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