anonymousland-synapse/synapse/storage/databases/main/user_directory.py

856 lines
31 KiB
Python
Raw Normal View History

2017-05-31 06:51:01 -04:00
# -*- coding: utf-8 -*-
# Copyright 2017 Vector Creations Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
2018-07-09 02:09:20 -04:00
import logging
import re
Add /user/{user_id}/shared_rooms/ api (#7785) * Add shared_rooms api * Add changelog * Add . * Wrap response in {"rooms": } * linting * Add unstable_features key * Remove options from isort that aren't part of 5.x `-y` and `-rc` are now default behaviour and no longer exist. `dont-skip` is no longer required https://timothycrosley.github.io/isort/CHANGELOG/#500-penny-july-4-2020 * Update imports to make isort happy * Add changelog * Update tox.ini file with correct invocation * fix linting again for isort * Vendor prefix unstable API * Fix to match spec * import Codes * import Codes * Use FORBIDDEN * Update changelog.d/7785.feature Co-authored-by: Andrew Morgan <1342360+anoadragon453@users.noreply.github.com> * Implement get_shared_rooms_for_users * a comma * trailing whitespace * Handle the easy feedback * Switch to using runInteraction * Add tests * Feedback * Seperate unstable endpoint from v2 * Add upgrade node * a line * Fix style by adding a blank line at EOF. * Update synapse/storage/databases/main/user_directory.py Co-authored-by: Tulir Asokan <tulir@maunium.net> * Update synapse/storage/databases/main/user_directory.py Co-authored-by: Andrew Morgan <1342360+anoadragon453@users.noreply.github.com> * Update UPGRADE.rst Co-authored-by: Andrew Morgan <1342360+anoadragon453@users.noreply.github.com> * Fix UPGRADE/CHANGELOG unstable paths unstable unstable unstable Co-authored-by: Andrew Morgan <1342360+anoadragon453@users.noreply.github.com> Co-authored-by: Tulir Asokan <tulir@maunium.net> Co-authored-by: Andrew Morgan <1342360+anoadragon453@users.noreply.github.com> Co-authored-by: Patrick Cloke <clokep@users.noreply.github.com> Co-authored-by: Tulir Asokan <tulir@maunium.net>
2020-09-02 08:18:40 -04:00
from typing import Any, Dict, Iterable, Optional, Set, Tuple
2017-05-31 06:51:01 -04:00
from synapse.api.constants import EventTypes, HistoryVisibility, JoinRules
from synapse.storage.database import DatabasePool
from synapse.storage.databases.main.state import StateFilter
from synapse.storage.databases.main.state_deltas import StateDeltasStore
2019-03-11 09:39:12 -04:00
from synapse.storage.engines import PostgresEngine, Sqlite3Engine
2017-05-31 09:29:32 -04:00
from synapse.types import get_domain_from_id, get_localpart_from_id
2019-03-12 06:47:14 -04:00
from synapse.util.caches.descriptors import cached
2017-05-31 06:51:01 -04:00
logger = logging.getLogger(__name__)
2017-05-31 06:51:01 -04:00
TEMP_TABLE = "_temp_populate_user_directory"
class UserDirectoryBackgroundUpdateStore(StateDeltasStore):
# How many records do we calculate before sending it to
# add_users_who_share_private_rooms?
SHARE_PRIVATE_WORKING_SET = 500
def __init__(self, database: DatabasePool, db_conn, hs):
super().__init__(database, db_conn, hs)
self.server_name = hs.hostname
self.db_pool.updates.register_background_update_handler(
"populate_user_directory_createtables",
self._populate_user_directory_createtables,
)
self.db_pool.updates.register_background_update_handler(
"populate_user_directory_process_rooms",
self._populate_user_directory_process_rooms,
)
self.db_pool.updates.register_background_update_handler(
"populate_user_directory_process_users",
self._populate_user_directory_process_users,
)
self.db_pool.updates.register_background_update_handler(
"populate_user_directory_cleanup", self._populate_user_directory_cleanup
)
async def _populate_user_directory_createtables(self, progress, batch_size):
