NIRSPY — NIRS Processing in Python

GUI-first, fNIRS-focused pipeline builder in Python. Runs locally in your browser. Wraps MNE-NIRS so every pipeline remains reproducible and inspectable.

Status

Pre-alpha — v0.1.0 in progress. Core pipeline engine, preprocessing, quality control, and HRF block-average blocks are implemented and exercised by 280+ tests. The CLI (nirspy run, nirspy serve) works end-to-end. The Dash GUI foundation is in place; the visual block builder ships in the next sub-release. See docs/roadmap.md for the full plan.

Why

Researchers who need a GUI for fNIRS today rely on Homer3, which requires a paid MATLAB license. The Python ecosystem already has excellent libraries — MNE-NIRS, Cedalion, pysnirf2 — but every one of them assumes the user can program.

NIRSPY fills that gap: a GUI-first, open-source (BSD-3) tool with reproducible YAML pipelines, running locally in the browser via Dash.

Features (v0.1)

• Linear, modular pipeline builder (drag-and-drop, coming in 5B)
• Optical-density conversion + modified Beer–Lambert law
• Bandpass filtering (configurable cutoffs)
• Automated quality control: Scalp Coupling Index + channel pruning
• HRF Block Average across conditions
• Probe / montage visualization (planned for 5C)
• Save and reload pipelines as YAML (round-trip stable)
• .nirs ↔ .snirf converter

Architecture details: docs/architecture.md.

Installation

NIRSPY is not yet on PyPI. Install from source:

git clone https://github.com/BrunoFurlanetto/nirspy.git
cd nirspy
python -m venv .venv
source .venv/bin/activate    # PowerShell: .venv\Scripts\Activate.ps1
pip install -e ".[dev]"

Once installed:

nirspy --version
nirspy serve                              # GUI on http://127.0.0.1:8050
nirspy run pipeline.yml --input data.snirf --output result.snirf

Development

pytest                       # 280+ tests, ~5 s on a laptop
ruff check .                 # lint
mypy src/nirspy              # type check (strict)

The Python 3.10 / 3.11 / 3.12 matrix runs on every PR via GitHub Actions.

Architecture

Three layers with strictly one-directional dependencies:

• domain/ — pure types (Pipeline, Block, DataType). No UI / MNE imports.
• engine/ — thin adapter over MNE-NIRS, plus the cache layer.
• gui/ — the Dash application.

Pipelines are persisted as YAML and treated as a public API: a breaking schema change bumps the minor version and ships an ADR.

Contributing

Read CONTRIBUTING.md and the Code of Conduct first. Bug reports and feature requests use the issue templates.

Security

Found a vulnerability? Do not open a public issue. Use the private security advisory flow.

License

BSD-3-Clause — same license as MNE-NIRS. Third-party license notices live in third_party/licenses/.

Acknowledgements

Built on top of MNE-NIRS and MNE-Python.