receipts.py 17 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516
  1. # -*- coding: utf-8 -*-
  2. # Copyright 2014-2016 OpenMarket Ltd
  3. # Copyright 2018 New Vector Ltd
  4. #
  5. # Licensed under the Apache License, Version 2.0 (the "License");
  6. # you may not use this file except in compliance with the License.
  7. # You may obtain a copy of the License at
  8. #
  9. # http://www.apache.org/licenses/LICENSE-2.0
  10. #
  11. # Unless required by applicable law or agreed to in writing, software
  12. # distributed under the License is distributed on an "AS IS" BASIS,
  13. # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  14. # See the License for the specific language governing permissions and
  15. # limitations under the License.
  16. from ._base import SQLBaseStore
  17. from .util.id_generators import StreamIdGenerator
  18. from synapse.util.caches.descriptors import cachedInlineCallbacks, cachedList, cached
  19. from synapse.util.caches.stream_change_cache import StreamChangeCache
  20. from twisted.internet import defer
  21. import abc
  22. import logging
  23. import simplejson as json
  24. logger = logging.getLogger(__name__)
  25. class ReceiptsWorkerStore(SQLBaseStore):
  26. """This is an abstract base class where subclasses must implement
  27. `get_max_receipt_stream_id` which can be called in the initializer.
  28. """
  29. # This ABCMeta metaclass ensures that we cannot be instantiated without
  30. # the abstract methods being implemented.
  31. __metaclass__ = abc.ABCMeta
  32. def __init__(self, db_conn, hs):
  33. super(ReceiptsWorkerStore, self).__init__(db_conn, hs)
  34. self._receipts_stream_cache = StreamChangeCache(
  35. "ReceiptsRoomChangeCache", self.get_max_receipt_stream_id()
  36. )
  37. @abc.abstractmethod
  38. def get_max_receipt_stream_id(self):
  39. """Get the current max stream ID for receipts stream
  40. Returns:
  41. int
  42. """
  43. raise NotImplementedError()
  44. @cachedInlineCallbacks()
  45. def get_users_with_read_receipts_in_room(self, room_id):
  46. receipts = yield self.get_receipts_for_room(room_id, "m.read")
  47. defer.returnValue(set(r['user_id'] for r in receipts))
  48. @cached(num_args=2)
  49. def get_receipts_for_room(self, room_id, receipt_type):
  50. return self._simple_select_list(
  51. table="receipts_linearized",
  52. keyvalues={
  53. "room_id": room_id,
  54. "receipt_type": receipt_type,
  55. },
  56. retcols=("user_id", "event_id"),
  57. desc="get_receipts_for_room",
  58. )
  59. @cached(num_args=3)
  60. def get_last_receipt_event_id_for_user(self, user_id, room_id, receipt_type):
  61. return self._simple_select_one_onecol(
  62. table="receipts_linearized",
  63. keyvalues={
  64. "room_id": room_id,
  65. "receipt_type": receipt_type,
  66. "user_id": user_id
  67. },
  68. retcol="event_id",
  69. desc="get_own_receipt_for_user",
  70. allow_none=True,
  71. )
  72. @cachedInlineCallbacks(num_args=2)
  73. def get_receipts_for_user(self, user_id, receipt_type):
  74. rows = yield self._simple_select_list(
  75. table="receipts_linearized",
  76. keyvalues={
  77. "user_id": user_id,
  78. "receipt_type": receipt_type,
  79. },
  80. retcols=("room_id", "event_id"),
  81. desc="get_receipts_for_user",
  82. )
  83. defer.returnValue({row["room_id"]: row["event_id"] for row in rows})
  84. @defer.inlineCallbacks
  85. def get_receipts_for_user_with_orderings(self, user_id, receipt_type):
  86. def f(txn):
  87. sql = (
  88. "SELECT rl.room_id, rl.event_id,"
  89. " e.topological_ordering, e.stream_ordering"
  90. " FROM receipts_linearized AS rl"
  91. " INNER JOIN events AS e USING (room_id, event_id)"
  92. " WHERE rl.room_id = e.room_id"
  93. " AND rl.event_id = e.event_id"
  94. " AND user_id = ?"
