Closed rahulbot closed 1 month ago
Realizing that we compute aggregations for "significant" and "rare" terms as well as for "top"- my instinct is that we could excise them, as we never use them on the frontend
Yeah - we included those early to experiment with different results. The "top" seems most useful and stable, so perhaps we should just commit to those.
Also, the api.py file is really bloated- I'm thinking it would be a worthwhile improvement in the developer experience if the the fastapi views and the actual elasticsearch logic were separated into two files
Feeling like this is probably ready to close after the most recent round of polish I did
This code has seen a few different authors at this point, and been forked and reassesed with some changing assumptions. I think we need to revisit this code to clean it up in a few ways: