mygpo/api/advanced/__init__.py

   1 #
   2 # This file is part of my.gpodder.org.
   3 #
   4 # my.gpodder.org is free software: you can redistribute it and/or modify it
   5 # under the terms of the GNU Affero General Public License as published by
   6 # the Free Software Foundation, either version 3 of the License, or (at your
   7 # option) any later version.
   8 #
   9 # my.gpodder.org is distributed in the hope that it will be useful, but
  10 # WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
  11 # or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Affero General Public
  12 # License for more details.
  13 #
  14 # You should have received a copy of the GNU Affero General Public License
  15 # along with my.gpodder.org. If not, see <http://www.gnu.org/licenses/>.
  16 #
  17
  18 from functools import partial
  19 from itertools import imap, chain
  20 from collections import defaultdict, namedtuple
  21 from datetime import datetime
  22 from importlib import import_module
  23
  24 import dateutil.parser
  25
  26 try:
  27     import gevent
  28 except ImportError:
  29     gevent = None
  30
  31 from django.http import HttpResponse, HttpResponseBadRequest, Http404, HttpResponseNotFound
  32 from django.contrib.sites.models import RequestSite
  33 from django.views.decorators.csrf import csrf_exempt
  34 from django.views.decorators.cache import never_cache
  35 from django.utils.decorators import method_decorator
  36 from django.views.generic.base import View
  37 from django.conf import settings
  38
  39 from mygpo.api.constants import EPISODE_ACTION_TYPES, DEVICE_TYPES
  40 from mygpo.api.httpresponse import JsonResponse
  41 from mygpo.api.sanitizing import sanitize_url, sanitize_urls
  42 from mygpo.api.advanced.directory import episode_data, podcast_data
  43 from mygpo.api.backend import get_device, BulkSubscribe
  44 from mygpo.log import log
  45 from mygpo.utils import parse_time, format_time, parse_bool, get_timestamp
  46 from mygpo.decorators import allowed_methods, repeat_on_conflict
  47 from mygpo.core import models
  48 from mygpo.core.models import SanitizingRule, Podcast
  49 from mygpo.core.tasks import auto_flattr_episode
  50 from mygpo.users.models import PodcastUserState, EpisodeAction, \
  51      EpisodeUserState, DeviceDoesNotExist, DeviceUIDException, \
  52      InvalidEpisodeActionAttributes
  53 from mygpo.users.settings import FLATTR_AUTO
  54 from mygpo.core.json import json, JSONDecodeError
  55 from mygpo.api.basic_auth import require_valid_user, check_username
  56 from mygpo.db.couchdb import BulkException, bulk_save_retry
  57 from mygpo.db.couchdb.episode import episode_by_id, \
  58          favorite_episodes_for_user, episodes_for_podcast
  59 from mygpo.db.couchdb.podcast import podcast_for_url
  60 from mygpo.db.couchdb.podcast_state import subscribed_podcast_ids_by_device
  61 from mygpo.db.couchdb.episode_state import get_podcasts_episode_states, \
  62          episode_state_for_ref_urls, get_episode_actions
  63
  64
  65 # keys that are allowed in episode actions
  66 EPISODE_ACTION_KEYS = ('position', 'episode', 'action', 'device', 'timestamp',
  67                        'started', 'total', 'podcast')
  68
  69
  70 @csrf_exempt
  71 @require_valid_user
  72 @check_username
  73 @never_cache
  74 @allowed_methods(['GET', 'POST'])
  75 def subscriptions(request, username, device_uid):
  76
  77     now = datetime.now()
  78     now_ = get_timestamp(now)
  79
  80     if request.method == 'GET':
  81
  82         try:
  83             device = request.user.get_device_by_uid(device_uid)
  84         except DeviceDoesNotExist as e:
  85             return HttpResponseNotFound(str(e))
  86
  87         since_ = request.GET.get('since', None)
  88         if since_ is None:
  89             return HttpResponseBadRequest('parameter since missing')
  90         try:
  91             since = datetime.fromtimestamp(float(since_))
  92         except ValueError:
  93             return HttpResponseBadRequest('since-value is not a valid timestamp')
  94
  95         changes = get_subscription_changes(request.user, device, since, now)
  96
  97         return JsonResponse(changes)
  98
  99     elif request.method == 'POST':
 100         d = get_device(request.user, device_uid,
 101                 request.META.get('HTTP_USER_AGENT', ''))
 102
 103         if not request.body:
 104             return HttpResponseBadRequest('POST data must not be empty')
 105
 106         actions = json.loads(request.body)
 107         add = actions['add'] if 'add' in actions else []
