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 from mygpo.decorators import allowed_methods, repeat_on_conflict
  46 from mygpo.core import models
  47 from mygpo.core.models import SanitizingRule, Podcast
  48 from mygpo.core.tasks import auto_flattr_episode
  49 from mygpo.users.models import PodcastUserState, EpisodeAction, \
  50      EpisodeUserState, DeviceDoesNotExist, DeviceUIDException, \
  51      InvalidEpisodeActionAttributes
  52 from mygpo.users.settings import FLATTR_AUTO
  53 from mygpo.core.json import json, JSONDecodeError
  54 from mygpo.api.basic_auth import require_valid_user, check_username
  55 from mygpo.db.couchdb import BulkException, bulk_save_retry
  56 from mygpo.db.couchdb.episode import episode_by_id, \
  57          favorite_episodes_for_user, episodes_for_podcast
  58 from mygpo.db.couchdb.podcast import podcast_for_url
  59 from mygpo.db.couchdb.podcast_state import subscribed_podcast_ids_by_device
  60 from mygpo.db.couchdb.episode_state import get_podcasts_episode_states, \
  61          episode_state_for_ref_urls, get_episode_actions
  62
  63
  64 import logging
  65 logger = logging.getLogger(__name__)
  66
  67
  68 # keys that are allowed in episode actions
  69 EPISODE_ACTION_KEYS = ('position', 'episode', 'action', 'device', 'timestamp',
  70                        'started', 'total', 'podcast')
  71
  72
  73 @csrf_exempt
  74 @require_valid_user
  75 @check_username
  76 @never_cache
  77 @allowed_methods(['GET', 'POST'])
  78 def subscriptions(request, username, device_uid):
  79
  80     now = datetime.now()
  81     now_ = get_timestamp(now)
  82
  83     if request.method == 'GET':
  84
  85         try:
  86             device = request.user.get_device_by_uid(device_uid)
  87         except DeviceDoesNotExist as e:
  88             return HttpResponseNotFound(str(e))
  89
  90         since_ = request.GET.get('since', None)
  91         if since_ is None:
  92             return HttpResponseBadRequest('parameter since missing')
  93         try:
  94             since = datetime.fromtimestamp(float(since_))
  95         except ValueError:
  96             return HttpResponseBadRequest('since-value is not a valid timestamp')
  97
  98         changes = get_subscription_changes(request.user, device, since, now)
  99
 100         return JsonResponse(changes)
 101
 102     elif request.method == 'POST':
 103         d = get_device(request.user, device_uid,
 104                 request.META.get('HTTP_USER_AGENT', ''))
 105
 106         if not request.body:
 107             return HttpResponseBadRequest('POST data must not be empty')
 108
 109         try:
 110             actions = json.loads(request.body)
 111         except (JSONDecodeError, UnicodeDecodeError, ValueError) as e:
 112             msg = ('Could not decode subscription update POST data for ' +
 113                    'user %s: %s') % (username, request.body)
 114             logger.exception(msg)
 115             return HttpResponseBadRequest(msg)
 116
 117         add = actions['add'] if 'add' in actions else []
 118         rem = actions['remove'] if 'remove' in actions else []
 119
 120         add = filter(None, add)
 121         rem = filter(None, rem)
 122
 123         try:
 124             update_urls = update_subscriptions(request.user, d, add, rem)
 125         except ValueError, e:
 126             return HttpResponseBadRequest(e)
 127
 128         return JsonResponse({
 129             'timestamp': now_,
 130             'update_urls': update_urls,
 131             })
 132
 133
 134 def update_subscriptions(user, device, add, remove):
 135
 136     for a in add:
 137         if a in remove:
 138             raise ValueError('can not add and remove %s at the same time' % a)
 139
 140     add_s = list(sanitize_urls(add, 'podcast'))
 141     rem_s = list(sanitize_urls(remove, 'podcast'))
 142
 143     assert len(add) == len(add_s) and len(remove) == len(rem_s)
 144
 145     updated_urls = filter(lambda (a, b): a != b, zip(add + remove, add_s + rem_s))
 146
 147     add_s = filter(None, add_s)
 148     rem_s = filter(None, rem_s)
 149
 150     # If two different URLs (in add and remove) have
 151     # been sanitized to the same, we ignore the removal
 152     rem_s = filter(lambda x: x not in add_s, rem_s)
 153
