From fa0dcbc8fa396fa78aabc524728e08fd84b70a0c Mon Sep 17 00:00:00 2001 From: Brendan Abolivier Date: Tue, 29 Oct 2019 18:35:49 +0000 Subject: Store labels for new events --- synapse/api/constants.py | 3 +++ 1 file changed, 3 insertions(+) (limited to 'synapse/api/constants.py') diff --git a/synapse/api/constants.py b/synapse/api/constants.py index 312196675e..999ec02fd9 100644 --- a/synapse/api/constants.py +++ b/synapse/api/constants.py @@ -138,3 +138,6 @@ class LimitBlockingTypes(object): MONTHLY_ACTIVE_USER = "monthly_active_user" HS_DISABLED = "hs_disabled" + + +LabelsField = "org.matrix.labels" -- cgit 1.5.1 From c6dbca2422bf77ccbf0b52d9245d28c258dac4f3 Mon Sep 17 00:00:00 2001 From: Brendan Abolivier Date: Fri, 1 Nov 2019 10:30:51 +0000 Subject: Incorporate review --- changelog.d/6301.feature | 2 +- synapse/api/constants.py | 5 ++++- synapse/api/filtering.py | 6 ++++-- synapse/storage/data_stores/main/events.py | 12 ++++++++++-- tests/api/test_filtering.py | 10 +++++----- tests/rest/client/v1/test_rooms.py | 10 +++++----- tests/rest/client/v2_alpha/test_sync.py | 10 +++++----- 7 files changed, 34 insertions(+), 21 deletions(-) (limited to 'synapse/api/constants.py') diff --git a/changelog.d/6301.feature b/changelog.d/6301.feature index b7ff3fad3b..78a187a1dc 100644 --- a/changelog.d/6301.feature +++ b/changelog.d/6301.feature @@ -1 +1 @@ -Implement label-based filtering. +Implement label-based filtering on `/sync` and `/messages` ([MSC2326](https://github.com/matrix-org/matrix-doc/pull/2326)). diff --git a/synapse/api/constants.py b/synapse/api/constants.py index 999ec02fd9..cf4ce5f5a2 100644 --- a/synapse/api/constants.py +++ b/synapse/api/constants.py @@ -140,4 +140,7 @@ class LimitBlockingTypes(object): HS_DISABLED = "hs_disabled" -LabelsField = "org.matrix.labels" +class EventContentFields(object): + """Fields found in events' content, regardless of type.""" + # Labels for the event, cf https://github.com/matrix-org/matrix-doc/pull/2326 + Labels = "org.matrix.labels" diff --git a/synapse/api/filtering.py b/synapse/api/filtering.py index bd91b9f018..30a7ee0a7a 100644 --- a/synapse/api/filtering.py +++ b/synapse/api/filtering.py @@ -20,7 +20,7 @@ from jsonschema import FormatChecker from twisted.internet import defer -from synapse.api.constants import LabelsField +from synapse.api.constants import EventContentFields from synapse.api.errors import SynapseError from synapse.storage.presence import UserPresenceState from synapse.types import RoomID, UserID @@ -67,6 +67,8 @@ ROOM_EVENT_FILTER_SCHEMA = { "contains_url": {"type": "boolean"}, "lazy_load_members": {"type": "boolean"}, "include_redundant_members": {"type": "boolean"}, + # Include or exclude events with the provided labels. + # cf https://github.com/matrix-org/matrix-doc/pull/2326 "org.matrix.labels": {"type": "array", "items": {"type": "string"}}, "org.matrix.not_labels": {"type": "array", "items": {"type": "string"}}, }, @@ -307,7 +309,7 @@ class Filter(object): content = event.get("content", {}) # check if there is a string url field in the content for filtering purposes contains_url = isinstance(content.get("url"), text_type) - labels = content.get(LabelsField, []) + labels = content.get(EventContentFields.Labels, []) return self.check_fields(room_id, sender, ev_type, labels, contains_url) diff --git a/synapse/storage/data_stores/main/events.py b/synapse/storage/data_stores/main/events.py