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 as dsettings
  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.utils import parse_time, format_time, parse_bool, get_timestamp, \
  45     parse_request_body
  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 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 import logging
  66 logger = logging.getLogger(__name__)
  67
  68
  69 # keys that are allowed in episode actions
  70 EPISODE_ACTION_KEYS = ('position', 'episode', 'action', 'device', 'timestamp',
  71                        'started', 'total', 'podcast')
  72
  73
  74 @csrf_exempt
  75 @require_valid_user
  76 @check_username
  77 @never_cache
  78 @allowed_methods(['GET', 'POST'])
  79 def subscriptions(request, username, device_uid):
  80
  81     now = datetime.now()
  82     now_ = get_timestamp(now)
  83
  84     if request.method == 'GET':
  85
  86         try:
  87             device = request.user.get_device_by_uid(device_uid)
  88         except DeviceDoesNotExist as e:
  89             return HttpResponseNotFound(str(e))
  90
  91         since_ = request.GET.get('since', None)
  92         if since_ is None:
  93             return HttpResponseBadRequest('parameter since missing')
  94         try:
  95             since = datetime.fromtimestamp(float(since_))
  96         except ValueError:
  97             return HttpResponseBadRequest('since-value is not a valid timestamp')
  98
  99         changes = get_subscription_changes(request.user, device, since, now)
 100
 101         return JsonResponse(changes)
 102
 103     elif request.method == 'POST':
 104         d = get_device(request.user, device_uid,
 105                 request.META.get('HTTP_USER_AGENT', ''))
 106
 107         if not request.body:
 108             return HttpResponseBadRequest('POST data must not be empty')
 109
 110         try:
 111             actions = parse_request_body(request)
 112         except (JSONDecodeError, UnicodeDecodeError, ValueError) as e:
 113             msg = (u'Could not decode subscription update POST data for ' +
 114                    'user %s: %s') % (username,
 115                    request.body.decode('ascii', errors='replace'))
 116             logger.exception(msg)
 117             return HttpResponseBadRequest(msg)
 118
 119         add = actions['add'] if 'add' in actions else []
 120         rem = actions['remove'] if 'remove' in actions else []
 121
 122         add = filter(None, add)
 123         rem = filter(None, rem)
 124
 125         try:
 126             update_urls = update_subscriptions(request.user, d, add, rem)
 127         except ValueError, e:
 128             return HttpResponseBadRequest(e)
 129
 130         return JsonResponse({
 131             'timestamp': now_,
 132             'update_urls': update_urls,
 133             })
 134
 135
 136 def update_subscriptions(user, device, add, remove):
 137
 138     for a in add:
 139         if a in remove:
 140             raise ValueError('can not add and remove %s at the same time' % a)
 141
 142     add_s = list(sanitize_urls(add, 'podcast'))
 143     rem_s = list(sanitize_urls(remove, 'podcast'))
 144
 145     assert len(add) == len(add_s) and len(remove) == len(rem_s)
 146
 147     updated_urls = filter(lambda (a, b): a != b, zip(add + remove, add_s + rem_s))
 148
 149     add_s = filter(None, add_s)
 150     rem_s = filter(None, rem_s)
 151
 152     # If two different URLs (in add and remove) have
 153     # been sanitized to the same, we ignore the removal
 154     rem_s = filter(lambda x: x not in add_s, rem_s)
 155
 156     subscriber = BulkSubscribe(user, device)
 157
 158     for a in add_s:
 159         subscriber.add_action(a, 'subscribe')
 160
 161     for r in rem_s:
 162         subscriber.add_action(r, 'unsubscribe')
 163
 164     try:
 165         subscriber.execute()
 166     except BulkException as be:
 167         for err in be.errors:
 168             loger.error('Advanced API: %(username)s: Updating subscription for '
 169                     '%(podcast_url)s on %(device_uid)s failed: '
 170                     '%(rerror)s (%(reason)s)'.format(username=user.username,
 171                         podcast_url=err.doc, device_uid=device.uid,
