Kedro is a toolbox for production-ready data science. It uses software engineering best practices to help you create data engineering and data science pipelines that are reproducible, maintainable, and modular.
Unit tests for the tools selection have become hard to maintain see e.g. #3594, because framework needs to pull the starters. If we can somehow decouple the tools selection process better the tests might become simpler too.
Description
With the addition of the tools flow some coupling between framework and the starters was added in the post_gen_project scripts: see https://github.com/kedro-org/kedro-starters/blob/main/spaceflights-pyspark/hooks/post_gen_project.py
Unit tests for the tools selection have become hard to maintain see e.g. #3594, because framework needs to pull the starters. If we can somehow decouple the tools selection process better the tests might become simpler too.
Context
See https://github.com/kedro-org/kedro/issues/3594#issuecomment-1927510672
Possible Implementation
The checks here: https://github.com/kedro-org/kedro-starters/blob/main/spaceflights-pyspark/hooks/post_gen_project.py#L30 shouldn't be needed because at this point the tool selection should already have been checked and verified on the Kedro side. This might mean we can move the
main
function to Kedro.