  95. )
  96. txn.execute(sql, (user_id,))
  97. return txn.fetchall()
  98. rows = yield self.runInteraction(
  99. "get_receipts_for_user_with_orderings", f
  100. )
  101. defer.returnValue({
  102. row[0]: {
  103. "event_id": row[1],
  104. "topological_ordering": row[2],
  105. "stream_ordering": row[3],
  106. } for row in rows
  107. })
  108. @defer.inlineCallbacks
  109. def get_linearized_receipts_for_rooms(self, room_ids, to_key, from_key=None):
  110. """Get receipts for multiple rooms for sending to clients.
  111. Args:
  112. room_ids (list): List of room_ids.
  113. to_key (int): Max stream id to fetch receipts upto.
  114. from_key (int): Min stream id to fetch receipts from. None fetches
  115. from the start.
  116. Returns:
  117. list: A list of receipts.
  118. """
  119. room_ids = set(room_ids)
  120. if from_key:
  121. room_ids = yield self._receipts_stream_cache.get_entities_changed(
  122. room_ids, from_key
  123. )
  124. results = yield self._get_linearized_receipts_for_rooms(
  125. room_ids, to_key, from_key=from_key
  126. )
  127. defer.returnValue([ev for res in results.values() for ev in res])
  128. @cachedInlineCallbacks(num_args=3, tree=True)
  129. def get_linearized_receipts_for_room(self, room_id, to_key, from_key=None):
  130. """Get receipts for a single room for sending to clients.
  131. Args:
  132. room_ids (str): The room id.
  133. to_key (int): Max stream id to fetch receipts upto.
  134. from_key (int): Min stream id to fetch receipts from. None fetches
  135. from the start.
  136. Returns:
  137. list: A list of receipts.
  138. """
  139. def f(txn):
  140. if from_key:
  141. sql = (
  142. "SELECT * FROM receipts_linearized WHERE"
  143. " room_id = ? AND stream_id > ? AND stream_id <= ?"
  144. )
  145. txn.execute(
  146. sql,
  147. (room_id, from_key, to_key)
  148. )
  149. else:
  150. sql = (
  151. "SELECT * FROM receipts_linearized WHERE"
  152. " room_id = ? AND stream_id <= ?"
  153. )
  154. txn.execute(
  155. sql,
  156. (room_id, to_key)
  157. )
  158. rows = self.cursor_to_dict(txn)
  159. return rows
  160. rows = yield self.runInteraction(
  161. "get_linearized_receipts_for_room", f
  162. )
  163. if not rows:
  164. defer.returnValue([])
  165. content = {}
  166. for row in rows:
  167. content.setdefault(
  168. row["event_id"], {}
  169. ).setdefault(
  170. row["receipt_type"], {}
  171. )[row["user_id"]] = json.loads(row["data"])
  172. defer.returnValue([{
  173. "type": "m.receipt",
  174. "room_id": room_id,
  175. "content": content,
  176. }])
  177. @cachedList(cached_method_name="get_linearized_receipts_for_room",
  178. list_name="room_ids", num_args=3, inlineCallbacks=True)
  179. def _get_linearized_receipts_for_rooms(self, room_ids, to_key, from_key=None):
  180. if not room_ids:
  181. defer.returnValue({})
  182. def f(txn):
  183. if from_key:
  184. sql = (
  185. "SELECT * FROM receipts_linearized WHERE"
  186. " room_id IN (%s) AND stream_id > ? AND stream_id <= ?"
  187. ) % (
  188. ",".join(["?"] * len(room_ids))
  189. )
  190. args = list(room_ids)
  191. args.extend([from_key, to_key])
  192. txn.execute(sql, args)
  193. else:
  194. sql = (
  195. "SELECT * FROM receipts_linearized WHERE"
  196. " room_id IN (%s) AND stream_id <= ?"
  197. ) % (
  198. ",".join(["?"] * len(room_ids))
  199. )
  200. args = list(room_ids)
  201. args.append(to_key)
  202. txn.execute(sql, args)
  203. return self.cursor_to_dict(txn)
  204. txn_results = yield self.runInteraction(
  205. "_get_linearized_receipts_for_rooms", f
  206. )
  207. results = {}
  208. for row in txn_results:
  209. # We want a single event per room, since we want to batch the
  210. # receipts by room, event and type.
