Improve the performance of calculating ignored users in large rooms (#9024)

This allows for efficiently finding which users ignore a particular
user.

Co-authored-by: Erik Johnston <erik@matrix.org>

synapse/storage/databases/main/account_data.py

@@ -16,7 +16,7 @@
 
 import abc
 import logging
-from typing import Dict, List, Optional, Tuple
+from typing import Dict, List, Optional, Set, Tuple
 
 from synapse.api.constants import AccountDataTypes
 from synapse.storage._base import SQLBaseStore, db_to_json
@@ -24,7 +24,7 @@ from synapse.storage.database import DatabasePool
 from synapse.storage.util.id_generators import StreamIdGenerator
 from synapse.types import JsonDict
 from synapse.util import json_encoder
-from synapse.util.caches.descriptors import _CacheContext, cached
+from synapse.util.caches.descriptors import cached
 from synapse.util.caches.stream_change_cache import StreamChangeCache
 
 logger = logging.getLogger(__name__)
@@ -287,23 +287,25 @@ class AccountDataWorkerStore(SQLBaseStore, metaclass=abc.ABCMeta):
             "get_updated_account_data_for_user", get_updated_account_data_for_user_txn
         )
 
-    @cached(num_args=2, cache_context=True, max_entries=5000)
-    async def is_ignored_by(
-        self, ignored_user_id: str, ignorer_user_id: str, cache_context: _CacheContext
-    ) -> bool:
-        ignored_account_data = await self.get_global_account_data_by_type_for_user(
-            AccountDataTypes.IGNORED_USER_LIST,
-            ignorer_user_id,
-            on_invalidate=cache_context.invalidate,
-        )
-        if not ignored_account_data:
-            return False
+    @cached(max_entries=5000, iterable=True)
+    async def ignored_by(self, user_id: str) -> Set[str]:
+        """
+        Get users which ignore the given user.
 
-        try:
-            return ignored_user_id in ignored_account_data.get("ignored_users", {})
-        except TypeError:
-            # The type of the ignored_users field is invalid.
-            return False
+        Params:
+            user_id: The user ID which might be ignored.
+
+        Return:
+            The user IDs which ignore the given user.
+        """
+        return set(
+            await self.db_pool.simple_select_onecol(
+                table="ignored_users",
+                keyvalues={"ignored_user_id": user_id},
+                retcol="ignorer_user_id",
+                desc="ignored_by",
+            )
+        )
 
 
 class AccountDataStore(AccountDataWorkerStore):
@@ -390,18 +392,14 @@ class AccountDataStore(AccountDataWorkerStore):
         Returns:
             The maximum stream ID.
         """
-        content_json = json_encoder.encode(content)
-
         async with self._account_data_id_gen.get_next() as next_id:
-            # no need to lock here as account_data has a unique constraint on
-            # (user_id, account_data_type) so simple_upsert will retry if
-            # there is a conflict.
-            await self.db_pool.simple_upsert(
-                desc="add_user_account_data",
-                table="account_data",
-                keyvalues={"user_id": user_id, "account_data_type": account_data_type},
-                values={"stream_id": next_id, "content": content_json},
-                lock=False,
+            await self.db_pool.runInteraction(
+                "add_user_account_data",
+                self._add_account_data_for_user,
+                next_id,
+                user_id,
+                account_data_type,
+                content,
             )
 
             # it's theoretically possible for the above to succeed and the
@@ -424,6 +422,71 @@ class AccountDataStore(AccountDataWorkerStore):
 
         return self._account_data_id_gen.get_current_token()
 
+    def _add_account_data_for_user(
+        self,
+        txn,
+        next_id: int,
+        user_id: str,
+        account_data_type: str,
+        content: JsonDict,
+    ) -> None:
+        content_json = json_encoder.encode(content)
+
+        # no need to lock here as account_data has a unique constraint on
+        # (user_id, account_data_type) so simple_upsert will retry if
+        # there is a conflict.
+        self.db_pool.simple_upsert_txn(
+            txn,
+            table="account_data",
+            keyvalues={"user_id": user_id, "account_data_type": account_data_type},
+            values={"stream_id": next_id, "content": content_json},
+            lock=False,
+        )
+
+        # Ignored users get denormalized into a separate table as an optimisation.
+        if account_data_type != AccountDataTypes.IGNORED_USER_LIST:
+            return
+
+        # Insert / delete to sync the list of ignored users.
+        previously_ignored_users = set(
+            self.db_pool.simple_select_onecol_txn(
+                txn,
+                table="ignored_users",
+                keyvalues={"ignorer_user_id": user_id},
+                retcol="ignored_user_id",
+            )
+        )
+
+        # If the data is invalid, no one is ignored.
+        ignored_users_content = content.get("ignored_users", {})
+        if isinstance(ignored_users_content, dict):
+            currently_ignored_users = set(ignored_users_content)
+        else:
+            currently_ignored_users = set()
+
+        # Delete entries which are no longer ignored.
+        self.db_pool.simple_delete_many_txn(
+            txn,
+            table="ignored_users",
+            column="ignored_user_id",
+            iterable=previously_ignored_users - currently_ignored_users,
+            keyvalues={"ignorer_user_id": user_id},
+        )
+
+        # Add entries which are newly ignored.
+        self.db_pool.simple_insert_many_txn(
+            txn,
+            table="ignored_users",
+            values=[
+                {"ignorer_user_id": user_id, "ignored_user_id": u}
+                for u in currently_ignored_users - previously_ignored_users
+            ],
+        )
+
+        # Invalidate the cache for any ignored users which were added or removed.
+        for ignored_user_id in previously_ignored_users ^ currently_ignored_users:
+            self._invalidate_cache_and_stream(txn, self.ignored_by, (ignored_user_id,))
+
     async def _update_max_stream_id(self, next_id: int) -> None:
         """Update the max stream_id