search

stefan-grafberger / mlwhatif

Data-Centric What-If Analysis for Native Machine Learning Pipelines
Apache License 2.0
14 stars 4 forks source link

Make Debugging Easier if DAG execution crashes #23

Open stefan-grafberger opened 1 year ago

stefan-grafberger commented 1 year ago

Description

We can add a few strategic try-catch statements and throw exceptions that include the necessary data to visualize the DAGs for debugging purposes.

stefan-grafberger commented 1 year ago

Note: previously, we didn't have this problem because the DAG generation code was called directly in the analysis execution code. However, this was refactored to reduce the complexity of the analysis execution code, and because it is cleaner to have the expensive (if used to visualize higher number of DAGs) visualization calls in a separate place.

stefan-grafberger commented 1 year ago

(DAG execution can only crash if there is a bug in mlwhat.)