triage icon indicating copy to clipboard operation
triage copied to clipboard

General Purpose Risk Modeling and Prediction Toolkit for Policy and Social Good Problems

Results 140 triage issues
Sort by recently updated
recently updated
newest added

The postgres service in the github test suite is unnecessary: https://github.com/dssg/triage/blob/ab052a7a09a430e905823fb5cf4b47cb2ffbc361/.github/workflows/test.yaml#L9-L22 `testing.postgresql` initializes a local database. The service is created but goes unused. I ran `test.yaml` without the service and...

Looks like it will need a `pip install` of triage in addition to checking out the repo to build the API docs: ``` ERROR - mkdocstrings: Importing 'triage.component.audition.__init__' failed, possible...

documentation

black is still in requirement/dev.txt. Is it actually being used any more? I recall the discussion in #470 concluding with not using it going forward

dependencies

This may involve some considerable refactoring, especially of unit tests, but by way of documentation, here are a couple of notes to consider from discussion with @jtwalsh0 and @ecsalomon on...

dependencies

The additional estimators provided by catwalk are mentioned in the example experiment config, but should also be described in the documentation as well.

documentation
shovel-ready

add support for baselines that include scorecard type models: - give a number of points for a feature being a certain value (or threshold) - either sort by the sum...

shovel-ready

We've been using SQLAlchemy text type for UUID columns rather than Postgres's native UUID column type. (See [here](https://github.com/dssg/triage/blob/f15b86a15484663e948255a5b2251f1c41a38595/src/triage/component/results_schema/schema.py#L169) for an example.) The latter is 1) more efficient and 2) more...

More model types than LR should have scaled features, and scaling can aid in postmodeling analysis for all model types. Abstract the scaler that @shaycrk wrote and allow it to...

Our decision not to persist features limits the flexibility and reproducibility of the system. Triage is designed for batch processing, which means that we _could_ follow [functional data engineering principles](https://maximebeauchemin.medium.com/functional-data-engineering-a-modern-paradigm-for-batch-data-processing-2327ec32c42a)...

https://github.com/dssg/triage/blob/fa8d576c548b35d89dd4550ebb053ad6bfb37a38/src/triage/component/postmodeling/contrast/model_group_evaluator.py#L233