From 1f6d6f918a57183083c2dd1bba6179373102b918 Mon Sep 17 00:00:00 2001
From: Richard van der Hoff <richard@matrix.org>
Date: Wed, 27 Mar 2019 15:18:28 +0000
Subject: Make EventStream rows have a type

... as a precursor to combining it with the CurrentStateDelta stream.
---
 synapse/replication/tcp/streams/events.py | 98 +++++++++++++++++++++++++++----
 1 file changed, 86 insertions(+), 12 deletions(-)

(limited to 'synapse/replication/tcp/streams')

diff --git a/synapse/replication/tcp/streams/events.py b/synapse/replication/tcp/streams/events.py
index 511dd6bcc7..928028e893 100644
--- a/synapse/replication/tcp/streams/events.py
+++ b/synapse/replication/tcp/streams/events.py
@@ -13,28 +13,102 @@
 # 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.
-from collections import namedtuple
+
+import attr
+
+from twisted.internet import defer
 
 from ._base import Stream
 
-EventStreamRow = namedtuple("EventStreamRow", (
-    "event_id",  # str
-    "room_id",  # str
-    "type",  # str
-    "state_key",  # str, optional
-    "redacts",  # str, optional
-))
+
+"""Handling of the 'events' replication stream
+
+This stream contains rows of various types. Each row therefore contains a 'type'
+identifier before the real data. For example::
+
+    RDATA events 12345 ["ev", ["$event:id", "!room:id", "m.type", null, null]]
+
+An "ev" row is sent for each new event. The fields in the data part are:
+
+ * The new event id
+ * The room id for the event
+ * The type of the new event
+ * The state key of the event, for state events
+ * The event id of an event which is redacted by this event.
+
+"""
+
+
+@attr.s(slots=True, frozen=True)
+class EventsStreamRow(object):
+    """A parsed row from the events replication stream"""
+    type = attr.ib()  # str: the TypeId of one of the *EventsStreamRows
+    data = attr.ib()  # BaseEventsStreamRow
+
+
+class BaseEventsStreamRow(object):
+    """Base class for rows to be sent in the events stream.
+
+    Specifies how to identify, serialize and deserialize the different types.
+    """
+
+    TypeId = None  # Unique string that ids the type. Must be overriden in sub classes.
+
+    @classmethod
+    def from_data(cls, data):
+        """Parse the data from the replication stream into a row.
+
+        By default we just call the constructor with the data list as arguments
+
+        Args:
+            data: The value of the data object from the replication stream
+        """
+        return cls(*data)
+
+
+@attr.s(slots=True, frozen=True)
+class EventsStreamEventRow(BaseEventsStreamRow):
+    TypeId = "ev"
+
+    event_id = attr.ib()   # str
+    room_id = attr.ib()    # str
+    type = attr.ib()       # str
+    state_key = attr.ib()  # str, optional
+    redacts = attr.ib()    # str, optional
+
+
+TypeToRow = {
+    Row.TypeId: Row
+    for Row in (
+        EventsStreamEventRow,
+    )
+}
 
 
 class EventsStream(Stream):
     """We received a new event, or an event went from being an outlier to not
     """
     NAME = "events"
-    ROW_TYPE = EventStreamRow
 
     def __init__(self, hs):
-        store = hs.get_datastore()
-        self.current_token = store.get_current_events_token
-        self.update_function = store.get_all_new_forward_event_rows
+        self._store = hs.get_datastore()
+        self.current_token = self._store.get_current_events_token
 
         super(EventsStream, self).__init__(hs)
+
+    @defer.inlineCallbacks
+    def update_function(self, from_token, current_token, limit=None):
+        event_rows = yield self._store.get_all_new_forward_event_rows(
+            from_token, current_token, limit,
+        )
+        event_updates = (
+            (row[0], EventsStreamEventRow.TypeId, row[1:])
+            for row in event_rows
+        )
+        defer.returnValue(event_updates)
+
+    @classmethod
+    def parse_row(cls, row):
+        (typ, data) = row
+        data = TypeToRow[typ].from_data(data)
+        return EventsStreamRow(typ, data)
-- 
cgit 1.4.1