 108         rem = actions['remove'] if 'remove' in actions else []
 109
 110         add = filter(None, add)
 111         rem = filter(None, rem)
 112
 113         try:
 114             update_urls = update_subscriptions(request.user, d, add, rem)
 115         except ValueError, e:
 116             return HttpResponseBadRequest(e)
 117
 118         return JsonResponse({
 119             'timestamp': now_,
 120             'update_urls': update_urls,
 121             })
 122
 123
 124 def update_subscriptions(user, device, add, remove):
 125
 126     for a in add:
 127         if a in remove:
 128             raise ValueError('can not add and remove %s at the same time' % a)
 129
 130     add_s = list(sanitize_urls(add, 'podcast'))
 131     rem_s = list(sanitize_urls(remove, 'podcast'))
 132
 133     assert len(add) == len(add_s) and len(remove) == len(rem_s)
 134
 135     updated_urls = filter(lambda (a, b): a != b, zip(add + remove, add_s + rem_s))
 136
 137     add_s = filter(None, add_s)
 138     rem_s = filter(None, rem_s)
 139
 140     # If two different URLs (in add and remove) have
 141     # been sanitized to the same, we ignore the removal
 142     rem_s = filter(lambda x: x not in add_s, rem_s)
 143
 144     subscriber = BulkSubscribe(user, device)
 145
 146     for a in add_s:
 147         subscriber.add_action(a, 'subscribe')
 148
 149     for r in rem_s:
 150         subscriber.add_action(r, 'unsubscribe')
 151
 152     try:
 153         subscriber.execute()
 154     except BulkException as be:
 155         for err in be.errors:
 156             log('Advanced API: %(username)s: Updating subscription for '
 157                     '%(podcast_url)s on %(device_uid)s failed: '
 158                     '%(rerror)s (%(reason)s)'.format(username=user.username,
 159                         podcast_url=err.doc, device_uid=device.uid,
 160                         error=err.error, reason=err.reason)
 161                 )
 162
 163     return updated_urls
 164
 165
 166 def get_subscription_changes(user, device, since, until):
 167     add_urls, rem_urls = device.get_subscription_changes(since, until)
 168     until_ = get_timestamp(until)
 169     return {'add': add_urls, 'remove': rem_urls, 'timestamp': until_}
 170
 171
 172 @csrf_exempt
 173 @require_valid_user
 174 @check_username
 175 @never_cache
 176 @allowed_methods(['GET', 'POST'])
 177 def episodes(request, username, version=1):
 178
 179     version = int(version)
 180     now = datetime.now()
 181     now_ = get_timestamp(now)
 182     ua_string = request.META.get('HTTP_USER_AGENT', '')
 183
 184     if request.method == 'POST':
 185         try:
 186             actions = json.loads(request.body)
 187         except (JSONDecodeError, UnicodeDecodeError) as e:
 188             msg = 'Advanced API: could not decode episode update POST data for user %s: %s' % (username, e)
 189             log(msg)
 190             return HttpResponseBadRequest(msg)
 191
 192         log('start: user %s: %d actions from %s' % (request.user._id, len(actions), ua_string))
 193
 194         # handle in background
 195         if len(actions) > settings.API_ACTIONS_MAX_NONBG:
 196             bg_handler = settings.API_ACTIONS_BG_HANDLER
 197             if bg_handler is not None:
 198
 199                 modname, funname = bg_handler.rsplit('.', 1)
 200                 mod = import_module(modname)
 201                 fun = getattr(mod, funname)
 202
 203                 fun(request.user, actions, now, ua_string)
 204
 205                 # TODO: return 202 Accepted
 206                 return JsonResponse({'timestamp': now_, 'update_urls': []})
 207
 208
 209         try:
 210             update_urls = update_episodes(request.user, actions, now, ua_string)
 211         except DeviceUIDException as e:
 212             import traceback
 213             s = u'could not update episodes for user %s: %s %s: %s' % (username, e, traceback.format_exc(), actions)
 214             log(s.decode('utf-8', errors='ignore'))
 215             return HttpResponseBadRequest(str(e))
 216         except InvalidEpisodeActionAttributes as e:
 217             import traceback
 218             log(u'could not update episodes for user %s: %s %s: %s' % (username, e, traceback.format_exc(), actions))
 219             return HttpResponseBadRequest(str(e))
 220
 221         log('done:  user %s: %d actions from %s' % (request.user._id, len(actions), ua_string))
 222         return JsonResponse({'timestamp': now_, 'update_urls': update_urls})
 223
