summary refs log tree commit diff
path: root/synapse/storage/search.py
diff options
context:
space:
mode:
authorErik Johnston <erik@matrix.org>2015-11-11 17:51:14 +0000
committerErik Johnston <erik@matrix.org>2015-11-11 17:51:14 +0000
commit6341be45c6ceaa1b01a0b98ac6e24d2684841a26 (patch)
treee3bd4addb574125ee81931c20feb57c0712d46b0 /synapse/storage/search.py
parentRemove anonymous access, since its not ready yet (diff)
parentMerge pull request #359 from matrix-org/markjh/incremental_indexing (diff)
downloadsynapse-6341be45c6ceaa1b01a0b98ac6e24d2684841a26.tar.xz
Merge branch 'develop' of github.com:matrix-org/synapse into release-v0.11.0
Diffstat (limited to 'synapse/storage/search.py')
-rw-r--r--synapse/storage/search.py105
1 files changed, 102 insertions, 3 deletions
diff --git a/synapse/storage/search.py b/synapse/storage/search.py
index 3cea2011fa..2e88c51ad0 100644
--- a/synapse/storage/search.py
+++ b/synapse/storage/search.py
@@ -15,7 +15,7 @@
 
 from twisted.internet import defer
 
-from _base import SQLBaseStore
+from .background_updates import BackgroundUpdateStore
 from synapse.api.errors import SynapseError
 from synapse.storage.engines import PostgresEngine, Sqlite3Engine
 
@@ -25,7 +25,106 @@ import logging
 logger = logging.getLogger(__name__)
 
 
-class SearchStore(SQLBaseStore):
+class SearchStore(BackgroundUpdateStore):
+
+    EVENT_SEARCH_UPDATE_NAME = "event_search"
+
+    def __init__(self, hs):
+        super(SearchStore, self).__init__(hs)
+        self.register_background_update_handler(
+            self.EVENT_SEARCH_UPDATE_NAME, self._background_reindex_search
+        )
+
+    @defer.inlineCallbacks
+    def _background_reindex_search(self, progress, batch_size):
+        target_min_stream_id = progress["target_min_stream_id_inclusive"]
+        max_stream_id = progress["max_stream_id_exclusive"]
+        rows_inserted = progress.get("rows_inserted", 0)
+
+        INSERT_CLUMP_SIZE = 1000
+        TYPES = ["m.room.name", "m.room.message", "m.room.topic"]
+
+        def reindex_search_txn(txn):
+            sql = (
+                "SELECT stream_ordering, event_id FROM events"
+                " WHERE ? <= stream_ordering AND stream_ordering < ?"
+                " AND (%s)"
+                " ORDER BY stream_ordering DESC"
+                " LIMIT ?"
+            ) % (" OR ".join("type = '%s'" % (t,) for t in TYPES),)
+
+            txn.execute(sql, (target_min_stream_id, max_stream_id, batch_size))
+
+            rows = txn.fetchall()
+            if not rows:
+                return 0
+
+            min_stream_id = rows[-1][0]
+            event_ids = [row[1] for row in rows]
+
+            events = self._get_events_txn(txn, event_ids)
+
+            event_search_rows = []
+            for event in events:
+                try:
+                    event_id = event.event_id
+                    room_id = event.room_id
+                    content = event.content
+                    if event.type == "m.room.message":
+                        key = "content.body"
+                        value = content["body"]
+                    elif event.type == "m.room.topic":
+                        key = "content.topic"
+                        value = content["topic"]
+                    elif event.type == "m.room.name":
+                        key = "content.name"
+                        value = content["name"]
+                except (KeyError, AttributeError):
+                    # If the event is missing a necessary field then
+                    # skip over it.
+                    continue
+
+                event_search_rows.append((event_id, room_id, key, value))
+
+            if isinstance(self.database_engine, PostgresEngine):
+                sql = (
+                    "INSERT INTO event_search (event_id, room_id, key, vector)"
+                    " VALUES (?,?,?,to_tsvector('english', ?))"
+                )
+            elif isinstance(self.database_engine, Sqlite3Engine):
+                sql = (
+                    "INSERT INTO event_search (event_id, room_id, key, value)"
+                    " VALUES (?,?,?,?)"
+                )
+            else:
+                # This should be unreachable.
+                raise Exception("Unrecognized database engine")
+
+            for index in range(0, len(event_search_rows), INSERT_CLUMP_SIZE):
+                clump = event_search_rows[index:index + INSERT_CLUMP_SIZE]
+                txn.executemany(sql, clump)
+
+            progress = {
+                "target_min_stream_id_inclusive": target_min_stream_id,
+                "max_stream_id_exclusive": min_stream_id,
+                "rows_inserted": rows_inserted + len(event_search_rows)
+            }
+
+            self._background_update_progress_txn(
+                txn, self.EVENT_SEARCH_UPDATE_NAME, progress
+            )
+
+            return len(event_search_rows)
+
+        result = yield self.runInteraction(
+            self.EVENT_SEARCH_UPDATE_NAME, reindex_search_txn
+        )
+
+        if not result:
+            yield self._end_background_update(self.EVENT_SEARCH_UPDATE_NAME)
+
+        defer.returnValue(result)
+
     @defer.inlineCallbacks
     def search_msgs(self, room_ids, search_term, keys):
         """Performs a full text search over events with given keys.
@@ -154,7 +253,7 @@ class SearchStore(SQLBaseStore):
             )
         elif isinstance(self.database_engine, Sqlite3Engine):
             sql = (
-                "SELECT rank(matchinfo(event_search)) as rank, room_id, event_id"
+                "SELECT rank(matchinfo(event_search)) as rank, room_id, event_id,"
                 " topological_ordering, stream_ordering"
                 " FROM event_search"
                 " NATURAL JOIN events"