# Get all the rooms that we want to process.
def _make_staging_area(txn):
sql = (
"CREATE TABLE IF NOT EXISTS "
+ TEMP_TABLE
+ "_rooms(room_id TEXT NOT NULL, events BIGINT NOT NULL)"
)
txn.execute(sql)
sql = (
"CREATE TABLE IF NOT EXISTS "
+ TEMP_TABLE
+ "_position(position TEXT NOT NULL)"
)
txn.execute(sql)
# Get rooms we want to process from the database
sql = """
SELECT room_id, count(*) FROM current_state_events
GROUP BY room_id
"""
txn.execute(sql)
rooms = [{"room_id": x[0], "events": x[1]} for x in txn.fetchall()]
self.db_pool.simple_insert_many_txn(txn, TEMP_TABLE + "_rooms", rooms)
del rooms
# If search all users is on, get all the users we want to add.
if self.hs.config.user_directory_search_all_users:
sql = (
"CREATE TABLE IF NOT EXISTS "
+ TEMP_TABLE
+ "_users(user_id TEXT NOT NULL)"
)
txn.execute(sql)
txn.execute("SELECT name FROM users")
users = [{"user_id": x[0]} for x in txn.fetchall()]
self.db_pool.simple_insert_many_txn(txn, TEMP_TABLE + "_users", users)
new_pos = await self.get_max_stream_id_in_current_state_deltas()
await self.db_pool.runInteraction(
"populate_user_directory_temp_build", _make_staging_area
)
await self.db_pool.simple_insert(
TEMP_TABLE + "_position", {"position": new_pos}
)
await self.db_pool.updates._end_background_update(
"populate_user_directory_createtables"
)
return 1
async def _populate_user_directory_cleanup(self, progress, batch_size):
"""
Update the user directory stream position, then clean up the old tables.
"""
position = await self.db_pool.simple_select_one_onecol(
TEMP_TABLE + "_position", None, "position"
)
await self.update_user_directory_stream_pos(position)
def _delete_staging_area(txn):
txn.execute("DROP TABLE IF EXISTS " + TEMP_TABLE + "_rooms")
txn.execute("DROP TABLE IF EXISTS " + TEMP_TABLE + "_users")
txn.execute("DROP TABLE IF EXISTS " + TEMP_TABLE + "_position")
await self.db_pool.runInteraction(
"populate_user_directory_cleanup", _delete_staging_area
)
await self.db_pool.updates._end_background_update(
"populate_user_directory_cleanup"
)
return 1
async def _populate_user_directory_process_rooms(self, progress, batch_size):
"""
Args:
progress (dict)
batch_size (int): Maximum number of state events to process
per cycle.
"""
state = self.hs.get_state_handler()
# If we don't have progress filed, delete everything.
if not progress:
await self.delete_all_from_user_dir()
def _get_next_batch(txn):
# Only fetch 250 rooms, so we don't fetch too many at once, even
# if those 250 rooms have less than batch_size state events.
sql = """
SELECT room_id, events FROM %s
ORDER BY events DESC
LIMIT 250
""" % (
TEMP_TABLE + "_rooms",
)
txn.execute(sql)
rooms_to_work_on = txn.fetchall()
if not rooms_to_work_on:
return None
# Get how many are left to process, so we can give status on how
# far we are in processing
txn.execute("SELECT COUNT(*) FROM " + TEMP_TABLE + "_rooms")
progress["remaining"] = txn.fetchone()[0]
return rooms_to_work_on
rooms_to_work_on = await self.db_pool.runInteraction(
"populate_user_directory_temp_read", _get_next_batch
)
# No more rooms -- complete the transaction.
if not rooms_to_work_on:
await self.db_pool.updates._end_background_update(
"populate_user_directory_process_rooms"
)
return 1
logger.debug(
"Processing the next %d rooms of %d remaining"
% (len(rooms_to_work_on), progress["remaining"])
)
processed_event_count = 0
for room_id, event_count in rooms_to_work_on:
is_in_room = await self.is_host_joined(room_id, self.server_name)
if is_in_room:
is_public = await self.is_room_world_readable_or_publicly_joinable(
room_id
)
users_with_profile = await state.get_current_users_in_room(room_id)
user_ids = set(users_with_profile)