 172                         error=err.error, reason=err.reason)
 173                 )
 174
 175     return updated_urls
 176
 177
 178 def get_subscription_changes(user, device, since, until):
 179     add_urls, rem_urls = device.get_subscription_changes(since, until)
 180     until_ = get_timestamp(until)
 181     return {'add': add_urls, 'remove': rem_urls, 'timestamp': until_}
 182
 183
 184 @csrf_exempt
 185 @require_valid_user
 186 @check_username
 187 @never_cache
 188 @allowed_methods(['GET', 'POST'])
 189 def episodes(request, username, version=1):
 190
 191     version = int(version)
 192     now = datetime.now()
 193     now_ = get_timestamp(now)
 194     ua_string = request.META.get('HTTP_USER_AGENT', '')
 195
 196     if request.method == 'POST':
 197         try:
 198             actions = parse_request_body(request)
 199         except (JSONDecodeError, UnicodeDecodeError, ValueError) as e:
 200             msg = ('Could not decode episode update POST data for ' +
 201                    'user %s: %s') % (username,
 202                    request.body.decode('ascii', errors='replace'))
 203             logger.exception(msg)
 204             return HttpResponseBadRequest(msg)
 205
 206         logger.info('start: user %s: %d actions from %s' % (request.user._id, len(actions), ua_string))
 207
 208         # handle in background
 209         if len(actions) > dsettings.API_ACTIONS_MAX_NONBG:
 210             bg_handler = dsettings.API_ACTIONS_BG_HANDLER
 211             if bg_handler is not None:
 212
 213                 modname, funname = bg_handler.rsplit('.', 1)
 214                 mod = import_module(modname)
 215                 fun = getattr(mod, funname)
 216
 217                 fun(request.user, actions, now, ua_string)
 218
 219                 # TODO: return 202 Accepted
 220                 return JsonResponse({'timestamp': now_, 'update_urls': []})
 221
 222
 223         try:
 224             update_urls = update_episodes(request.user, actions, now, ua_string)
 225         except DeviceUIDException as e:
 226             logger.warn('invalid device UID while uploading episode actions for user %s', username)
 227             return HttpResponseBadRequest(str(e))
 228
 229         except InvalidEpisodeActionAttributes as e:
 230             logger.exception('invalid episode action attributes while uploading episode actions for user %s: %s' % (username,))
 231             return HttpResponseBadRequest(str(e))
 232
 233         logger.info('done:  user %s: %d actions from %s' % (request.user._id, len(actions), ua_string))
 234         return JsonResponse({'timestamp': now_, 'update_urls': update_urls})
 235
 236     elif request.method == 'GET':
 237         podcast_url= request.GET.get('podcast', None)
 238         device_uid = request.GET.get('device', None)
 239         since_     = request.GET.get('since', None)
 240         aggregated = parse_bool(request.GET.get('aggregated', False))
 241
 242         try:
 243             since = int(since_) if since_ else None
 244         except ValueError:
 245             return HttpResponseBadRequest('since-value is not a valid timestamp')
 246
 247         if podcast_url:
 248             podcast = podcast_for_url(podcast_url)
 249             if not podcast:
 250                 raise Http404
 251         else:
 252             podcast = None
 253
 254         if device_uid:
 255
 256             try:
 257                 device = request.user.get_device_by_uid(device_uid)
 258             except DeviceDoesNotExist as e:
 259                 return HttpResponseNotFound(str(e))
 260
 261         else:
 262             device = None
 263
 264         changes = get_episode_changes(request.user, podcast, device, since,
 265                 now_, aggregated, version)
 266
 267         return JsonResponse(changes)
 268
 269
 270
 271 def convert_position(action):
 272     """ convert position parameter for API 1 compatibility """
 273     pos = getattr(action, 'position', None)
 274     if pos is not None:
 275         action.position = format_time(pos)
 276     return action
 277
 278
 279