 154     subscriber = BulkSubscribe(user, device)
 155
 156     for a in add_s:
 157         subscriber.add_action(a, 'subscribe')
 158
 159     for r in rem_s:
 160         subscriber.add_action(r, 'unsubscribe')
 161
 162     try:
 163         subscriber.execute()
 164     except BulkException as be:
 165         for err in be.errors:
 166             loger.error('Advanced API: %(username)s: Updating subscription for '
 167                     '%(podcast_url)s on %(device_uid)s failed: '
 168                     '%(rerror)s (%(reason)s)'.format(username=user.username,
 169                         podcast_url=err.doc, device_uid=device.uid,
 170                         error=err.error, reason=err.reason)
 171                 )
 172
 173     return updated_urls
 174
 175
 176 def get_subscription_changes(user, device, since, until):
 177     add_urls, rem_urls = device.get_subscription_changes(since, until)
 178     until_ = get_timestamp(until)
 179     return {'add': add_urls, 'remove': rem_urls, 'timestamp': until_}
 180
 181
 182 @csrf_exempt
 183 @require_valid_user
 184 @check_username
 185 @never_cache
 186 @allowed_methods(['GET', 'POST'])
 187 def episodes(request, username, version=1):
 188
 189     version = int(version)
 190     now = datetime.now()
 191     now_ = get_timestamp(now)
 192     ua_string = request.META.get('HTTP_USER_AGENT', '')
 193
 194     if request.method == 'POST':
 195         try:
 196             actions = json.loads(request.body)
 197         except (JSONDecodeError, UnicodeDecodeError, ValueError) as e:
 198             msg = ('Could not decode episode update POST data for ' +
 199                    'user %s: %s') % (username, request.body)
 200             logger.exception(msg)
 201             return HttpResponseBadRequest(msg)
 202
 203         logger.info('start: user %s: %d actions from %s' % (request.user._id, len(actions), ua_string))
 204
 205         # handle in background
 206         if len(actions) > dsettings.API_ACTIONS_MAX_NONBG:
 207             bg_handler = dsettings.API_ACTIONS_BG_HANDLER
 208             if bg_handler is not None:
 209
 210                 modname, funname = bg_handler.rsplit('.', 1)
 211                 mod = import_module(modname)
 212                 fun = getattr(mod, funname)
 213
 214                 fun(request.user, actions, now, ua_string)
 215
 216                 # TODO: return 202 Accepted
 217                 return JsonResponse({'timestamp': now_, 'update_urls': []})
 218
 219
 220         try:
 221             update_urls = update_episodes(request.user, actions, now, ua_string)
 222         except DeviceUIDException as e:
 223             logger.exception('invalid device UID while uploading episode actions for user %s' % (username,))
 224             return HttpResponseBadRequest(str(e))
 225
 226         except InvalidEpisodeActionAttributes as e:
 227             logger.exception('invalid episode action attributes while uploading episode actions for user %s: %s' % (username,))
 228             return HttpResponseBadRequest(str(e))
 229
 230         logger.info('done:  user %s: %d actions from %s' % (request.user._id, len(actions), ua_string))
 231         return JsonResponse({'timestamp': now_, 'update_urls': update_urls})
 232
 233     elif request.method == 'GET':
 234         podcast_url= request.GET.get('podcast', None)
 235         device_uid = request.GET.get('device', None)
 236         since_     = request.GET.get('since', None)
 237         aggregated = parse_bool(request.GET.get('aggregated', False))
 238
 239         try:
 240             since = int(since_) if since_ else None
 241         except ValueError:
 242             return HttpResponseBadRequest('since-value is not a valid timestamp')
 243
 244         if podcast_url:
 245             podcast = podcast_for_url(podcast_url)
 246             if not podcast:
 247                 raise Http404
 248         else:
 249             podcast = None
 250
 251         if device_uid:
 252
 253             try:
 254                 device = request.user.get_device_by_uid(device_uid)
 255             except DeviceDoesNotExist as e:
 256                 return HttpResponseNotFound(str(e))
 257
 258         else:
 259             device = None
 260
 261         changes = get_episode_changes(request.user, podcast, device, since,
