utils: Utility Functions

This module provides a collection of helper functions for data processing, model training, and other tasks.

• data: Contains the Dataset class, which handles loading, preprocessing, and sampling of data from various sources, including GeoTIFF files and synthetic data generation.

• gpflow: Provides utility functions for training GPflow models, such as get_model_params for finding optimal kernel hyperparameters and optimize_model for running the optimization. It also includes a TraceInducingPts class to monitor the movement of inducing points during training.

• misc: A collection of miscellaneous helper functions, including get_inducing_pts for selecting initial inducing points (via k-means or random sampling), cont2disc for mapping continuous solutions to a discrete set of candidates, and polygon2candidates for sampling points within a polygon.

• tsp: Provides functionality for solving the Traveling Salesperson Problem (TSP) using Google's OR-Tools. The run_tsp function can find effient paths for single or multiple vehicles, with optional start and end nodes, and can resample the resulting paths to have a fixed number of points.