 280 def get_episode_changes(user, podcast, device, since, until, aggregated, version):
 281
 282     devices = dict( (dev.id, dev.uid) for dev in user.devices )
 283
 284     args = {}
 285     if podcast is not None:
 286         args['podcast_id'] = podcast.get_id()
 287
 288     if device is not None:
 289         args['device_id'] = device.id
 290
 291     actions = get_episode_actions(user._id, since, until, **args)
 292
 293     if version == 1:
 294         actions = imap(convert_position, actions)
 295
 296     clean_data = partial(clean_episode_action_data,
 297             user=user, devices=devices)
 298
 299     actions = map(clean_data, actions)
 300     actions = filter(None, actions)
 301
 302     if aggregated:
 303         actions = dict( (a['episode'], a) for a in actions ).values()
 304
 305     return {'actions': actions, 'timestamp': until}
 306
 307
 308
 309
 310 def clean_episode_action_data(action, user, devices):
 311
 312     if None in (action.get('podcast', None), action.get('episode', None)):
 313         return None
 314
 315     if 'device_id' in action:
 316         device_id = action['device_id']
 317         device_uid = devices.get(device_id)
 318         if device_uid:
 319             action['device'] = device_uid
 320
 321         del action['device_id']
 322
 323     # remove superfluous keys
 324     for x in action.keys():
 325         if x not in EPISODE_ACTION_KEYS:
 326             del action[x]
 327
 328     # set missing keys to None
 329     for x in EPISODE_ACTION_KEYS:
 330         if x not in action:
 331             action[x] = None
 332
 333     if action['action'] != 'play':
 334         if 'position' in action:
 335             del action['position']
 336
 337         if 'total' in action:
 338             del action['total']
 339
 340         if 'started' in action:
 341             del action['started']
 342
 343         if 'playmark' in action:
 344             del action['playmark']
 345
 346     else:
 347         action['position'] = action.get('position', False) or 0
 348
 349     return action
 350
 351
 352
 353
 354
 355 def update_episodes(user, actions, now, ua_string):
 356     update_urls = []
 357
 358     grouped_actions = defaultdict(list)
 359
 360     # group all actions by their episode
 361     for action in actions:
 362
 363         podcast_url = action['podcast']
 364         podcast_url = sanitize_append(podcast_url, 'podcast', update_urls)
 365         if podcast_url == '':
 366             continue
 367
 368         episode_url = action['episode']
 369         episode_url = sanitize_append(episode_url, 'episode', update_urls)
 370         if episode_url == '':
 371             continue
 372
 373         act = parse_episode_action(action, user, update_urls, now, ua_string)
 374         grouped_actions[ (podcast_url, episode_url) ].append(act)
 375
 376
 377     auto_flattr_episodes = []
 378
 379     # Prepare the updates for each episode state
 380     obj_funs = []
 381
 382     for (p_url, e_url), action_list in grouped_actions.iteritems():
 383         episode_state = episode_state_for_ref_urls(user, p_url, e_url)
 384
 385         if any(a['action'] == 'play' for a in actions):
 386             auto_flattr_episodes.append(episode_state.episode)
 387
 388         fun = partial(update_episode_actions, action_list=action_list)
 389         obj_funs.append( (episode_state, fun) )
 390
 391     bulk_save_retry(obj_funs)
 392
 393     if user.get_wksetting(FLATTR_AUTO):
 394         for episode_id in auto_flattr_episodes:
 395             auto_flattr_episode.delay(user, episode_id)
 396
 397     return update_urls
 398
 399
 400 def update_episode_actions(episode_state, action_list):
 401     """ Adds actions to the episode state and saves if necessary """
 402
 403     len1 = len(episode_state.actions)
 404     episode_state.add_actions(action_list)
 405
 406     if len(episode_state.actions) == len1:
 407         return None
 408
 409     return episode_state
 410
 411
 412
 413 def parse_episode_action(action, user, update_urls, now, ua_string):
 414     action_str  = action.get('action', None)
 415     if not valid_episodeaction(action_str):
 416         raise Exception('invalid action %s' % action_str)