 262                 now_, aggregated, version)
 263
 264         return JsonResponse(changes)
 265
 266
 267
 268 def convert_position(action):
 269     """ convert position parameter for API 1 compatibility """
 270     pos = getattr(action, 'position', None)
 271     if pos is not None:
 272         action.position = format_time(pos)
 273     return action
 274
 275
 276
 277 def get_episode_changes(user, podcast, device, since, until, aggregated, version):
 278
 279     devices = dict( (dev.id, dev.uid) for dev in user.devices )
 280
 281     args = {}
 282     if podcast is not None:
 283         args['podcast_id'] = podcast.get_id()
 284
 285     if device is not None:
 286         args['device_id'] = device.id
 287
 288     actions = get_episode_actions(user._id, since, until, **args)
 289
 290     if version == 1:
 291         actions = imap(convert_position, actions)
 292
 293     clean_data = partial(clean_episode_action_data,
 294             user=user, devices=devices)
 295
 296     actions = map(clean_data, actions)
 297     actions = filter(None, actions)
 298
 299     if aggregated:
 300         actions = dict( (a['episode'], a) for a in actions ).values()
 301
 302     return {'actions': actions, 'timestamp': until}
 303
 304
 305
 306
 307 def clean_episode_action_data(action, user, devices):
 308
 309     if None in (action.get('podcast', None), action.get('episode', None)):
 310         return None
 311
 312     if 'device_id' in action:
 313         device_id = action['device_id']
 314         device_uid = devices.get(device_id)
 315         if device_uid:
 316             action['device'] = device_uid
 317
 318         del action['device_id']
 319
 320     # remove superfluous keys
 321     for x in action.keys():
 322         if x not in EPISODE_ACTION_KEYS:
 323             del action[x]
 324
 325     # set missing keys to None
 326     for x in EPISODE_ACTION_KEYS:
 327         if x not in action:
 328             action[x] = None
 329
 330     if action['action'] != 'play':
 331         if 'position' in action:
 332             del action['position']
 333
 334         if 'total' in action:
 335             del action['total']
 336
 337         if 'started' in action:
 338             del action['started']
 339
 340         if 'playmark' in action:
 341             del action['playmark']
 342
 343     else:
 344         action['position'] = action.get('position', False) or 0
 345
 346     return action
 347
 348
 349
 350
 351
 352 def update_episodes(user, actions, now, ua_string):
 353     update_urls = []
 354
 355     grouped_actions = defaultdict(list)
 356
 357     # group all actions by their episode
 358     for action in actions:
 359
 360         podcast_url = action['podcast']
 361         podcast_url = sanitize_append(podcast_url, 'podcast', update_urls)
 362         if podcast_url == '':
 363             continue
 364
 365         episode_url = action['episode']
 366         episode_url = sanitize_append(episode_url, 'episode', update_urls)
 367         if episode_url == '':
 368             continue
 369
 370         act = parse_episode_action(action, user, update_urls, now, ua_string)
 371         grouped_actions[ (podcast_url, episode_url) ].append(act)
 372
 373
 374     auto_flattr_episodes = []
 375
 376     # Prepare the updates for each episode state
 377     obj_funs = []
 378
 379     for (p_url, e_url), action_list in grouped_actions.iteritems():
 380         episode_state = episode_state_for_ref_urls(user, p_url, e_url)
 381
 382         if any(a['action'] == 'play' for a in actions):
 383             auto_flattr_episodes.append(episode_state.episode)
 384
 385         fun = partial(update_episode_actions, action_list=action_list)
 386         obj_funs.append( (episode_state, fun) )
 387
 388     bulk_save_retry(obj_funs)
 389
 390     if user.get_wksetting(FLATTR_AUTO):
 391         for episode_id in auto_flattr_episodes:
 392             auto_flattr_episode.delay(user, episode_id)
 393
 394     return update_urls
 395
 396
 397 def update_episode_actions(episode_state, action_list):
 398     """ Adds actions to the episode state and saves if necessary """
 399
 400     len1 = len(episode_state.actions)
 401     episode_state.add_actions(action_list)
 402
 403     if len(episode_state.actions) == len1:
