Merge pull request #4415 from matrix-org/anoa/full_search_upgraded_rooms

Ability to search entire room history after upgrading room

3 files changed, 85 insertions, 0 deletions

diff --git a/synapse/api/filtering.py b/synapse/api/filtering.py
index 16ad654864..3906475403 100644
--- a/synapse/api/filtering.py
+++ b/synapse/api/filtering.py
@@ -444,6 +444,20 @@ class Filter(object):
     def include_redundant_members(self):
         return self.filter_json.get("include_redundant_members", False)
 
+    def with_room_ids(self, room_ids):
+        """Returns a new filter with the given room IDs appended.
+
+        Args:
+            room_ids (iterable[unicode]): The room_ids to add
+
+        Returns:
+            filter: A new filter including the given rooms and the old
+                    filter's rooms.
+        """
+        newFilter = Filter(self.filter_json)
+        newFilter.rooms += room_ids
+        return newFilter
+
 
 def _matches_wildcard(actual_value, filter_value):
     if filter_value.endswith("*"):
diff --git a/synapse/handlers/search.py b/synapse/handlers/search.py
index ec936bbb4e..49c439313e 100644
--- a/synapse/handlers/search.py
+++ b/synapse/handlers/search.py
@@ -38,6 +38,41 @@ class SearchHandler(BaseHandler):
         super(SearchHandler, self).__init__(hs)
 
     @defer.inlineCallbacks
+    def get_old_rooms_from_upgraded_room(self, room_id):
+        """Retrieves room IDs of old rooms in the history of an upgraded room.
+
+        We do so by checking the m.room.create event of the room for a
+        `predecessor` key. If it exists, we add the room ID to our return
+        list and then check that room for a m.room.create event and so on
+        until we can no longer find any more previous rooms.
+
+        The full list of all found rooms in then returned.
+
+        Args:
+            room_id (str): id of the room to search through.
+
+        Returns:
+            Deferred[iterable[unicode]]: predecessor room ids
+        """
+
+        historical_room_ids = []
+
+        while True:
+            predecessor = yield self.store.get_room_predecessor(room_id)
+
+            # If no predecessor, assume we've hit a dead end
+            if not predecessor:
+                break
+
+            # Add predecessor's room ID
+            historical_room_ids.append(predecessor["room_id"])
+
+            # Scan through the old room for further predecessors
+            room_id = predecessor["room_id"]
+
+        defer.returnValue(historical_room_ids)
+
+    @defer.inlineCallbacks
     def search(self, user, content, batch=None):
         """Performs a full text search for a user.
 
@@ -137,6 +172,18 @@ class SearchHandler(BaseHandler):
         )
         room_ids = set(r.room_id for r in rooms)
 
+        # If doing a subset of all rooms seearch, check if any of the rooms
+        # are from an upgraded room, and search their contents as well
+        if search_filter.rooms:
+            historical_room_ids = []
+            for room_id in search_filter.rooms:
+                # Add any previous rooms to the search if they exist
+                ids = yield self.get_old_rooms_from_upgraded_room(room_id)
+                historical_room_ids += ids
+
+            # Prevent any historical events from being filtered
+            search_filter = search_filter.with_room_ids(historical_room_ids)
+
         room_ids = search_filter.filter_rooms(room_ids)
 
         if batch_group == "room_id":
diff --git a/synapse/storage/state.py b/synapse/storage/state.py
index a134e9b3e8..c3ab7db7ae 100644
--- a/synapse/storage/state.py
+++ b/synapse/storage/state.py
@@ -437,6 +437,30 @@ class StateGroupWorkerStore(EventsWorkerStore, SQLBaseStore):
         create_event = yield self.get_event(create_id)
         defer.returnValue(create_event.content.get("room_version", "1"))
 
+    @defer.inlineCallbacks
+    def get_room_predecessor(self, room_id):
+        """Get the predecessor room of an upgraded room if one exists.
+        Otherwise return None.
+
+        Args:
+            room_id (str)
+
+        Returns:
+            Deferred[unicode|None]: predecessor room id
+        """
+        state_ids = yield self.get_current_state_ids(room_id)
+        create_id = state_ids.get((EventTypes.Create, ""))
+
+        # If we can't find the create event, assume we've hit a dead end
+        if not create_id:
+            defer.returnValue(None)
+
+        # Retrieve the room's create event
+        create_event = yield self.get_event(create_id)
+
+        # Return predecessor if present
+        defer.returnValue(create_event.content.get("predecessor", None))
+
     @cached(max_entries=100000, iterable=True)
     def get_current_state_ids(self, room_id):
         """Get the current state event ids for a room based on the