index 2b900f1ce1..42ffa9066a 100644 --- a/synapse/storage/data_stores/main/events.py +++ b/synapse/storage/data_stores/main/events.py @@ -29,7 +29,7 @@ from prometheus_client import Counter, Histogram from twisted.internet import defer import synapse.metrics -from synapse.api.constants import EventTypes, LabelsField +from synapse.api.constants import EventTypes, EventContentFields from synapse.api.errors import SynapseError from synapse.events import EventBase # noqa: F401 from synapse.events.snapshot import EventContext # noqa: F401 @@ -1491,7 +1491,7 @@ class EventsStore( self._handle_event_relations(txn, event) # Store the labels for this event. - labels = event.content.get(LabelsField) + labels = event.content.get(EventContentFields.Labels) if labels: self.insert_labels_for_event_txn(txn, event.event_id, labels) @@ -2483,6 +2483,14 @@ class EventsStore( ) def insert_labels_for_event_txn(self, txn, event_id, labels): + """Store the mapping between an event's ID and its labels, with one row per + (event_id, label) tuple. + + Args: + txn (LoggingTransaction): The transaction to execute. + event_id (str): The event's ID. + labels (list[str]): A list of text labels. + """ return self._simple_insert_many_txn( txn=txn, table="event_labels", diff --git a/tests/api/test_filtering.py b/tests/api/test_filtering.py index e004ab1ee5..8ec48c4154 100644 --- a/tests/api/test_filtering.py +++ b/tests/api/test_filtering.py @@ -19,7 +19,7 @@ import jsonschema from twisted.internet import defer -from synapse.api.constants import LabelsField +from synapse.api.constants import EventContentFields from synapse.api.errors import SynapseError from synapse.api.filtering import Filter from synapse.events import FrozenEvent @@ -329,7 +329,7 @@ class FilteringTestCase(unittest.TestCase): sender="@foo:bar", type="m.room.message", room_id="!secretbase:unknown", - content={LabelsField: ["#fun"]}, + content={EventContentFields.Labels: ["#fun"]}, ) self.assertTrue(Filter(definition).check(event)) @@ -338,7 +338,7 @@ class FilteringTestCase(unittest.TestCase): sender="@foo:bar", type="m.room.message", room_id="!secretbase:unknown", - content={LabelsField: ["#notfun"]}, + content={EventContentFields.Labels: ["#notfun"]}, ) self.assertFalse(Filter(definition).check(event)) @@ -349,7 +349,7 @@ class FilteringTestCase(unittest.TestCase): sender="@foo:bar", type="m.room.message", room_id="!secretbase:unknown", - content={LabelsField: ["#fun"]}, + content={EventContentFields.Labels: ["#fun"]}, ) self.assertFalse(Filter(definition).check(event)) @@ -358,7 +358,7 @@ class FilteringTestCase(unittest.TestCase): sender="@foo:bar", type="m.room.message", room_id="!secretbase:unknown", - content={LabelsField: ["#notfun"]}, + content={EventContentFields.Labels: ["#notfun"]}, ) self.assertTrue(Filter(definition).check(event)) diff --git a/tests/rest/client/v1/test_rooms.py b/tests/rest/client/v1/test_rooms.py index 188f47bd7d..0dc0faa0e5 100644 --- a/tests/rest/client/v1/test_rooms.py +++ b/tests/rest/client/v1/test_rooms.py @@ -24,7 +24,7 @@ from six.moves.urllib import parse as urlparse from twisted.internet import defer import synapse.rest.admin -from synapse.api.constants import EventTypes, LabelsField, Membership +from synapse.api.constants import EventContentFields, EventTypes, Membership from synapse.rest.client.v1 import login, profile, room from tests import unittest @@ -860,7 +860,7 @@ class