# Update each user in the user directory.
for user_id, profile in users_with_profile.items():
await self.update_profile_in_user_dir(
user_id, profile.display_name, profile.avatar_url
)
to_insert = set()
if is_public:
for user_id in user_ids:
if self.get_if_app_services_interested_in_user(user_id):
continue
to_insert.add(user_id)
if to_insert:
await self.add_users_in_public_rooms(room_id, to_insert)
to_insert.clear()
else:
for user_id in user_ids:
if not self.hs.is_mine_id(user_id):
continue
if self.get_if_app_services_interested_in_user(user_id):
continue
for other_user_id in user_ids:
if user_id == other_user_id:
continue
user_set = (user_id, other_user_id)
to_insert.add(user_set)
# If it gets too big, stop and write to the database
# to prevent storing too much in RAM.
if len(to_insert) >= self.SHARE_PRIVATE_WORKING_SET:
await self.add_users_who_share_private_room(
room_id, to_insert
)
to_insert.clear()
if to_insert:
await self.add_users_who_share_private_room(room_id, to_insert)
to_insert.clear()
# We've finished a room. Delete it from the table.
await self.db_pool.simple_delete_one(
TEMP_TABLE + "_rooms", {"room_id": room_id}
)
# Update the remaining counter.
progress["remaining"] -= 1
await self.db_pool.runInteraction(
"populate_user_directory",
self.db_pool.updates._background_update_progress_txn,
"populate_user_directory_process_rooms",
progress,
)
processed_event_count += event_count
if processed_event_count > batch_size:
# Don't process any more rooms, we've hit our batch size.
return processed_event_count
return processed_event_count
async def _populate_user_directory_process_users(self, progress, batch_size):
"""
If search_all_users is enabled, add all of the users to the user directory.
"""
if not self.hs.config.user_directory_search_all_users:
await self.db_pool.updates._end_background_update(
"populate_user_directory_process_users"
)
return 1
def _get_next_batch(txn):
sql = "SELECT user_id FROM %s LIMIT %s" % (
TEMP_TABLE + "_users",
str(batch_size),
)
txn.execute(sql)
users_to_work_on = txn.fetchall()
if not users_to_work_on:
return None
users_to_work_on = [x[0] for x in users_to_work_on]
# Get how many are left to process, so we can give status on how
# far we are in processing
sql = "SELECT COUNT(*) FROM " + TEMP_TABLE + "_users"
txn.execute(sql)
progress["remaining"] = txn.fetchone()[0]
return users_to_work_on
users_to_work_on = await self.db_pool.runInteraction(
"populate_user_directory_temp_read", _get_next_batch
)
# No more users -- complete the transaction.
if not users_to_work_on:
await self.db_pool.updates._end_background_update(
"populate_user_directory_process_users"
)
return 1
logger.debug(
"Processing the next %d users of %d remaining"
% (len(users_to_work_on), progress["remaining"])
)
for user_id in users_to_work_on:
profile = await self.get_profileinfo(get_localpart_from_id(user_id))
await self.update_profile_in_user_dir(
user_id, profile.display_name, profile.avatar_url
)
# We've finished processing a user. Delete it from the table.
await self.db_pool.simple_delete_one(
TEMP_TABLE + "_users", {"user_id": user_id}
)
# Update the remaining counter.
progress["remaining"] -= 1
await self.db_pool.runInteraction(
"populate_user_directory",
self.db_pool.updates._background_update_progress_txn,
"populate_user_directory_process_users",
progress,
)
return len(users_to_work_on)
async def is_room_world_readable_or_publicly_joinable(self, room_id):
2017-05-31 10:00:29 -04:00
"""Check if the room is either world_readable or publically joinable
"""
# Create a state filter that only queries join and history state event
types_to_filter = (
(EventTypes.JoinRules, ""),
(EventTypes.RoomHistoryVisibility, ""),
)
current_state_ids = await self.get_filtered_current_state_ids(
room_id, StateFilter.from_types(types_to_filter)
2017-05-31 06:51:01 -04:00
)
join_rules_id = current_state_ids.get((EventTypes.JoinRules, ""))
if join_rules_id:
join_rule_ev = await self.get_event(join_rules_id, allow_none=True)
2017-05-31 06:51:01 -04:00
if join_rule_ev:
if join_rule_ev.content.get("join_rule") == JoinRules.PUBLIC:
return True
2017-05-31 06:51:01 -04:00
hist_vis_id = current_state_ids.get((EventTypes.RoomHistoryVisibility, ""))
if hist_vis_id:
hist_vis_ev = await self.get_event(hist_vis_id, allow_none=True)
2017-05-31 06:51:01 -04:00
if hist_vis_ev:
if (
hist_vis_ev.content.get("history_visibility")
== HistoryVisibility.WORLD_READABLE
):
return True
2017-05-31 06:51:01 -04:00
return False
2017-05-31 06:51:01 -04:00
async def update_profile_in_user_dir(
self, user_id: str, display_name: str, avatar_url: str
) -> None:
"""
Update or add a user's profile in the user directory.
