Is it somehow possible to subclass

SQLStream

to create a "well-known" stream that defines a

replication_key

that is respected (as in standard `RestStream`/`Stream` implementations)?

client.py

class TestStream(SQLStream):
    connector_class = TestConnector

class AStream(TestStream):
    name = "a"
    replication_key = "updated_at"

class BStream(TestStream):
    name = "b"
    replication_key = "updated_at"

tap.py

KNOWN_STREAMS = {
  "a": stream.AStream,
  "b": streams.BStream,
}


...


    def discover_streams(self):
        for stream in super().discover_streams():
            if stream_class := KNOWN_STREAMS.get(stream.name):
                stream = stream_class(
                    tap=self,
                    catalog_entry=stream.catalog_entry,
                    connector=self.tap_connector,
                )

            yield stream

I've played around with this and a few other variations, but can't seem to get it to work - the sync runs as if no replication key was defined. Is there another (possibly simpler) way to do this?