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