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.
conf_source shouldn't be Path.cwd() but pointing to conf
base_env & default_run_env are missing.
This solution works but the complication is we introduce these "default" argument which is invisible to Kedro Framework user. Should we do this or instead use KedroSession? IMO KedroSession will be more consistent and it respect all the custom settings built in settings.py.
We may have to update some documentation in the test as well.
Your Environment
Kedro version used (pip show kedro or kedro -V): 0.19.6
Description
Related: https://github.com/kedro-org/kedro/issues/3954 Slack: https://linen-slack.kedro.org/t/22301026/question-is-there-an-extra-step-required-to-run-pytest-with-#661646b5-a2e7-4b77-a3db-ab585f408793
https://github.com/kedro-org/kedro-starters/blob/main/spaceflights-pandas-viz/%7B%7B%20cookiecutter.repo_name%20%7D%7D/tests/test_run.py#L27
The config loader is using incorrect arguments.
Context
Steps to Reproduce
Try to print out
parameters
and find that it loads incorrrect config (use with globals. for example)Expected Result
The default test should have similar behavior as
kedro new
by defaultPotential Solutions:
OmegaConfigLoader
with the correct arguments, i.e.The previous test has 2 mistakes:
Path.cwd()
but pointing toconf
This solution works but the complication is we introduce these "default" argument which is invisible to Kedro Framework user. Should we do this or instead use
KedroSession
? IMOKedroSession
will be more consistent and it respect all the custom settings built insettings.py
.We may have to update some documentation in the test as well.
Your Environment
pip show kedro
orkedro -V
): 0.19.6python -V
) 3.10