2017-05-31 10:00:29 -04:00
"""
# If the display name or avatar URL are unexpected types, overwrite them.
if not isinstance(display_name, str):
display_name = None
if not isinstance(avatar_url, str):
avatar_url = None
2019-03-07 04:22:53 -05:00
def _update_profile_in_user_dir_txn(txn):
new_entry = self.db_pool.simple_upsert_txn(
txn,
table="user_directory",
keyvalues={"user_id": user_id},
values={"display_name": display_name, "avatar_url": avatar_url},
lock=False, # We're only inserter
)
if isinstance(self.database_engine, PostgresEngine):
2017-11-29 13:27:05 -05:00
# We weight the localpart most highly, then display name and finally
# server name
if self.database_engine.can_native_upsert:
sql = """
INSERT INTO user_directory_search(user_id, vector)
VALUES (?,
setweight(to_tsvector('simple', ?), 'A')
|| setweight(to_tsvector('simple', ?), 'D')
|| setweight(to_tsvector('simple', COALESCE(?, '')), 'B')
) ON CONFLICT (user_id) DO UPDATE SET vector=EXCLUDED.vector
"""
2017-06-13 06:19:18 -04:00
txn.execute(
sql,
(
user_id,
get_localpart_from_id(user_id),
get_domain_from_id(user_id),
display_name,
),
)
else:
# TODO: Remove this code after we've bumped the minimum version
# of postgres to always support upserts, so we can get rid of
# `new_entry` usage
if new_entry is True:
sql = """
INSERT INTO user_directory_search(user_id, vector)
VALUES (?,
setweight(to_tsvector('simple', ?), 'A')
|| setweight(to_tsvector('simple', ?), 'D')
|| setweight(to_tsvector('simple', COALESCE(?, '')), 'B')
)
"""
txn.execute(
sql,
(
user_id,
get_localpart_from_id(user_id),
get_domain_from_id(user_id),
display_name,
),
)
elif new_entry is False:
sql = """
UPDATE user_directory_search
SET vector = setweight(to_tsvector('simple', ?), 'A')
|| setweight(to_tsvector('simple', ?), 'D')
|| setweight(to_tsvector('simple', COALESCE(?, '')), 'B')
WHERE user_id = ?
"""
txn.execute(
sql,
(
get_localpart_from_id(user_id),
get_domain_from_id(user_id),
display_name,
user_id,
),
)
else:
raise RuntimeError(
"upsert returned None when 'can_native_upsert' is False"
2017-06-13 06:19:18 -04:00
)
elif isinstance(self.database_engine, Sqlite3Engine):
value = "%s %s" % (user_id, display_name) if display_name else user_id
self.db_pool.simple_upsert_txn(
txn,
table="user_directory_search",
keyvalues={"user_id": user_id},
values={"value": value},
lock=False, # We're only inserter
)
else:
# This should be unreachable.
raise Exception("Unrecognized database engine")
txn.call_after(self.get_user_in_directory.invalidate, (user_id,))
await self.db_pool.runInteraction(
"update_profile_in_user_dir", _update_profile_in_user_dir_txn
)
2017-06-01 09:50:46 -04:00
async def add_users_who_share_private_room(
self, room_id: str, user_id_tuples: Iterable[Tuple[str, str]]
) -> None:
2019-03-12 06:47:14 -04:00
"""Insert entries into the users_who_share_private_rooms table. The first
user should be a local user.
Args:
room_id
user_id_tuples: iterable of 2-tuple of user IDs.
