BlockInvGPFA

This package provides a Python implementation of Gaussian Process Factor Analysis (GPFA) that incorporates a novel approach to efficiently handle variable-length time series data. Building on the original method by Byron Yu et al. (2009), this implementation introduces a block-matrix–based inversion strategy that reuses kernel computations across trials of different lengths.

The package also includes scikit-learn–compatible API for integration into existing ML workflows such as:

• A Modular Preprocessing — Separates data preprocessing from model logic via a dedicated EventTimesToCounts transformer.
  • Accepts standard array-like inputs (not Neo objects), simplifying integration with other tools.
  • Follows scikit-learn's transformer–estimator interface for clean, reusable workflows.
• A new variance-explained metric to evaluate GPFA model fits.

This implementation is adapted from Elephant’s GPFA codebase with substantial modifications to improve performance, modularity, and usability in Python-based pipelines.

Usage

Installation

---

Install directly from GitHub using pip:

$ pip install git+https://github.com/CausalityInMotion/BlockInverseGPFA.git

Or clone the repo and install locally:

$ git clone https://github.com/CausalityInMotion/BlockInverseGPFA.git
$ cd BlockInverseGPFA
$ pip install . # or pip install .[test,docs] to include optional dependencies

You are now set to use the package.

---

Building the documentation

---

Building the documentation requires the following packages:

• Sphinx
• Read the Docs Sphinx Theme
• numpydoc
• Jupyter Notebook Tools for Sphinx

You can install the required documentation dependencies by:

$ pip install .[docs]

or manually by calling

$ pip install sphinx sphinx-rtd-theme numpydoc nbsphinx

Finally, to view the documentation locally, run

$ cd docs
$ make html
$ open _build/html/index.html

or view them online: BlockInvGPFA docs

A detailed walkthrough of the package — including how to fit the model to real neural data — is available in the Jupyter notebook example: (link)

---

Tests

---

To run the full test suite in the test folder, use:

$ pip install .[test]
$ pytest test/

Tests are automatically run via GitHub Actions on every push and pull request.

License

Modified BSD License based on Elephant, see LICENSE.txt for details.

Copyright

©️ 2021-2025 Brooks M. Musangu and Jan Drugowitsch

Acknowledgments

See acknowledgments.