 417
 418     new_action = EpisodeAction()
 419
 420     new_action.action = action['action']
 421
 422     if action.get('device', False):
 423         device = get_device(user, action['device'], ua_string)
 424         new_action.device = device.id
 425
 426     if action.get('timestamp', False):
 427         new_action.timestamp = dateutil.parser.parse(action['timestamp'])
 428     else:
 429         new_action.timestamp = now
 430     new_action.timestamp = new_action.timestamp.replace(microsecond=0)
 431
 432     new_action.upload_timestamp = get_timestamp(now)
 433
 434     new_action.started = action.get('started', None)
 435     new_action.playmark = action.get('position', None)
 436     new_action.total = action.get('total', None)
 437
 438     return new_action
 439
 440
 441 @csrf_exempt
 442 @require_valid_user
 443 @check_username
 444 @never_cache
 445 # Workaround for mygpoclient 1.0: It uses "PUT" requests
 446 # instead of "POST" requests for uploading device settings
 447 @allowed_methods(['POST', 'PUT'])
 448 def device(request, username, device_uid):
 449     d = get_device(request.user, device_uid,
 450             request.META.get('HTTP_USER_AGENT', ''))
 451
 452     try:
 453         data = parse_request_body(request)
 454     except (JSONDecodeError, UnicodeDecodeError, ValueError) as e:
 455         msg = ('Could not decode device update POST data for ' +
 456                'user %s: %s') % (username,
 457                request.body.decode('ascii', errors='replace'))
 458         logger.exception(msg)
 459         return HttpResponseBadRequest(msg)
 460
 461     if 'caption' in data:
 462         if not data['caption']:
 463             return HttpResponseBadRequest('caption must not be empty')
 464         d.name = data['caption']
 465
 466     if 'type' in data:
 467         if not valid_devicetype(data['type']):
 468             return HttpResponseBadRequest('invalid device type %s' % data['type'])
 469         d.type = data['type']
 470
 471
 472     request.user.update_device(d)
 473
 474     return HttpResponse()
 475
 476
 477 def valid_devicetype(type):
 478     for t in DEVICE_TYPES:
 479         if t[0] == type:
 480             return True
 481     return False
 482
 483 def valid_episodeaction(type):
 484     for t in EPISODE_ACTION_TYPES:
 485         if t[0] == type:
 486             return True
 487     return False
 488
 489
 490 @csrf_exempt
 491 @require_valid_user
 492 @check_username
 493 @never_cache
 494 @allowed_methods(['GET'])
 495 def devices(request, username):
 496     devices = filter(lambda d: not d.deleted, request.user.devices)
 497     devices = map(device_data, devices)
 498     return JsonResponse(devices)
 499
 500
 501 def device_data(device):
 502     return dict(
 503         id           = device.uid,
 504         caption      = device.name,
 505         type         = device.type,
 506         subscriptions= len(subscribed_podcast_ids_by_device(device)),
 507     )
 508
 509
 510
 511 def get_podcast_data(podcasts, domain, url):
 512     """ Gets podcast data for a URL from a dict of podcasts """
 513     podcast = podcasts.get(url)
 514     return podcast_data(podcast, domain)
 515
 516
 517 def get_episode_data(podcasts, domain, clean_action_data, include_actions, episode_status):
 518     """ Get episode data for an episode status object """
 519     podcast_id = episode_status.episode.podcast
 520     podcast = podcasts.get(podcast_id, None)
 521     t = episode_data(episode_status.episode, domain, podcast)
 522     t['status'] = episode_status.status
 523
 524     # include latest action (bug 1419)
 525     if include_actions and episode_status.action:
 526         t['action'] = clean_action_data(episode_status.action)
 527
 528     return t
 529
 530
 531
 532 class DeviceUpdates(View):
 533
 534     @method_decorator(csrf_exempt)
 535     @method_decorator(require_valid_user)
 536     @method_decorator(check_username)
 537     @method_decorator(never_cache)
 538     def get(self, request, username, device_uid):
 539         now = datetime.now()
 540         now_ = get_timestamp(now)
 541
 542         try:
 543             device = request.user.get_device_by_uid(device_uid)
