Implement order and group by

1 files changed, 96 insertions, 0 deletions

diff --git a/synapse/storage/search.py b/synapse/storage/search.py
index cdf003502f..e37e56c1f2 100644
--- a/synapse/storage/search.py
+++ b/synapse/storage/search.py
@@ -20,6 +20,12 @@ from synapse.storage.engines import PostgresEngine, Sqlite3Engine
 
 from collections import namedtuple
 
+import logging
+
+
+logger = logging.getLogger(__name__)
+
+
 """The result of a search.
 
 Fields:
@@ -109,3 +115,93 @@ class SearchStore(SQLBaseStore):
             event_map,
             None
         ))
+
+    @defer.inlineCallbacks
+    def search_room(self, room_id, search_term, keys, limit, pagination_token=None):
+        """Performs a full text search over events with given keys.
+
+        Args:
+            room_id (str): The room_id to search in
+            search_term (str): Search term to search for
+            keys (list): List of keys to search in, currently supports
+                "content.body", "content.name", "content.topic"
+            pagination_token (str): A pagination token previously returned
+
+        Returns:
+            SearchResult
+        """
+        clauses = []
+        args = [search_term, room_id]
+
+        local_clauses = []
+        for key in keys:
+            local_clauses.append("key = ?")
+            args.append(key)
+
+        clauses.append(
+            "(%s)" % (" OR ".join(local_clauses),)
+        )
+
+        if pagination_token:
+            topo, stream = pagination_token.split(",")
+            clauses.append(
+                "(topological_ordering < ?"
+                " OR (topological_ordering = ? AND stream_ordering < ?))"
+            )
+            args.extend([topo, topo, stream])
+
+        if isinstance(self.database_engine, PostgresEngine):
+            sql = (
+                "SELECT ts_rank_cd(vector, query) as rank,"
+                " topological_ordering, stream_ordering, room_id, event_id"
+                " FROM plainto_tsquery('english', ?) as query, event_search"
+                " NATURAL JOIN events"
+                " WHERE vector @@ query AND room_id = ?"
+            )
+        elif isinstance(self.database_engine, Sqlite3Engine):
+            sql = (
+                "SELECT rank(matchinfo(event_search)) as rank, room_id, event_id"
+                " topological_ordering, stream_ordering"
+                " FROM event_search"
+                " NATURAL JOIN events"
+                " WHERE value MATCH ? AND room_id = ?"
+            )
+        else:
+            # This should be unreachable.
+            raise Exception("Unrecognized database engine")
+
+        for clause in clauses:
+            sql += " AND " + clause
+
+        # We add an arbitrary limit here to ensure we don't try to pull the
+        # entire table from the database.
+        sql += " ORDER BY topological_ordering DESC, stream_ordering DESC LIMIT ?"
+
+        args.append(limit)
+
+        results = yield self._execute(
+            "search_rooms", self.cursor_to_dict, sql, *args
+        )
+
+        events = yield self._get_events([r["event_id"] for r in results])
+
+        event_map = {
+            ev.event_id: ev
+            for ev in events
+        }
+
+        pagination_token = None
+        if results:
+            topo = results[-1]["topological_ordering"]
+            stream = results[-1]["stream_ordering"]
+            pagination_token = "%s,%s" % (topo, stream)
+
+        defer.returnValue(SearchResult(
+            {
+                r["event_id"]: r["rank"]
+                for r in results
+                if r["event_id"] in event_map
+            },
+            event_map,
+            pagination_token
+        ))