 404         return None
 405
 406     return episode_state
 407
 408
 409
 410 def parse_episode_action(action, user, update_urls, now, ua_string):
 411     action_str  = action.get('action', None)
 412     if not valid_episodeaction(action_str):
 413         raise Exception('invalid action %s' % action_str)
 414
 415     new_action = EpisodeAction()
 416
 417     new_action.action = action['action']
 418
 419     if action.get('device', False):
 420         device = get_device(user, action['device'], ua_string)
 421         new_action.device = device.id
 422
 423     if action.get('timestamp', False):
 424         new_action.timestamp = dateutil.parser.parse(action['timestamp'])
 425     else:
 426         new_action.timestamp = now
 427     new_action.timestamp = new_action.timestamp.replace(microsecond=0)
 428
 429     new_action.upload_timestamp = get_timestamp(now)
 430
 431     new_action.started = action.get('started', None)
 432     new_action.playmark = action.get('position', None)
 433     new_action.total = action.get('total', None)
 434
 435     return new_action
 436
 437
 438 @csrf_exempt
 439 @require_valid_user
 440 @check_username
 441 @never_cache
 442 # Workaround for mygpoclient 1.0: It uses "PUT" requests
 443 # instead of "POST" requests for uploading device settings
 444 @allowed_methods(['POST', 'PUT'])
 445 def device(request, username, device_uid):
 446     d = get_device(request.user, device_uid,
 447             request.META.get('HTTP_USER_AGENT', ''))
 448
 449     try:
 450         data = json.loads(request.body)
 451     except (JSONDecodeError, UnicodeDecodeError, ValueError) as e:
 452         msg = ('Could not decode device update POST data for ' +
 453                'user %s: %s') % (username, request.body)
 454         logger.exception(msg)
 455         return HttpResponseBadRequest(msg)
 456
 457     if 'caption' in data:
 458         if not data['caption']:
 459             return HttpResponseBadRequest('caption must not be empty')
 460         d.name = data['caption']
 461
 462     if 'type' in data:
 463         if not valid_devicetype(data['type']):
 464             return HttpResponseBadRequest('invalid device type %s' % data['type'])
 465         d.type = data['type']
 466
 467
 468     request.user.update_device(d)
 469
 470     return HttpResponse()
 471
 472
 473 def valid_devicetype(type):
 474     for t in DEVICE_TYPES:
 475         if t[0] == type:
 476             return True
 477     return False
 478
 479 def valid_episodeaction(type):
 480     for t in EPISODE_ACTION_TYPES:
 481         if t[0] == type:
 482             return True
 483     return False
 484
 485
 486 @csrf_exempt
 487 @require_valid_user
 488 @check_username
 489 @never_cache
 490 @allowed_methods(['GET'])
 491 def devices(request, username):
 492     devices = filter(lambda d: not d.deleted, request.user.devices)
 493     devices = map(device_data, devices)
 494     return JsonResponse(devices)
 495
 496
 497 def device_data(device):
 498     return dict(
 499         id           = device.uid,
 500         caption      = device.name,
 501         type         = device.type,
 502         subscriptions= len(subscribed_podcast_ids_by_device(device)),
 503     )
 504
 505
 506
 507 def get_podcast_data(podcasts, domain, url):
 508     """ Gets podcast data for a URL from a dict of podcasts """
 509     podcast = podcasts.get(url)
 510     return podcast_data(podcast, domain)
 511
 512
 513 def get_episode_data(podcasts, domain, clean_action_data, include_actions, episode_status):
 514     """ Get episode data for an episode status object """
 515     podcast_id = episode_status.episode.podcast
 516     podcast = podcasts.get(podcast_id, None)
 517     t = episode_data(episode_status.episode, domain, podcast)
 518     t['status'] = episode_status.status
 519
 520     # include latest action (bug 1419)
 521     if include_actions and episode_status.action:
 522         t['action'] = clean_action_data(episode_status.action)
 523
 524     return t
 525
 526
 527
 528 class DeviceUpdates(View):
 529
 530     @method_decorator(csrf_exempt)
 531     @method_decorator(require_valid_user)
 532     @method_decorator(check_username)
 533     @method_decorator(never_cache)
 534     def get(self, request, username, device_uid):
