parrallize the map/reduce chain

[hwchase17]

make the calls in the "map" part concurrently

[vshulman]

might be helpful to also make the intermediate outputs of the map step accessible to the caller

[devonbrackbill]

I understand the map/reduce chain to currently do 1 step of map and then "stuff" all of the results into 1 final query. I'm running into token limits on the final "stuff". Is there an obvious way to call map on the intermediate summaries until the input to the final "stuff" is below the token threshold?

[hwchase17]

sorry @devonbrackbill i did not see this earlier! i think this should be addressed now - could you confirm?

[devonbrackbill]

Yes, this is addressed. Thank you