Improvements from Cass project

[rvandewater]

Summary by CodeRabbit

• New Features

  • HPC benchmarking configs & SLURM scripts; GPU XGBoost option; alarm-style prediction conversion; CLI dict-style dataset/file parsing; toggleable feature-explanation with explainer-value saving; per-sample prediction indicators export; calibration APIs for model probabilities.

• Enhancements

  • Expanded hyperparameter search spaces and tuning repetitions; improved preprocessing (modality selection, infinite-value handling, row-indicators); new classification metrics (sensitivity, specificity, PPV, incidence); aggregation of fold-wise predictions.

• Documentation

  • Added full Sphinx docs, API reference, and how-to guides.

• Chores

  • Dependency/version bumps, package version update, removed CI workflow.