spectral-library¶

Prepared Spectral Mapping

Map between sensors. Reconstruct spectra. Keep the runtime stable.¶

`spectral-library` is a Python package and CLI for preparing row-aligned mapping runtimes from SIAC-style hyperspectral exports, retrieving nearest spectral neighbors, simulating target-sensor reflectance, and reconstructing VNIR, SWIR, or full `400-2500 nm` spectra.

Get Started See Official Examples Read The Theory

What This Site Covers¶

The public documentation is organized around one practical flow: prepare a runtime, validate it, map reflectance, and understand the retrieval model well enough to trust the outputs.

1. Prepare

Build a compact prepared runtime from a SIAC-style spectral library and one or more sensor SRF definitions.

Prepared runtime workflow

2. Map

Run single-sample or batch mapping to a target sensor, or reconstruct VNIR, SWIR, or the full hyperspectral spectrum.

CLI mapping examples

3. Understand

Follow the exact forward model, retrieval distance, overlap blending, and benchmark definitions used by the implementation.

Mathematical foundations

Start Here¶

Getting StartedInstall the package, prepare a runtime, validate it, and run your first mappings. Official Sensor ExamplesMODIS, Sentinel-2A, Landsat 8, and Landsat 9 examples built from official response functions. Mathematical FoundationsThe equations and assumptions behind retrieval, target simulation, and overlap blending. Prepared Runtime ContractThe stable on-disk standard used by the mapper.

Public Documentation Map¶

Guides¶

Getting Started for installation, runtime preparation, validation, single-sample mapping, batch mapping, and minimal Python usage
Official Sensor Examples for reproducible examples and figures built from official MODIS, Sentinel-2A, Landsat 8, and Landsat 9 spectral response functions
Troubleshooting for common issues, confidence score interpretation, and KNN backend selection

Concepts¶

Mathematical Foundations for the sensor forward model, segment-wise nearest-neighbor retrieval, target-sensor simulation, and benchmark metrics
Prepared Runtime Contract for the stable files, schema rules, checksum behavior, and compatibility policy
Official Example Bundle for the bundled example data and provenance artifacts

Reference¶

CLI Reference for the public commands, required flags, CSV layouts, and outputs
Python API Reference for the stable imports, result objects, and public errors
FAQ for answers to common questions

Project¶

Package Facts¶

Distribution`spectral-library`

Import`spectral_library`

CLI`spectral-library`

Python`3.9` to `3.14`

Repository Layout¶

The repository code is split into four top-level package areas:

spectral_library.mapping public mapping runtime, prepared-runtime build, and retrieval engine
spectral_library.distribution runtime download helpers
spectral_library.sources maintainer-oriented source manifests, fetchers, fetch batching, and catalog assembly
spectral_library.normalization maintainer-oriented normalization, coverage filtering, quality plots, and SIAC package export

Public vs internal material

This site includes a small internal section for maintainers. If you are using the package, you can ignore the internal build/design pages and stay within the Start Here, Concepts, and Reference sections.