RoomMessageListTestCase(RoomBase): content={ "msgtype": "m.text", "body": "with right label", - LabelsField: ["#fun"], + EventContentFields.Labels: ["#fun"], }, ) @@ -876,7 +876,7 @@ class RoomMessageListTestCase(RoomBase): content={ "msgtype": "m.text", "body": "with wrong label", - LabelsField: ["#work"], + EventContentFields.Labels: ["#work"], }, ) @@ -886,7 +886,7 @@ class RoomMessageListTestCase(RoomBase): content={ "msgtype": "m.text", "body": "with two wrong labels", - LabelsField: ["#work", "#notfun"], + EventContentFields.Labels: ["#work", "#notfun"], }, ) @@ -896,7 +896,7 @@ class RoomMessageListTestCase(RoomBase): content={ "msgtype": "m.text", "body": "with right label", - LabelsField: ["#fun"], + EventContentFields.Labels: ["#fun"], }, ) diff --git a/tests/rest/client/v2_alpha/test_sync.py b/tests/rest/client/v2_alpha/test_sync.py index c5c199d412..c3c6f75ced 100644 --- a/tests/rest/client/v2_alpha/test_sync.py +++ b/tests/rest/client/v2_alpha/test_sync.py @@ -17,7 +17,7 @@ import json from mock import Mock import synapse.rest.admin -from synapse.api.constants import EventTypes, LabelsField +from synapse.api.constants import EventContentFields, EventTypes from synapse.rest.client.v1 import login, room from synapse.rest.client.v2_alpha import sync @@ -157,7 +157,7 @@ class SyncFilterTestCase(unittest.HomeserverTestCase): content={ "msgtype": "m.text", "body": "with right label", - LabelsField: ["#fun"], + EventContentFields.Labels: ["#fun"], }, tok=tok, ) @@ -175,7 +175,7 @@ class SyncFilterTestCase(unittest.HomeserverTestCase): content={ "msgtype": "m.text", "body": "with wrong label", - LabelsField: ["#work"], + EventContentFields.Labels: ["#work"], }, tok=tok, ) @@ -186,7 +186,7 @@ class SyncFilterTestCase(unittest.HomeserverTestCase): content={ "msgtype": "m.text", "body": "with two wrong labels", - LabelsField: ["#work", "#notfun"], + EventContentFields.Labels: ["#work", "#notfun"], }, tok=tok, ) @@ -197,7 +197,7 @@ class SyncFilterTestCase(unittest.HomeserverTestCase): content={ "msgtype": "m.text", "body": "with right label", - LabelsField: ["#fun"], + EventContentFields.Labels: ["#fun"], }, tok=tok, ) -- cgit 1.5.1 From 57cdb046e48c6837fc9b41ade9b06d3ff68ec91b Mon Sep 17 00:00:00 2001 From: Brendan Abolivier Date: Fri, 1 Nov 2019 10:39:14 +0000 Subject: Lint --- synapse/api/constants.py | 1 + synapse/storage/data_stores/main/events.py | 2 +- 2 files changed, 2 insertions(+), 1 deletion(-) (limited to 'synapse/api/constants.py') diff --git a/synapse/api/constants.py b/synapse/api/constants.py index cf4ce5f5a2..066ce18704 100644 --- a/synapse/api/constants.py +++ b/synapse/api/constants.py @@ -142,5 +142,6 @@ class LimitBlockingTypes(object): class EventContentFields(object): """Fields found in events' content, regardless of type.""" + # Labels for the event, cf https://github.com/matrix-org/matrix-doc/pull/2326 Labels = "org.matrix.labels" diff --git a/synapse/storage/data_stores/main/events.py b/synapse/storage/data_stores/main/events.py index 42ffa9066a..0480161056 100644 --- a/synapse/storage/data_stores/main/events.py +++ b/synapse/storage/data_stores/main/events.py @@ -29,7 +29,7 @@ from prometheus_client import Counter, Histogram from twisted.internet import defer import synapse.metrics -from synapse.api.constants import EventTypes, EventContentFields +from synapse.api.constants import EventContentFields, EventTypes from synapse.api.errors