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