Running queries from the Python API currently works as follows:

• Query is sent to the query SQS queue where it is picked up by a lambda and then the results written to S3.
• Whilst this is happening the Python client is polling Dynamo to see if the query has finished.
• When the query is finished, the results are read from S3 into a list.

The need to wait for the results to have been written to S3 means that the queries are slower than they could be. This could be improved by using the same websockets approach as is used in QueryWebSocketClient.