ID IE IL IM IN IO IQ IR IS IT JE JM JO JP KE
 332         KG KH KI KM KN KP KR KW KY KZ LA LB LC LI LK LR LS LT LU LV LY MA MC MD
 333         ME MF MG MH MK ML MM MN MO MP MQ MR MS MT MU MV MW MX MY MZ NA NC NE NF
 334         NG NI NL NO NP NR NU NZ OM PA PE PF PG PH PK PL PM PN PR PS PT PW PY QA
 335         RE RO RS RU RW SA SB SC SD SE SG SH SI SJ SK SL SM SN SO SR SS ST SV SX
 336         SY SZ TC TD TF TG TH TJ TK TL TM TN TO TR TT TV TW TZ UA UG UM US UY UZ
 337         VA VC VE VG VI VN VU WF WS YE YT ZA ZM ZW""".split()
 338     whitelisted = sorted(meta2.get('availableCountries',[]))
 339     blacklisted = sorted(set(all_countries) - set(whitelisted))
 340
 341     published_at = f"{meta2['publishDate']}T00:00:00Z" # yyyy-mm-dd
 342     # 'premiere' videos (and livestreams?) have a ISO8601 date available:
 343     if 'liveBroadcastDetails' in meta2 and 'startTimestamp' in meta2['liveBroadcastDetails']: # TODO: tighten up
 344         published_at = meta2['liveBroadcastDetails']['startTimestamp']
 345
 346     return {
 347         'title': meta1['title'],
 348         'author': meta1['author'],
 349         'channel_id': meta1['channelId'],
 350         'description': meta1['shortDescription'],
 351         'published': published_at,
 352         'views': meta1['viewCount'],
 353         'length': int(meta1['lengthSeconds']),
 354         'rating': meta1['averageRating'],
 355         'category': meta2['category'],
 356         'aspectr': aspect_ratio,
 357         'unlisted': meta2['isUnlisted'],
 358         'whitelisted': whitelisted,
 359         'blacklisted': blacklisted,
 360         'poster': meta2['thumbnail']['thumbnails'][0]['url'],
 361         'infocards': infocards,
 362         'endcards': endcards,
 363         'all_cards': allcards,
 364         'subtitles': subtitles,
 365     }
 366
 367 def store_video_metadata(video_id):
 368     # check if we know about it, and if not, fetch and store video metadata
 369     with sqlite3.connect(cf['global']['database']) as conn:
 370         c = conn.cursor()
 371         c.execute("SELECT 1 from videos where id = ?", (video_id,))
 372         new_video = len(c.fetchall()) < 1
 373         if new_video:
 374             _, _, meta, _, _ = get_video_info(video_id)
 375             if meta:
 376                 meta = prepare_metadata(meta)
 377                 c.execute("""
 378                     INSERT OR IGNORE INTO videos (id, channel_id, title, published, crawled)
 379                                    VALUES (?, ?, ?, datetime(?), datetime(?))
 380                 """, (
 381                     video_id,
 382                     meta['channel_id'],
 383                     meta['title'],
 384                     meta['published'],
 385                     meta['published'],
 386                 ))
 387                 c.execute("""
 388                     INSERT OR REPLACE INTO channels (id, name)
 389                                     VALUES (?, ?)
 390                 """, (meta['channel_id'], meta['author']))
 391
 392 from werkzeug.exceptions import NotFound
 393 class NoFallbackException(NotFound): pass
 394 def fallback_route(*args, **kwargs): # TODO: worthy as a flask-extension?
 395     """
 396     finds the next route that matches the current url rule, and executes it.
 397     args, kwargs: pass all arguments of the current route
 398     """
 399     from flask import current_app, request, g
 400
 401     # build a list of endpoints that match the current request's url rule:
 402     matching = [
 403         rule.endpoint
 404         for rule in current_app.url_map.iter_rules()
 405         if rule.rule == request.url_rule.rule
 406     ]
 407     current = matching.index(request.endpoint)
 408
 409     # since we can't change request.endpoint, we always get the original
 410     # endpoint back. so for repeated fall throughs, we use the g object to
 411     # increment how often we want to fall through.
 412     if not '_fallback_next' in g:
 413         g._fallback_next = 0
 414     g._fallback_next += 1
 415
 416     next_ep = current + g._fallback_next
 417
 418     if next_ep < len(matching):
 419         return current_app.view_functions[matching[next_ep]](*args, **kwargs)
 420     else:
 421         raise NoFallbackException
 422
 423 def websub_url_hmac(key, feed_id, timestamp, nonce):
 424     """ generate sha1 hmac, as required by websub/pubsubhubbub """
 425     sig_input = f"{feed_id}:{timestamp}:{nonce}".encode('ascii')
 426     return hmac.new(key.encode('ascii'), sig_input, hashlib.sha1).hexdigest()
 427
 428 def websub_body_hmac(key, body):
 429     return hmac.new(key.encode('ascii'), body, hashlib.sha1).hexdigest()
 430
 431 def pp(*args):
 432     from pprint import pprint
 433     import sys, codecs
 434     pprint(args, stream=codecs.getwriter("utf-8")(sys.stderr.buffer))