GitAtbx

A modular Nextflow pipeline framework for automated data analysis — with built-in Git synchronization of results.

What is GitAtbx?

GitAtbx provides infrastructure for data pipelines that process multiple datasets in parallel, automatically version their results via Git, and track progress per participant. It is built on Nextflow for scalable, reproducible workflows.

The framework is domain-agnostic — modules follow simple input/output conventions (Parquet / FIF) and can implement any processing logic. gitatbx init sets up a new analysis project: directory layout, a parameters config file, and an optional GitHub remote for the results folder.

Core Commands

Command	Purpose
`gitatbx init <dir>`	Create a new analysis project
`gitatbx run [pattern]`	Find and launch a pipeline by project name
`gitatbx serve`	View result HTML locally in the browser
`gitatbx reinject <id>`	Re-inject corrected output for a participant
`gitatbx move <dest>`	Move the modules symlink
`gitatbx config show`	Show current configuration

Download & Installation

Install via pip:

pip install GitAtbx

Requirements: Python 3.10+, Java Runtime + Nextflow, Git
License: MIT
Status: Active development

Analysis Toolbox

A versatile Python toolkit for data analysis, statistical modeling, and scientific computing across research domains.

What is the Analysis Toolbox?

The Analysis Toolbox is a continuously evolving collection of analysis methods and computational tools built for scientific research. While it originated from work with physiological signals (EEG, fNIRS, ECG), it has grown into a general-purpose analysis framework designed to handle diverse data analysis challenges across different research domains.

Rather than being a fixed toolkit, it's an active development project that expands based on real research needs. Missing modules are added as new analytical challenges arise, and the architecture is designed to accommodate everything from time-series analysis and statistical modeling to machine learning pipelines and custom computational methods.

The toolbox provides validated, reusable components built on scientific Python libraries (NumPy, SciPy, Pandas, Statsmodels). Whether you're processing sensor data, running statistical analyses, building data pipelines, or developing custom algorithms, the toolbox offers tested building blocks that integrate into your workflow. All code follows open science principles and welcomes contributions—if you're missing a module or have suggestions, they're actively considered for implementation.

Core Capabilities

Signal Processing — Time-series analysis, filtering, spectral methods, and feature extraction
Statistical Analysis — Univariate and multivariate testing, regression, and statistical modeling
Data Pipeline Tools — Preprocessing, quality assurance, and automated workflows
Machine Learning Integration — Scikit-learn compatible transformers and utilities
Physiological Computing — Specialized methods for EEG, fNIRS, ECG, and other biosignals
Extensible Architecture — Modular design for adding custom analysis methods
Open Development — MIT licensed, documented, and accepting contributions

Download & Installation

The toolbox is available as an open-source Python package. Download the repository directly or clone it for development. Full documentation and example workflows are included.

Requirements: Python 3.8+, NumPy, SciPy, Pandas, Statsmodels
License: MIT
Status: Active development — suggestions welcome