"""
Use autocommit mode for single statement DB functions. (#8542) Autocommit means that we don't wrap the functions in transactions, and instead get executed directly. Introduced in #8456. This will help: 1. reduce the number of `could not serialize access due to concurrent delete` errors that we see (though there are a few functions that often cause serialization errors that we don't fix here); 2. improve the DB performance, as it no longer needs to deal with the overhead of `REPEATABLE READ` isolation levels; and 3. improve wall clock speed of these functions, as we no longer need to send `BEGIN` and `COMMIT` to the DB. Some notes about the differences between autocommit mode and our default `REPEATABLE READ` transactions: 1. Currently `autocommit` only applies when using PostgreSQL, and is ignored when using SQLite (due to silliness with [Twisted DB classes](https://twistedmatrix.com/trac/ticket/9998)). 2. Autocommit functions may get retried on error, which means they can get applied *twice* (or more) to the DB (since they are not in a transaction the previous call would not get rolled back). This means that the functions need to be idempotent (or otherwise not care about being called multiple times). Read queries, simple deletes, and updates/upserts that replace rows (rather than generating new values from existing rows) are all idempotent. 3. Autocommit functions no longer get executed in [`REPEATABLE READ`](https://www.postgresql.org/docs/current/transaction-iso.html) isolation level, and so data can change queries, which is fine for single statement queries.
2020-10-14 10:50:59 -04:00
await self.db_pool.simple_upsert_many(
table="users_who_share_private_rooms",
key_names=["user_id", "other_user_id", "room_id"],
key_values=[
(user_id, other_user_id, room_id)
for user_id, other_user_id in user_id_tuples
],
value_names=(),
value_values=None,
desc="add_users_who_share_room",
2019-03-12 06:47:14 -04:00
)
async def add_users_in_public_rooms(
self, room_id: str, user_ids: Iterable[str]
) -> None:
2019-03-12 06:47:14 -04:00
"""Insert entries into the users_who_share_private_rooms table. The first
user should be a local user.
Args:
room_id
user_ids
2019-03-12 06:47:14 -04:00
"""
Use autocommit mode for single statement DB functions. (#8542) Autocommit means that we don't wrap the functions in transactions, and instead get executed directly. Introduced in #8456. This will help: 1. reduce the number of `could not serialize access due to concurrent delete` errors that we see (though there are a few functions that often cause serialization errors that we don't fix here); 2. improve the DB performance, as it no longer needs to deal with the overhead of `REPEATABLE READ` isolation levels; and 3. improve wall clock speed of these functions, as we no longer need to send `BEGIN` and `COMMIT` to the DB. Some notes about the differences between autocommit mode and our default `REPEATABLE READ` transactions: 1. Currently `autocommit` only applies when using PostgreSQL, and is ignored when using SQLite (due to silliness with [Twisted DB classes](https://twistedmatrix.com/trac/ticket/9998)). 2. Autocommit functions may get retried on error, which means they can get applied *twice* (or more) to the DB (since they are not in a transaction the previous call would not get rolled back). This means that the functions need to be idempotent (or otherwise not care about being called multiple times). Read queries, simple deletes, and updates/upserts that replace rows (rather than generating new values from existing rows) are all idempotent. 3. Autocommit functions no longer get executed in [`REPEATABLE READ`](https://www.postgresql.org/docs/current/transaction-iso.html) isolation level, and so data can change queries, which is fine for single statement queries.
2020-10-14 10:50:59 -04:00
await self.db_pool.simple_upsert_many(
table="users_in_public_rooms",
key_names=["user_id", "room_id"],
key_values=[(user_id, room_id) for user_id in user_ids],
value_names=(),
value_values=None,
desc="add_users_in_public_rooms",
)
async def delete_all_from_user_dir(self) -> None:
"""Delete the entire user directory
"""
def _delete_all_from_user_dir_txn(txn):
txn.execute("DELETE FROM user_directory")
txn.execute("DELETE FROM user_directory_search")
txn.execute("DELETE FROM users_in_public_rooms")
txn.execute("DELETE FROM users_who_share_private_rooms")
txn.call_after(self.get_user_in_directory.invalidate_all)
await self.db_pool.runInteraction(
"delete_all_from_user_dir", _delete_all_from_user_dir_txn
)
@cached()
async def get_user_in_directory(self, user_id: str) -> Optional[Dict[str, Any]]:
return await self.db_pool.simple_select_one(
table="user_directory",
keyvalues={"user_id": user_id},
retcols=("display_name", "avatar_url"),
allow_none=True,
desc="get_user_in_directory",
)
async def update_user_directory_stream_pos(self, stream_id: str) -> None:
await self.db_pool.simple_update_one(
table="user_directory_stream_pos",
keyvalues={},
updatevalues={"stream_id": stream_id},
desc="update_user_directory_stream_pos",
)
class UserDirectoryStore(UserDirectoryBackgroundUpdateStore):
