search

dssg / triage

General Purpose Risk Modeling and Prediction Toolkit for Policy and Social Good Problems
Other
187 stars 61 forks source link

Deleting all entries pertaining to a particular experiment_run #803

Open hemanklamba opened 4 years ago

hemanklamba commented 4 years ago

This functionality might be good to make control the size of predictions table. Might also help in removing unwanted (buggy) runs altogether from the database.

rayidghani commented 3 years ago

listing out which tables to drop rows from. we could do it everywhere but here's a starting point

  1. triage_metadata.models
  2. triage_metadata.model_groups
  3. train_results.predictions
  4. test_results.predictions
  5. train_results.evaluations
  6. test_results.evaluations
  7. train_results.feature_importances
  8. train_results.aequitas
  9. test_results.aequitas

other high priority ones? should we log when this happens in case we find orphan model information later on?