Using the CLI Tool

The Chap CLI provides commands for evaluating disease prediction models, visualizing results, and exporting metrics.

Quick Start

The main workflow consists of three commands:

1. chap eval - Run a backtest and export results to NetCDF format
2. chap plot-backtest - Generate visualizations from evaluation results
3. chap export-metrics - Export and compare metrics across evaluations

See the Evaluation Workflow guide for detailed usage and examples.

To set up a model and dataset once and then run chap eval without options, create a Chap project with chap init.

Documentation

• Setup - How to install and configure the CLI
• Evaluation Workflow - Complete guide to evaluating and comparing models
• Chap Projects - Keep a model, dataset and settings in a directory and run chap eval without options

Command Reference

• eval - Full reference for the eval command with all parameters
• report - Generate a PDF report from a trained model
• explain-lime - Generate a LIME importance weighting for a single prediction
• simulate - Generate a synthetic dataset from a simulation model
• sanity-check-model - Check quickly that a model can be trained and make forecasts
• plot-dataset - Plot a dataset's columns over time, in a browser or to a file