# How many records do we calculate before sending it to
# add_users_who_share_private_rooms?
SHARE_PRIVATE_WORKING_SET = 500
def __init__(self, database: DatabasePool, db_conn, hs):
super().__init__(database, db_conn, hs)
async def remove_from_user_dir(self, user_id: str) -> None:
def _remove_from_user_dir_txn(txn):
self.db_pool.simple_delete_txn(
txn, table="user_directory", keyvalues={"user_id": user_id}
)
self.db_pool.simple_delete_txn(
txn, table="user_directory_search", keyvalues={"user_id": user_id}
)
self.db_pool.simple_delete_txn(
txn, table="users_in_public_rooms", keyvalues={"user_id": user_id}
)
self.db_pool.simple_delete_txn(
txn,
table="users_who_share_private_rooms",
keyvalues={"user_id": user_id},
)
self.db_pool.simple_delete_txn(
txn,
table="users_who_share_private_rooms",
keyvalues={"other_user_id": user_id},
)
txn.call_after(self.get_user_in_directory.invalidate, (user_id,))
await self.db_pool.runInteraction(
"remove_from_user_dir", _remove_from_user_dir_txn
)
async def get_users_in_dir_due_to_room(self, room_id):
"""Get all user_ids that are in the room directory because they're
in the given room_id
"""
user_ids_share_pub = await self.db_pool.simple_select_onecol(
table="users_in_public_rooms",
keyvalues={"room_id": room_id},
retcol="user_id",
desc="get_users_in_dir_due_to_room",
)
user_ids_share_priv = await self.db_pool.simple_select_onecol(
table="users_who_share_private_rooms",
keyvalues={"room_id": room_id},
retcol="other_user_id",
desc="get_users_in_dir_due_to_room",
)
user_ids = set(user_ids_share_pub)
user_ids.update(user_ids_share_priv)
return user_ids
async def remove_user_who_share_room(self, user_id: str, room_id: str) -> None:
"""
2019-03-07 04:22:53 -05:00
Deletes entries in the users_who_share_*_rooms table. The first
user should be a local user.
Args:
user_id
room_id
"""
def _remove_user_who_share_room_txn(txn):
self.db_pool.simple_delete_txn(
txn,
2019-03-07 04:22:53 -05:00
table="users_who_share_private_rooms",
keyvalues={"user_id": user_id, "room_id": room_id},
)
self.db_pool.simple_delete_txn(
2019-03-07 04:22:53 -05:00
txn,
table="users_who_share_private_rooms",
keyvalues={"other_user_id": user_id, "room_id": room_id},
)
self.db_pool.simple_delete_txn(
2019-03-07 04:22:53 -05:00
txn,
2019-03-12 06:47:14 -04:00
table="users_in_public_rooms",
2019-03-07 04:22:53 -05:00
keyvalues={"user_id": user_id, "room_id": room_id},
)
2019-03-11 06:11:36 -04:00
await self.db_pool.runInteraction(
"remove_user_who_share_room", _remove_user_who_share_room_txn
2017-05-31 06:51:01 -04:00
)
async def get_user_dir_rooms_user_is_in(self, user_id):
2019-03-12 06:47:14 -04:00
"""
Returns the rooms that a user is in.