 535         now = datetime.now()
 536         now_ = get_timestamp(now)
 537
 538         try:
 539             device = request.user.get_device_by_uid(device_uid)
 540         except DeviceDoesNotExist as e:
 541             return HttpResponseNotFound(str(e))
 542
 543         since_ = request.GET.get('since', None)
 544         if since_ is None:
 545             return HttpResponseBadRequest('parameter since missing')
 546         try:
 547             since = datetime.fromtimestamp(float(since_))
 548         except ValueError:
 549             return HttpResponseBadRequest("'since' is not a valid timestamp")
 550
 551         include_actions = parse_bool(request.GET.get('include_actions', False))
 552
 553         ret = get_subscription_changes(request.user, device, since, now)
 554         domain = RequestSite(request).domain
 555
 556         subscriptions = list(device.get_subscribed_podcasts())
 557
 558         podcasts = dict( (p.url, p) for p in subscriptions )
 559         prepare_podcast_data = partial(get_podcast_data, podcasts, domain)
 560
 561         ret['add'] = map(prepare_podcast_data, ret['add'])
 562
 563         devices = dict( (dev.id, dev.uid) for dev in request.user.devices )
 564         clean_action_data = partial(clean_episode_action_data,
 565                 user=request.user, devices=devices)
 566
 567         # index subscribed podcasts by their Id for fast access
 568         podcasts = dict( (p.get_id(), p) for p in subscriptions )
 569         prepare_episode_data = partial(get_episode_data, podcasts, domain,
 570                 clean_action_data, include_actions)
 571
 572         episode_updates = self.get_episode_updates(request.user,
 573                 subscriptions, since)
 574         ret['updates'] = map(prepare_episode_data, episode_updates)
 575
 576         return JsonResponse(ret)
 577
 578
 579     def get_episode_updates(self, user, subscribed_podcasts, since,
 580             max_per_podcast=5):
 581         """ Returns the episode updates since the timestamp """
 582
 583         EpisodeStatus = namedtuple('EpisodeStatus', 'episode status action')
 584
 585         episode_status = {}
 586
 587         # get episodes
 588         if gevent:
 589             episode_jobs = [gevent.spawn(episodes_for_podcast, p, since,
 590                     limit=max_per_podcast) for p in subscribed_podcasts]
 591             gevent.joinall(episode_jobs)
 592             episodes = chain.from_iterable(job.get() for job in episode_jobs)
 593
 594         else:
 595             episodes = chain.from_iterable(episodes_for_podcast(p, since,
 596                     limit=max_per_podcast) for p in subscribed_podcasts)
 597
 598
 599         for episode in episodes:
 600             episode_status[episode._id] = EpisodeStatus(episode, 'new', None)
 601
 602
 603         # get episode states
 604         if gevent:
 605             e_action_jobs = [gevent.spawn(get_podcasts_episode_states, p,
 606                     user._id) for p in subscribed_podcasts]
 607             gevent.joinall(e_action_jobs)
 608             e_actions = chain.from_iterable(job.get() for job in e_action_jobs)
 609
 610         else:
 611             e_actions = chain.from_iterable(get_podcasts_episode_states(p,
 612                     user._id) for p in subscribed_podcasts)
 613
 614
 615         for action in e_actions:
 616             e_id = action['episode_id']
 617
 618             if e_id in episode_status:
 619                 episode = episode_status[e_id].episode
 620             else:
 621                 episode = episode_by_id(e_id)
 622
 623             episode_status[e_id] = EpisodeStatus(episode, action['action'],
 624                     action)
 625
 626         return episode_status.itervalues()
 627
 628
 629 @require_valid_user
 630 @check_username
 631 @never_cache
 632 def favorites(request, username):
 633     favorites = favorite_episodes_for_user(request.user)
 634     domain = RequestSite(request).domain
 635     e_data = lambda e: episode_data(e, domain)
 636     ret = map(e_data, favorites)
 637     return JsonResponse(ret)
 638
 639
 640 def sanitize_append(url, obj_type, sanitized_list):
 641     urls = sanitize_url(url, obj_type)
 642     if url != urls:
 643         sanitized_list.append( (url, urls) )
 644     return urls