nirs4all.utils.hashing module

Fast, deterministic content hashing for numpy arrays.

Provides compute_data_hash which hashes the raw bytes of a numpy array using xxhash (preferred for speed) with a SHA-256 fallback.

nirs4all.utils.hashing.compute_data_hash(X: ndarray) → str

Compute a deterministic hex digest for a numpy array.

Uses xxhash (xxh128) when available for speed, otherwise falls back to SHA-256. The array is made contiguous before hashing so that the result is independent of memory layout.

Parameters:

X – Numpy array to hash.

Returns:

Hex digest string.