 544         except DeviceDoesNotExist as e:
 545             return HttpResponseNotFound(str(e))
 546
 547         since_ = request.GET.get('since', None)
 548         if since_ is None:
 549             return HttpResponseBadRequest('parameter since missing')
 550         try:
 551             since = datetime.fromtimestamp(float(since_))
 552         except ValueError:
 553             return HttpResponseBadRequest("'since' is not a valid timestamp")
 554
 555         include_actions = parse_bool(request.GET.get('include_actions', False))
 556
 557         ret = get_subscription_changes(request.user, device, since, now)
 558         domain = RequestSite(request).domain
 559
 560         subscriptions = list(device.get_subscribed_podcasts())
 561
 562         podcasts = dict( (p.url, p) for p in subscriptions )
 563         prepare_podcast_data = partial(get_podcast_data, podcasts, domain)
 564
 565         ret['add'] = map(prepare_podcast_data, ret['add'])
 566
 567         devices = dict( (dev.id, dev.uid) for dev in request.user.devices )
 568         clean_action_data = partial(clean_episode_action_data,
 569                 user=request.user, devices=devices)
 570
 571         # index subscribed podcasts by their Id for fast access
 572         podcasts = dict( (p.get_id(), p) for p in subscriptions )
 573         prepare_episode_data = partial(get_episode_data, podcasts, domain,
 574                 clean_action_data, include_actions)
 575
 576         episode_updates = self.get_episode_updates(request.user,
 577                 subscriptions, since)
 578         ret['updates'] = map(prepare_episode_data, episode_updates)
 579
 580         return JsonResponse(ret)
 581
 582
 583     def get_episode_updates(self, user, subscribed_podcasts, since,
 584             max_per_podcast=5):
 585         """ Returns the episode updates since the timestamp """
 586
 587         EpisodeStatus = namedtuple('EpisodeStatus', 'episode status action')
 588
 589         episode_status = {}
 590
 591         # get episodes
 592         if gevent:
 593             episode_jobs = [gevent.spawn(episodes_for_podcast, p, since,
 594                     limit=max_per_podcast) for p in subscribed_podcasts]
 595             gevent.joinall(episode_jobs)
 596             episodes = chain.from_iterable(job.get() for job in episode_jobs)
 597
 598         else:
 599             episodes = chain.from_iterable(episodes_for_podcast(p, since,
 600                     limit=max_per_podcast) for p in subscribed_podcasts)
 601
 602
 603         for episode in episodes:
 604             episode_status[episode._id] = EpisodeStatus(episode, 'new', None)
 605
 606
 607         # get episode states
 608         if gevent:
 609             e_action_jobs = [gevent.spawn(get_podcasts_episode_states, p,
 610                     user._id) for p in subscribed_podcasts]
 611             gevent.joinall(e_action_jobs)
 612             e_actions = chain.from_iterable(job.get() for job in e_action_jobs)
 613
 614         else:
 615             e_actions = chain.from_iterable(get_podcasts_episode_states(p,
 616                     user._id) for p in subscribed_podcasts)
 617
 618
 619         for action in e_actions:
 620             e_id = action['episode_id']
 621
 622             if e_id in episode_status:
 623                 episode = episode_status[e_id].episode
 624             else:
 625                 episode = episode_by_id(e_id)
 626
 627             episode_status[e_id] = EpisodeStatus(episode, action['action'],
 628                     action)
 629
 630         return episode_status.itervalues()
 631
 632
 633 @require_valid_user
 634 @check_username
 635 @never_cache
 636 def favorites(request, username):
 637     favorites = favorite_episodes_for_user(request.user)
 638     domain = RequestSite(request).domain
 639     e_data = lambda e: episode_data(e, domain)
 640     ret = map(e_data, favorites)
 641     return JsonResponse(ret)
 642
 643
 644 def sanitize_append(url, obj_type, sanitized_list):
 645     urls = sanitize_url(url, obj_type)
 646     if url != urls:
 647         sanitized_list.append( (url, urls) )
 648     return urls