  211. room_event = results.setdefault(row["room_id"], {
  212. "type": "m.receipt",
  213. "room_id": row["room_id"],
  214. "content": {},
  215. })
  216. # The content is of the form:
  217. # {"$foo:bar": { "read": { "@user:host": <receipt> }, .. }, .. }
  218. event_entry = room_event["content"].setdefault(row["event_id"], {})
  219. receipt_type = event_entry.setdefault(row["receipt_type"], {})
  220. receipt_type[row["user_id"]] = json.loads(row["data"])
  221. results = {
  222. room_id: [results[room_id]] if room_id in results else []
  223. for room_id in room_ids
  224. }
  225. defer.returnValue(results)
  226. def get_all_updated_receipts(self, last_id, current_id, limit=None):
  227. if last_id == current_id:
  228. return defer.succeed([])
  229. def get_all_updated_receipts_txn(txn):
  230. sql = (
  231. "SELECT stream_id, room_id, receipt_type, user_id, event_id, data"
  232. " FROM receipts_linearized"
  233. " WHERE ? < stream_id AND stream_id <= ?"
  234. " ORDER BY stream_id ASC"
  235. )
  236. args = [last_id, current_id]
  237. if limit is not None:
  238. sql += " LIMIT ?"
  239. args.append(limit)
  240. txn.execute(sql, args)
  241. return txn.fetchall()
  242. return self.runInteraction(
  243. "get_all_updated_receipts", get_all_updated_receipts_txn
  244. )
  245. def _invalidate_get_users_with_receipts_in_room(self, room_id, receipt_type,
  246. user_id):
  247. if receipt_type != "m.read":
  248. return
  249. # Returns either an ObservableDeferred or the raw result
  250. res = self.get_users_with_read_receipts_in_room.cache.get(
  251. room_id, None, update_metrics=False,
  252. )
  253. # first handle the Deferred case
  254. if isinstance(res, defer.Deferred):
  255. if res.called:
  256. res = res.result
  257. else:
  258. res = None
  259. if res and user_id in res:
  260. # We'd only be adding to the set, so no point invalidating if the
  261. # user is already there
  262. return
  263. self.get_users_with_read_receipts_in_room.invalidate((room_id,))
  264. class ReceiptsStore(ReceiptsWorkerStore):
  265. def __init__(self, db_conn, hs):
  266. # We instantiate this first as the ReceiptsWorkerStore constructor
  267. # needs to be able to call get_max_receipt_stream_id
  268. self._receipts_id_gen = StreamIdGenerator(
  269. db_conn, "receipts_linearized", "stream_id"
  270. )
  271. super(ReceiptsStore, self).__init__(db_conn, hs)
  272. def get_max_receipt_stream_id(self):
  273. return self._receipts_id_gen.get_current_token()
  274. def insert_linearized_receipt_txn(self, txn, room_id, receipt_type,
  275. user_id, event_id, data, stream_id):
  276. res = self._simple_select_one_txn(
  277. txn,
  278. table="events",
  279. retcols=["topological_ordering", "stream_ordering"],
  280. keyvalues={"event_id": event_id},
  281. allow_none=True
  282. )
  283. stream_ordering = int(res["stream_ordering"]) if res else None
  284. # We don't want to clobber receipts for more recent events, so we
  285. # have to compare orderings of existing receipts
  286. if stream_ordering is not None:
  287. sql = (
  288. "SELECT stream_ordering, event_id FROM events"
  289. " INNER JOIN receipts_linearized as r USING (event_id, room_id)"
  290. " WHERE r.room_id = ? AND r.receipt_type = ? AND r.user_id = ?"
