Proof of concept for a query execution service which could:
- [-] execute queries from diverse sources
- from local files
- from SQL databases
- from cloud storages (S3)
- from online services (HTTP APIs)
- combine queries from multiple sources
- push as much work as possible upstream (to the source DB) and execute the rest in-memory
- track time and resources used by query
- handle many concurrent clients
- limit memory usage per client
- save queries to a transient cache
- execute data processing jobs and stream the result to disk
poetry install
poetry run pytest
poetry run uvicorn api.main:api