hardhat

hardhat is a developer-focused R package designed to simplify the creation of new modeling packages while promoting consistent standards for R modeling interfaces. It provides infrastructure for preprocessing data at both fit time and prediction time.

The package offers four main capabilities: consistent data preprocessing with mold() and forge() functions, input validation to ensure prediction data matches training data structure, utility functions for common modeling tasks like adding intercepts and standardizing outputs, and a stricter reimagining of base R’s preprocessing functions. This standardization reduces the implementation burden on package developers and creates predictable interfaces that benefit end users.