import SynapseError from synapse.events import EventBase # noqa: F401 from synapse.events.snapshot import EventContext # noqa: F401 -- cgit 1.5.1 From 988d8d6507a0e8b34f2c352c77b5742197762190 Mon Sep 17 00:00:00 2001 From: Brendan Abolivier Date: Fri, 1 Nov 2019 16:22:44 +0000 Subject: Incorporate review --- synapse/api/constants.py | 2 +- synapse/api/filtering.py | 2 +- synapse/storage/data_stores/main/events.py | 2 +- synapse/storage/data_stores/main/schema/delta/56/event_labels.sql | 6 ++++++ tests/api/test_filtering.py | 8 ++++---- tests/rest/client/v1/test_rooms.py | 8 ++++---- tests/rest/client/v2_alpha/test_sync.py | 8 ++++---- 7 files changed, 21 insertions(+), 15 deletions(-) (limited to 'synapse/api/constants.py') diff --git a/synapse/api/constants.py b/synapse/api/constants.py index 066ce18704..49c4b85054 100644 --- a/synapse/api/constants.py +++ b/synapse/api/constants.py @@ -144,4 +144,4 @@ class EventContentFields(object): """Fields found in events' content, regardless of type.""" # Labels for the event, cf https://github.com/matrix-org/matrix-doc/pull/2326 - Labels = "org.matrix.labels" + LABELS = "org.matrix.labels" diff --git a/synapse/api/filtering.py b/synapse/api/filtering.py index 30a7ee0a7a..bec13f08d8 100644 --- a/synapse/api/filtering.py +++ b/synapse/api/filtering.py @@ -309,7 +309,7 @@ class Filter(object): content = event.get("content", {}) # check if there is a string url field in the content for filtering purposes contains_url = isinstance(content.get("url"), text_type) - labels = content.get(EventContentFields.Labels, []) + labels = content.get(EventContentFields.LABELS, []) return self.check_fields(room_id, sender, ev_type, labels, contains_url) diff --git a/synapse/storage/data_stores/main/events.py b/synapse/storage/data_stores/main/events.py index 577e79bcf9..1045c7fa2e 100644 --- a/synapse/storage/data_stores/main/events.py +++ b/synapse/storage/data_stores/main/events.py @@ -1491,7 +1491,7 @@ class EventsStore( self._handle_event_relations(txn, event) # Store the labels for this event. - labels = event.content.get(EventContentFields.Labels) + labels = event.content.get(EventContentFields.LABELS) if labels: self.insert_labels_for_event_txn( txn, event.event_id, labels, event.room_id, event.depth diff --git a/synapse/storage/data_stores/main/schema/delta/56/event_labels.sql b/synapse/storage/data_stores/main/schema/delta/56/event_labels.sql index 2acd8e1be5..5e29c1da19 100644 --- a/synapse/storage/data_stores/main/schema/delta/56/event_labels.sql +++ b/synapse/storage/data_stores/main/schema/delta/56/event_labels.sql @@ -13,6 +13,8 @@ * limitations under the License. */ +-- room_id and topoligical_ordering are denormalised from the events table in order to +-- make the index work. CREATE TABLE IF NOT EXISTS event_labels ( event_id TEXT, label TEXT, @@ -21,4 +23,8 @@ CREATE TABLE IF NOT EXISTS event_labels ( PRIMARY KEY(event_id, label) ); + +-- This index enables an event pagination looking for a particular label to index the +-- event_labels table first, which is much quicker than scanning the events table and then +-- filtering by label, if the label is rarely used relative to the size of the room. CREATE INDEX event_labels_room_id_label_idx ON event_labels(room_id, label, topological_ordering); diff --git a/tests/api/test_filtering.py b/tests/api/test_filtering.py index 8ec48c4154..2dc5052249 100644 --- a/tests/api/test_filtering.py +++ b/tests/api/test_filtering.py @@ -329,7 +329,7 @@ class FilteringTestCase(unittest.TestCase): sender="@foo:bar", type="m.room.message", room_id="!secretbase:unknown", - content={EventContentFields.Labels: ["#fun"]}, + content={EventContentFields.LABELS: ["#fun"]}, ) self.assertTrue(Filter(definition).check(event)) @@ -338,7 +338,7 @@ class FilteringTestCase(unittest.TestCase): sender="@foo:bar", type="m.room.message", room_id="!secretbase:unknown", - content={EventContentFields.Labels: ["#notfun"]}, + content={EventContentFields.LABELS: ["#notfun"]}, ) self.assertFalse(Filter(definition).check(event)) @@ -349,7 +349,7 @@ class FilteringTestCase(unittest.TestCase): sender="@foo:bar", type="m.room.message", room_id="!secretbase:unknown", - content={EventContentFields.Labels: ["#fun"]}, + content={EventContentFields.LABELS: ["#fun"]}, ) self.assertFalse(Filter(definition).check(event)) @@ -358,7 +358,7 @@ class FilteringTestCase(unittest.TestCase): sender="@foo:bar", type="m.room.message", room_id="!secretbase:unknown", - content={EventContentFields.Labels: ["#notfun"]}, + content={EventContentFields.LABELS: ["#notfun"]}, ) self.assertTrue(Filter(definition).check(event)) diff --git a/tests/rest/client/v1/test_rooms.py b/tests/rest/client/v1/test_rooms.py index 0dc0faa0e5..5e38fd6ced 100644 --- a/tests/rest/client/v1/test_rooms.py +++ b/tests/rest/client/v1/test_rooms.py @@ -860,7 +860,7 @@ class RoomMessageListTestCase(RoomBase): content={ "msgtype": "m.text", "body": "with right label", - EventContentFields.Labels: ["#fun"], + EventContentFields.LABELS: ["#fun"], }, ) @@ -876,7 +876,7 @@ class RoomMessageListTestCase(RoomBase): content={ "msgtype": "m.text", "body": "with wrong label", - EventContentFields.Labels: ["#work"], + EventContentFields.LABELS: ["#work"], }, ) @@ -886,7 +886,7 @@ class RoomMessageListTestCase(RoomBase): content={ "msgtype": "m.text", "body": "with two wrong labels", - EventContentFields.Labels: ["#work", "#notfun"], + EventContentFields.LABELS: ["#work", "#notfun"], }, ) @@ -896,7 +896,7 @@ class RoomMessageListTestCase(RoomBase): content={ "msgtype": "m.text", "body": "with right label", - EventContentFields.Labels: ["#fun"], + EventContentFields.LABELS: ["#fun"], }, ) diff --git a/tests/rest/client/v2_alpha/test_sync.py b/tests/rest/client/v2_alpha/test_sync.py index c3c6f75ced..3283c0e47b 100644 --- a/tests/rest/client/v2_alpha/test_sync.py +++ b/tests/rest/client/v2_alpha/test_sync.py @@ -157,7 +157,7 @@ class SyncFilterTestCase(unittest.HomeserverTestCase): content={ "msgtype": "m.text", "body": "with right label", - EventContentFields.Labels: ["#fun"], + EventContentFields.LABELS: ["#fun"], }, tok=tok, ) @@ -175,7 +175,7 @@ class SyncFilterTestCase(unittest.HomeserverTestCase): content={ "msgtype": "m.text", "body": "with wrong label", - EventContentFields.Labels: ["#work"], + EventContentFields.LABELS: ["#work"], }, tok=tok, ) @@ -186,7 +186,7 @@ class SyncFilterTestCase(unittest.HomeserverTestCase): content={ "msgtype": "m.text", "body": "with two wrong labels", - EventContentFields.Labels: ["#work", "#notfun"], + EventContentFields.LABELS: ["#work", "#notfun"], }, tok=tok, ) @@ -197,7 +197,7 @@ class SyncFilterTestCase(unittest.HomeserverTestCase): content={ "msgtype": "m.text", "body": "with right label", - EventContentFields.Labels: ["#fun"], + EventContentFields.LABELS: ["#fun"], }, tok=tok, ) -- cgit 1.5.1