 224     elif request.method == 'GET':
 225         podcast_url= request.GET.get('podcast', None)
 226         device_uid = request.GET.get('device', None)
 227         since_     = request.GET.get('since', None)
 228         aggregated = parse_bool(request.GET.get('aggregated', False))
 229
 230         try:
 231             since = int(since_) if since_ else None
 232         except ValueError:
 233             return HttpResponseBadRequest('since-value is not a valid timestamp')
 234
 235         if podcast_url:
 236             podcast = podcast_for_url(podcast_url)
 237             if not podcast:
 238                 raise Http404
 239         else:
 240             podcast = None
 241
 242         if device_uid:
 243
 244             try:
 245                 device = request.user.get_device_by_uid(device_uid)
 246             except DeviceDoesNotExist as e:
 247                 return HttpResponseNotFound(str(e))
 248
 249         else:
 250             device = None
 251
 252         changes = get_episode_changes(request.user, podcast, device, since,
 253                 now_, aggregated, version)
 254
 255         return JsonResponse(changes)
 256
 257
 258
 259 def convert_position(action):
 260     """ convert position parameter for API 1 compatibility """
 261     pos = getattr(action, 'position', None)
 262     if pos is not None:
 263         action.position = format_time(pos)
 264     return action
 265
 266
 267
 268 def get_episode_changes(user, podcast, device, since, until, aggregated, version):
 269
 270     devices = dict( (dev.id, dev.uid) for dev in user.devices )
 271
 272     args = {}
 273     if podcast is not None:
 274         args['podcast_id'] = podcast.get_id()
 275
 276     if device is not None:
 277         args['device_id'] = device.id
 278
 279     actions = get_episode_actions(user._id, since, until, **args)
 280
 281     if version == 1:
 282         actions = imap(convert_position, actions)
 283
 284     clean_data = partial(clean_episode_action_data,
 285             user=user, devices=devices)
 286
 287     actions = map(clean_data, actions)
 288     actions = filter(None, actions)
 289
 290     if aggregated:
 291         actions = dict( (a['episode'], a) for a in actions ).values()
 292
 293     return {'actions': actions, 'timestamp': until}
 294
 295
 296
 297
 298 def clean_episode_action_data(action, user, devices):
 299
 300     if None in (action.get('podcast', None), action.get('episode', None)):
 301         return None
 302
 303     if 'device_id' in action:
 304         device_id = action['device_id']
 305         device_uid = devices.get(device_id)
 306         if device_uid:
 307             action['device'] = device_uid
 308
 309         del action['device_id']
 310
 311     # remove superfluous keys
 312     for x in action.keys():
 313         if x not in EPISODE_ACTION_KEYS:
 314             del action[x]
 315
 316     # set missing keys to None
 317     for x in EPISODE_ACTION_KEYS:
 318         if x not in action:
 319             action[x] = None
 320
 321     if action['action'] != 'play':
 322         if 'position' in action:
 323             del action['position']
 324
 325         if 'total' in action:
 326             del action['total']
 327
 328         if 'started' in action:
 329             del action['started']
 330
 331         if 'playmark' in action:
 332             del action['playmark']
 333
 334     else:
 335         action['position'] = action.get('position', False) or 0
 336
 337     return action
 338
 339
 340
 341
 342
 343 def update_episodes(user, actions, now, ua_string):
 344     update_urls = []
 345
 346     grouped_actions = defaultdict(list)
 347
 348     # group all actions by their episode
 349     for action in actions:
 350
 351         podcast_url = action['podcast']
 352         podcast_url = sanitize_append(podcast_url, 'podcast', update_urls)
 353         if podcast_url == '':
 354             continue
 355
 356         episode_url = action['episode']
 357         episode_url = sanitize_append(episode_url, 'episode', update_urls)
 358         if episode_url == '':
 359             continue
 360
 361         act = parse_episode_action(action, user, update_urls, now, ua_string)
 362         grouped_actions[ (podcast_url, episode_url) ].append(act)
 363
 364
 365     auto_flattr_episodes = []
 366
 367     # Prepare the updates for each episode state
 368     obj_funs = []
 369
 370     for (p_url, e_url), action_list in grouped_actions.iteritems():
 371         episode_state = episode_state_for_ref_urls(user, p_url, e_url)
