nirs4all-formats

Rust-first, low-level readers for NIRS and spectroscopy file formats, with matching Python / R / WebAssembly / C bindings and conformance checks against reference loaders.

nirs4all-formats auto-detects each file by content, decodes it through a single Rust registry, and returns a canonical, provenance-tracked SpectralRecord model. It does no modelling itself — it produces the records that nirs4all consumes.

Start here: Getting started · Installation · Usage guide · Supported formats

Getting started

• Getting started
  • 30 seconds, three ways
  • What you get back
  • Next steps
• Installation
  • Python
  • Rust
  • Command-line tool
  • R
  • WebAssembly / JavaScript
  • C ABI
• Usage guide
  • 1. Probe — which reader, and why?
  • 2. Read records
  • 3. Project to numpy / pandas / sklearn / torch
  • 4. Sidecar formats (companion files)
  • 5. Image cubes — pick pixels, not the whole scene
  • 6. In-memory bytes
  • 7. Scan a directory
• Supported formats
  • Status legend
  • Generic & tabular
  • FT-NIR / FT-IR / UV-Vis vendors
  • Field spectroscopy
  • Hyperspectral cubes
  • Sun photometers
  • HDF5, containers & standards
  • Exchange & standards-track
  • Raman & adjacent (decoded, adjacent to core NIRS)
  • Detected & deliberately refused (not NIRS)
  • Not yet supported — help us by sending samples
  • Don’t see your format?

Reference

• Data Model
  • SpectralAxis
  • SpectralArray
  • SpectralRecord
  • Binding Exports
• User Outputs
  • Principles
  • Level 1: non-destructive canonical model
  • Level 2: simple data science outputs
  • Level 3: outputs for complex data
  • File Exports
  • Recommended Bundle
  • Interface Surfaces
  • Product Priorities
• CLI Contract
• Conformance Policy
  • Strict Checks
  • Reference Readers (M2, 2026-05-23)
  • No Reference Case
• Integration With nirs4all
• Release Procedure
  • Pipeline overview
  • Tag-to-release flow
  • Dry-run flow
  • Rollback / yank
  • Per-wheel capability matrix
  • C ABI versioning
• Versioning
• Security Policy
  • Runtime Rules
  • Reporting
• License Matrix
  • Runtime Rule
  • Reference Reader Rule
  • Fixture Rule

Bindings

• Python Binding
  • Raw access
  • Lossless object model
  • Projections (explicit, possibly lossy)
  • Transport
  • Examples
• R Binding
  • Installation
  • Raw access
  • Flat dataset
  • Walking a directory
  • Bytes and sidecars (native only)
  • Transport
  • Example
• WebAssembly / JS Binding
  • Build & package
  • API
  • Example
  • Scope
  • Tests
• C ABI
  • API
  • Build & link
  • Example
  • Memory & safety
  • Roadmap

Format reference

• Allotrope ADF
• Allotrope ASM JSON
• ANDI/MS NetCDF
• AnIML XML
• ASD FieldSpec
• Avantes AvaSoft
• Bruker OPUS
• BUCHI NIRCal
• DigitalSurf SUR / PRO
• ENVI Spectral Library & Standard Cubes
• ERDAS LAN / AVIRIS (Indian Pines)
• Excel Spectral Tables
• Felix Instruments F-750
• FGI HDF5 + XML
• Foss / WinISI
• Thermo / Galactic GRAMS SPC
• Hamamatsu HPD-TA IMG
• Generic HDF5 NIRS Datasets
• Horiba LabSpec / JobinYvon
• JASCO JWS
• JCAMP-DX
• MATLAB MAT / R RData Datasets
• Metrohm Vision / Vision Air
• EMSA/MAS MSA (ISO 22029)
• mzML
• Generic NetCDF NIRS Datasets
• Thermo Nicolet OMNIC
• NumPy NPY / NPZ
• Ocean Optics / Ocean Insight
• OpenSpecy Raman / (FT)IR .rds
• Parquet NIRS Tables
• PerkinElmer Spectrum / IR
• PP Systems UniSpec SC / DC
• Renishaw WDF
• Row-Oriented Spectral Tables
• Consumer Physics SCiO CSV
• Shimadzu UVProbe
• Si-Ware API JSON / CSV
• Si-Ware NeoSpectra
• Spectral Evolution / PSR SED
• Spectral Matrix Exports
• Sun Photometer Text Exports
• SVC / GER SIG
• Delimited Spectral Tables (one spectrum per row)
• Princeton TriVista TVF
• USGS SPECPR / Spectral Library Text
• VIAVI MicroNIR
• WiTec WIP / WID

The nirs4all ecosystem

nirs4all

Main Python modelling library — pipelines, SpectroDataset, predictions.

https://nirs4all.readthedocs.io/en/latest/

nirs4all-methods

Portable C-ABI PLS/NIRS engine (libn4m) + bindings.

https://nirs4all-methods.readthedocs.io/en/latest/

nirs4all-io

Dataset-assembly bridge → SpectroDataset.

https://nirs4all-io.readthedocs.io/en/latest/

nirs4all-datasets

Curated DOI-pinned NIRS dataset catalog (n4a-datasets).

https://nirs4all-datasets.readthedocs.io/en/latest/

nirs4all-lite

Portable aggregate distribution (Rust + bindings).

https://nirs4all-lite.readthedocs.io/en/latest/

dag-ml

Reproducible, OOF/leakage-safe ML coordinator.

https://dag-ml.readthedocs.io/en/latest/

dag-ml-data

Typed sample-aligned multi-source data contracts.

https://dag-ml-data.readthedocs.io/en/latest/