  291. )
  292. txn.execute(sql, (room_id, receipt_type, user_id))
  293. for so, eid in txn:
  294. if int(so) >= stream_ordering:
  295. logger.debug(
  296. "Ignoring new receipt for %s in favour of existing "
  297. "one for later event %s",
  298. event_id, eid,
  299. )
  300. return False
  301. txn.call_after(
  302. self.get_receipts_for_room.invalidate, (room_id, receipt_type)
  303. )
  304. txn.call_after(
  305. self._invalidate_get_users_with_receipts_in_room,
  306. room_id, receipt_type, user_id,
  307. )
  308. txn.call_after(
  309. self.get_receipts_for_user.invalidate, (user_id, receipt_type)
  310. )
  311. # FIXME: This shouldn't invalidate the whole cache
  312. txn.call_after(self.get_linearized_receipts_for_room.invalidate_many, (room_id,))
  313. txn.call_after(
  314. self._receipts_stream_cache.entity_has_changed,
  315. room_id, stream_id
  316. )
  317. txn.call_after(
  318. self.get_last_receipt_event_id_for_user.invalidate,
  319. (user_id, room_id, receipt_type)
  320. )
  321. self._simple_delete_txn(
  322. txn,
  323. table="receipts_linearized",
  324. keyvalues={
  325. "room_id": room_id,
  326. "receipt_type": receipt_type,
  327. "user_id": user_id,
  328. }
  329. )
  330. self._simple_insert_txn(
  331. txn,
  332. table="receipts_linearized",
  333. values={
  334. "stream_id": stream_id,
  335. "room_id": room_id,
  336. "receipt_type": receipt_type,
  337. "user_id": user_id,
  338. "event_id": event_id,
  339. "data": json.dumps(data),
  340. }
  341. )
  342. if receipt_type == "m.read" and stream_ordering is not None:
  343. self._remove_old_push_actions_before_txn(
  344. txn,
  345. room_id=room_id,
  346. user_id=user_id,
  347. stream_ordering=stream_ordering,
  348. )
  349. return True
  350. @defer.inlineCallbacks
  351. def insert_receipt(self, room_id, receipt_type, user_id, event_ids, data):
  352. """Insert a receipt, either from local client or remote server.
  353. Automatically does conversion between linearized and graph
  354. representations.
  355. """
  356. if not event_ids:
  357. return
  358. if len(event_ids) == 1:
  359. linearized_event_id = event_ids[0]
  360. else:
  361. # we need to points in graph -> linearized form.
  362. # TODO: Make this better.
  363. def graph_to_linear(txn):
  364. query = (
  365. "SELECT event_id WHERE room_id = ? AND stream_ordering IN ("
  366. " SELECT max(stream_ordering) WHERE event_id IN (%s)"
  367. ")"
  368. ) % (",".join(["?"] * len(event_ids)))
  369. txn.execute(query, [room_id] + event_ids)
  370. rows = txn.fetchall()
  371. if rows:
  372. return rows[0][0]
  373. else:
  374. raise RuntimeError("Unrecognized event_ids: %r" % (event_ids,))
  375. linearized_event_id = yield self.runInteraction(
  376. "insert_receipt_conv", graph_to_linear
  377. )
  378. stream_id_manager = self._receipts_id_gen.get_next()
  379. with stream_id_manager as stream_id:
  380. have_persisted = yield self.runInteraction(
  381. "insert_linearized_receipt",
  382. self.insert_linearized_receipt_txn,
  383. room_id, receipt_type, user_id, linearized_event_id,
  384. data,
  385. stream_id=stream_id,
  386. )
  387. if not have_persisted:
  388. defer.returnValue(None)
  389. yield self.insert_graph_receipt(
  390. room_id, receipt_type, user_id, event_ids, data
  391. )
  392. max_persisted_id = self._receipts_id_gen.get_current_token()
  393. defer.returnValue((stream_id, max_persisted_id))
  394. def insert_graph_receipt(self, room_id, receipt_type, user_id, event_ids,
  395. data):
  396. return self.runInteraction(
  397. "insert_graph_receipt",
  398. self.insert_graph_receipt_txn,
  399. room_id, receipt_type, user_id, event_ids, data
  400. )
  401. def insert_graph_receipt_txn(self, txn, room_id, receipt_type,
  402. user_id, event_ids, data):
  403. txn.call_after(
  404. self.get_receipts_for_room.invalidate, (room_id, receipt_type)
  405. )
  406. txn.call_after(
  407. self._invalidate_get_users_with_receipts_in_room,
  408. room_id, receipt_type, user_id,
  409. )
  410. txn.call_after(
  411. self.get_receipts_for_user.invalidate, (user_id, receipt_type)
  412. )
  413. # FIXME: This shouldn't invalidate the whole cache
  414. txn.call_after(self.get_linearized_receipts_for_room.invalidate_many, (room_id,))
  415. self._simple_delete_txn(
  416. txn,
  417. table="receipts_graph",
  418. keyvalues={
  419. "room_id": room_id,
  420. "receipt_type": receipt_type,
  421. "user_id": user_id,
  422. }
  423. )
  424. self._simple_insert_txn(
  425. txn,
  426. table="receipts_graph",
  427. values={
  428. "room_id": room_id,
  429. "receipt_type": receipt_type,
  430. "user_id": user_id,
  431. "event_ids": json.dumps(event_ids),
  432. "data": json.dumps(data),
  433. }
  434. )