 372
 373         if any(a['action'] == 'play' for a in actions):
 374             auto_flattr_episodes.append(episode_state.episode)
 375
 376         fun = partial(update_episode_actions, action_list=action_list)
 377         obj_funs.append( (episode_state, fun) )
 378
 379     bulk_save_retry(obj_funs)
 380
 381     if user.get_wksetting(FLATTR_AUTO):
 382         for episode_id in auto_flattr_episodes:
 383             auto_flattr_episode.delay(user, episode_id)
 384
 385     return update_urls
 386
 387
 388 def update_episode_actions(episode_state, action_list):
 389     """ Adds actions to the episode state and saves if necessary """
 390
 391     len1 = len(episode_state.actions)
 392     episode_state.add_actions(action_list)
 393
 394     if len(episode_state.actions) == len1:
 395         return None
 396
 397     return episode_state
 398
 399
 400
 401 def parse_episode_action(action, user, update_urls, now, ua_string):
 402     action_str  = action.get('action', None)
 403     if not valid_episodeaction(action_str):
 404         raise Exception('invalid action %s' % action_str)
 405
 406     new_action = EpisodeAction()
 407
 408     new_action.action = action['action']
 409
 410     if action.get('device', False):
 411         device = get_device(user, action['device'], ua_string)
 412         new_action.device = device.id
 413
 414     if action.get('timestamp', False):
 415         new_action.timestamp = dateutil.parser.parse(action['timestamp'])
 416     else:
 417         new_action.timestamp = now
 418     new_action.timestamp = new_action.timestamp.replace(microsecond=0)
 419
 420     new_action.upload_timestamp = get_timestamp(now)
 421
 422     new_action.started = action.get('started', None)
 423     new_action.playmark = action.get('position', None)
 424     new_action.total = action.get('total', None)
 425
 426     return new_action
 427
 428
 429 @csrf_exempt
 430 @require_valid_user
 431 @check_username
 432 @never_cache
 433 # Workaround for mygpoclient 1.0: It uses "PUT" requests
 434 # instead of "POST" requests for uploading device settings
 435 @allowed_methods(['POST', 'PUT'])
 436 def device(request, username, device_uid):
 437     d = get_device(request.user, device_uid,
 438             request.META.get('HTTP_USER_AGENT', ''))
 439
 440     data = json.loads(request.body)
 441
 442     if 'caption' in data:
 443         if not data['caption']:
 444             return HttpResponseBadRequest('caption must not be empty')
 445         d.name = data['caption']
 446
 447     if 'type' in data:
 448         if not valid_devicetype(data['type']):
 449             return HttpResponseBadRequest('invalid device type %s' % data['type'])
 450         d.type = data['type']
 451
 452
 453     request.user.update_device(d)
 454
 455     return HttpResponse()
 456
 457
 458 def valid_devicetype(type):
 459     for t in DEVICE_TYPES:
 460         if t[0] == type:
 461             return True
 462     return False
 463
 464 def valid_episodeaction(type):
 465     for t in EPISODE_ACTION_TYPES:
 466         if t[0] == type:
 467             return True
 468     return False
 469
 470
 471 @csrf_exempt
 472 @require_valid_user
 473 @check_username
 474 @never_cache
 475 @allowed_methods(['GET'])
 476 def devices(request, username):
 477     devices = filter(lambda d: not d.deleted, request.user.devices)
 478     devices = map(device_data, devices)
 479     return JsonResponse(devices)
 480
 481
 482 def device_data(device):
 483     return dict(
 484         id           = device.uid,
 485         caption      = device.name,
 486         type         = device.type,
 487         subscriptions= len(subscribed_podcast_ids_by_device(device)),
 488     )
 489
 490
 491
 492 def get_podcast_data(podcasts, domain, url):
 493     """ Gets podcast data for a URL from a dict of podcasts """
 494     podcast = podcasts.get(url)
 495     return podcast_data(podcast, domain)
 496
 497
 498 def get_episode_data(podcasts, domain, clean_action_data, include_actions, episode_status):
 499     """ Get episode data for an episode status object """
 500     podcast_id = episode_status.episode.podcast
 501     podcast = podcasts.get(podcast_id, None)
 502     t = episode_data(episode_status.episode, domain, podcast)
 503     t['status'] = episode_status.status
 504
 505     # include latest action (bug 1419)
 506     if include_actions and episode_status.action:
 507         t['action'] = clean_action_data(episode_status.action)
 508
 509     return t
 510
 511
 512
 513 class DeviceUpdates(View):
