API Reference Overview

The API docs are auto-generated from the codes package using Sphinx autodoc. Use the links below to jump directly into each area:

• codes.benchmark package — command-line entry points, benchmarking utilities, plotting helpers, and dataset utilities.

• codes.surrogates package — shared base classes plus module pages for every surrogate family (MultiONet, FCNN, LatentNeuralODE, LatentPoly, and helpers). Each surrogate package exposes its config dataclass, training loop, and evaluation helpers.

• codes.train package — task orchestration, queueing logic, and progress reporting for the training CLI.

• codes.tune package — Optuna integration, objective definitions, database utilities, and plotting for study analysis.

• codes.utils package — shared utilities for loading datasets, normalization, checkpointing, and miscellaneous helpers reused throughout the framework.

Each page mirrors the Python package tree and lists modules in alphabetical order. Collapse/expand sections in the sidebar to navigate classes, functions, and dataclasses quickly. When extending the framework, start from codes.surrogates to see how existing models integrate with AbstractSurrogateModel, or browse codes.benchmark to discover which helpers power the CLI scripts.