GA
Galileo-Galilei/kedro_mlflow_624
Investigate the bug https://github.com/Galileo-Galilei/kedro-mlflow/issues/624
spaceflights-pandas
This is the default spaceflights-pandas kedro starter, installed with
pip install uv
uv tool install kedro
uvx kedro new -s spaceflights-pandas -n spaceflights-pandas --telemetry no
cd spaceflights-pandas
uv venv
uv pip install -e .
I've added a custom hook in spaceflights_pandas/hooks.py which simply starts mlflow and registered it in settings.py
Results
Run the defaut pipeline ith kedro run and see this surprising result :
Conclusion : the after_context_created starts running in a different thread the ones where the node run, even whe using sequential runner.