 514
 515     @method_decorator(csrf_exempt)
 516     @method_decorator(require_valid_user)
 517     @method_decorator(check_username)
 518     @method_decorator(never_cache)
 519     def get(self, request, username, device_uid):
 520         now = datetime.now()
 521         now_ = get_timestamp(now)
 522
 523         try:
 524             device = request.user.get_device_by_uid(device_uid)
 525         except DeviceDoesNotExist as e:
 526             return HttpResponseNotFound(str(e))
 527
 528         since_ = request.GET.get('since', None)
 529         if since_ is None:
 530             return HttpResponseBadRequest('parameter since missing')
 531         try:
 532             since = datetime.fromtimestamp(float(since_))
 533         except ValueError:
 534             return HttpResponseBadRequest("'since' is not a valid timestamp")
 535
 536         include_actions = parse_bool(request.GET.get('include_actions', False))
 537
 538         ret = get_subscription_changes(request.user, device, since, now)
 539         domain = RequestSite(request).domain
 540
 541         subscriptions = list(device.get_subscribed_podcasts())
 542
 543         podcasts = dict( (p.url, p) for p in subscriptions )
 544         prepare_podcast_data = partial(get_podcast_data, podcasts, domain)
 545
 546         ret['add'] = map(prepare_podcast_data, ret['add'])
 547
 548         devices = dict( (dev.id, dev.uid) for dev in request.user.devices )
 549         clean_action_data = partial(clean_episode_action_data,
 550                 user=request.user, devices=devices)
 551
 552         # index subscribed podcasts by their Id for fast access
 553         podcasts = dict( (p.get_id(), p) for p in subscriptions )
 554         prepare_episode_data = partial(get_episode_data, podcasts, domain,
 555                 clean_action_data, include_actions)
 556
 557         episode_updates = self.get_episode_updates(request.user,
 558                 subscriptions, since)
 559         ret['updates'] = map(prepare_episode_data, episode_updates)
 560
 561         return JsonResponse(ret)
 562
 563
 564     def get_episode_updates(self, user, subscribed_podcasts, since,
 565             max_per_podcast=5):
 566         """ Returns the episode updates since the timestamp """
 567
 568         EpisodeStatus = namedtuple('EpisodeStatus', 'episode status action')
 569
 570         episode_status = {}
 571
 572         # get episodes
 573         if gevent:
 574             episode_jobs = [gevent.spawn(episodes_for_podcast, p, since,
 575                     limit=max_per_podcast) for p in subscribed_podcasts]
 576             gevent.joinall(episode_jobs)
 577             episodes = chain.from_iterable(job.get() for job in episode_jobs)
 578
 579         else:
 580             episodes = chain.from_iterable(episodes_for_podcast(p, since,
 581                     limit=max_per_podcast) for p in subscribed_podcasts)
 582
 583
 584         for episode in episodes:
 585             episode_status[episode._id] = EpisodeStatus(episode, 'new', None)
 586
 587
 588         # get episode states
 589         if gevent:
 590             e_action_jobs = [gevent.spawn(get_podcasts_episode_states, p,
 591                     user._id) for p in subscribed_podcasts]
 592             gevent.joinall(e_action_jobs)
 593             e_actions = chain.from_iterable(job.get() for job in e_action_jobs)
 594
 595         else:
 596             e_actions = chain.from_iterable(get_podcasts_episode_states(p,
 597                     user._id) for p in subscribed_podcasts)
 598
 599
 600         for action in e_actions:
 601             e_id = action['episode_id']
 602
 603             if e_id in episode_status:
 604                 episode = episode_status[e_id].episode
 605             else:
 606                 episode = episode_by_id(e_id)
 607
 608             episode_status[e_id] = EpisodeStatus(episode, action['action'],
 609                     action)
 610
 611         return episode_status.itervalues()
 612
 613
 614 @require_valid_user
 615 @check_username
 616 @never_cache
 617 def favorites(request, username):
 618     favorites = favorite_episodes_for_user(request.user)
 619     domain = RequestSite(request).domain
 620     e_data = lambda e: episode_data(e, domain)
 621     ret = map(e_data, favorites)
 622     return JsonResponse(ret)
 623
 624
 625 def sanitize_append(url, obj_type, sanitized_list):
 626     urls = sanitize_url(url, obj_type)
 627     if url != urls:
 628         sanitized_list.append( (url, urls) )
 629     return urls