Args:
user_id(str): Must be a local user
Returns:
2019-03-07 04:22:53 -05:00
list: user_id
"""
rows = await self.db_pool.simple_select_onecol(
2019-03-07 04:22:53 -05:00
table="users_who_share_private_rooms",
keyvalues={"user_id": user_id},
2019-03-12 06:47:14 -04:00
retcol="room_id",
desc="get_rooms_user_is_in",
2019-03-07 04:22:53 -05:00
)
pub_rows = await self.db_pool.simple_select_onecol(
2019-03-12 06:47:14 -04:00
table="users_in_public_rooms",
keyvalues={"user_id": user_id},
2019-03-12 06:47:14 -04:00
retcol="room_id",
desc="get_rooms_user_is_in",
)
2019-03-07 04:22:53 -05:00
users = set(pub_rows)
users.update(rows)
return list(users)
Add /user/{user_id}/shared_rooms/ api (#7785) * Add shared_rooms api * Add changelog * Add . * Wrap response in {"rooms": } * linting * Add unstable_features key * Remove options from isort that aren't part of 5.x `-y` and `-rc` are now default behaviour and no longer exist. `dont-skip` is no longer required https://timothycrosley.github.io/isort/CHANGELOG/#500-penny-july-4-2020 * Update imports to make isort happy * Add changelog * Update tox.ini file with correct invocation * fix linting again for isort * Vendor prefix unstable API * Fix to match spec * import Codes * import Codes * Use FORBIDDEN * Update changelog.d/7785.feature Co-authored-by: Andrew Morgan <1342360+anoadragon453@users.noreply.github.com> * Implement get_shared_rooms_for_users * a comma * trailing whitespace * Handle the easy feedback * Switch to using runInteraction * Add tests * Feedback * Seperate unstable endpoint from v2 * Add upgrade node * a line * Fix style by adding a blank line at EOF. * Update synapse/storage/databases/main/user_directory.py Co-authored-by: Tulir Asokan <tulir@maunium.net> * Update synapse/storage/databases/main/user_directory.py Co-authored-by: Andrew Morgan <1342360+anoadragon453@users.noreply.github.com> * Update UPGRADE.rst Co-authored-by: Andrew Morgan <1342360+anoadragon453@users.noreply.github.com> * Fix UPGRADE/CHANGELOG unstable paths unstable unstable unstable Co-authored-by: Andrew Morgan <1342360+anoadragon453@users.noreply.github.com> Co-authored-by: Tulir Asokan <tulir@maunium.net> Co-authored-by: Andrew Morgan <1342360+anoadragon453@users.noreply.github.com> Co-authored-by: Patrick Cloke <clokep@users.noreply.github.com> Co-authored-by: Tulir Asokan <tulir@maunium.net>
2020-09-02 08:18:40 -04:00
@cached()
async def get_shared_rooms_for_users(
self, user_id: str, other_user_id: str
) -> Set[str]:
"""
Returns the rooms that a local user shares with another local or remote user.
Args:
user_id: The MXID of a local user
other_user_id: The MXID of the other user
Returns:
A set of room ID's that the users share.
"""
def _get_shared_rooms_for_users_txn(txn):
txn.execute(
"""
SELECT p1.room_id
FROM users_in_public_rooms as p1
INNER JOIN users_in_public_rooms as p2
ON p1.room_id = p2.room_id
AND p1.user_id = ?
AND p2.user_id = ?
UNION
SELECT room_id
FROM users_who_share_private_rooms
WHERE
user_id = ?
AND other_user_id = ?
""",
(user_id, other_user_id, user_id, other_user_id),
)
rows = self.db_pool.cursor_to_dict(txn)
return rows
rows = await self.db_pool.runInteraction(
"get_shared_rooms_for_users", _get_shared_rooms_for_users_txn
)
return {row["room_id"] for row in rows}
async def get_user_directory_stream_pos(self) -> int:
return await self.db_pool.simple_select_one_onecol(
2017-05-31 06:51:01 -04:00
table="user_directory_stream_pos",
keyvalues={},
retcol="stream_id",
desc="get_user_directory_stream_pos",
)
async def search_user_dir(self, user_id, search_term, limit):
2017-05-31 10:00:29 -04:00
"""Searches for users in directory
Returns:
dict of the form::
{
"limited": <bool>, # whether there were more results or not
"results": [ # Ordered by best match first
{
"user_id": <user_id>,
"display_name": <display_name>,
"avatar_url": <avatar_url>
}
]
}
"""
if self.hs.config.user_directory_search_all_users:
join_args = (user_id,)
where_clause = "user_id != ?"
else:
join_args = (user_id,)
2019-03-12 10:17:51 -04:00
where_clause = """
(
EXISTS (select 1 from users_in_public_rooms WHERE user_id = t.user_id)
OR EXISTS (
SELECT 1 FROM users_who_share_private_rooms
WHERE user_id = ? AND other_user_id = t.user_id
)
)
"""
2017-05-31 09:00:01 -04:00
if isinstance(self.database_engine, PostgresEngine):
full_query, exact_query, prefix_query = _parse_query_postgres(search_term)
2017-05-31 13:17:47 -04:00
# We order by rank and then if they have profile info
# The ranking algorithm is hand tweaked for "best" results. Broadly
# the idea is we give a higher weight to exact matches.
# The array of numbers are the weights for the various part of the
# search: (domain, _, display name, localpart)
2017-05-31 09:00:01 -04:00
sql = """
SELECT d.user_id AS user_id, display_name, avatar_url
2019-03-12 10:17:51 -04:00
FROM user_directory_search as t
INNER JOIN user_directory AS d USING (user_id)
WHERE
%s
AND vector @@ to_tsquery('simple', ?)
2017-05-31 13:17:47 -04:00
ORDER BY
2019-03-07 04:22:53 -05:00
(CASE WHEN d.user_id IS NOT NULL THEN 4.0 ELSE 1.0 END)
* (CASE WHEN display_name IS NOT NULL THEN 1.2 ELSE 1.0 END)
* (CASE WHEN avatar_url IS NOT NULL THEN 1.2 ELSE 1.0 END)
* (
3 * ts_rank_cd(
'{0.1, 0.1, 0.9, 1.0}',
vector,
to_tsquery('simple', ?),
8
)
+ ts_rank_cd(
'{0.1, 0.1, 0.9, 1.0}',
vector,
to_tsquery('simple', ?),
8
)
)
DESC,
2017-05-31 13:17:47 -04:00
display_name IS NULL,
avatar_url IS NULL
2017-05-31 09:00:01 -04:00
LIMIT ?
""" % (
where_clause,
)
args = join_args + (full_query, exact_query, prefix_query, limit + 1)
2017-05-31 09:00:01 -04:00
elif isinstance(self.database_engine, Sqlite3Engine):
search_query = _parse_query_sqlite(search_term)
2017-05-31 09:00:01 -04:00
sql = """
SELECT d.user_id AS user_id, display_name, avatar_url
2019-03-12 10:17:51 -04:00
FROM user_directory_search as t
INNER JOIN user_directory AS d USING (user_id)
WHERE
%s
AND value MATCH ?
2017-05-31 13:17:47 -04:00
ORDER BY
2017-06-01 09:58:48 -04:00
rank(matchinfo(user_directory_search)) DESC,
2017-05-31 13:17:47 -04:00
display_name IS NULL,
avatar_url IS NULL
2017-05-31 09:00:01 -04:00
LIMIT ?
""" % (
where_clause,
)
args = join_args + (search_query, limit + 1)
2017-05-31 09:00:01 -04:00
else:
# This should be unreachable.
raise Exception("Unrecognized database engine")
results = await self.db_pool.execute(
"search_user_dir", self.db_pool.cursor_to_dict, sql, *args
)
2017-05-31 09:00:01 -04:00
limited = len(results) > limit
return {"limited": limited, "results": results}
2017-05-31 13:07:12 -04:00
def _parse_query_sqlite(search_term):
2017-05-31 13:07:12 -04:00
"""Takes a plain unicode string from the user and converts it into a form
that can be passed to database.
We use this so that we can add prefix matching, which isn't something
that is supported by default.
We specifically add both a prefix and non prefix matching term so that
exact matches get ranked higher.
"""
# Pull out the individual words, discarding any non-word characters.
results = re.findall(r"([\w\-]+)", search_term, re.UNICODE)
return " & ".join("(%s* OR %s)" % (result, result) for result in results)
def _parse_query_postgres(search_term):
"""Takes a plain unicode string from the user and converts it into a form
that can be passed to database.
We use this so that we can add prefix matching, which isn't something
that is supported by default.
"""
# Pull out the individual words, discarding any non-word characters.
results = re.findall(r"([\w\-]+)", search_term, re.UNICODE)
both = " & ".join("(%s:* | %s)" % (result, result) for result in results)
exact = " & ".join("%s" % (result,) for result in results)
prefix = " & ".join("%s:*" % (result,) for result in results)
2017-05-31 13:07:12 -04:00
return both, exact, prefix