Skip to content

Latest commit

 

History

History
54 lines (35 loc) · 3 KB

README.rst

File metadata and controls

54 lines (35 loc) · 3 KB

Philistine

A Python package for Phillip's helper and utility functions, especially for EEG and statistics.

Status

pipeline status coverage report documentation status license pypi

Overview

Philistine is a collection of hopefully useful functions in Python for statistics and analysis of EEG data using existing packages in the Python ecosystem. It is not intended to be a standalone package, but rather a convenient way to distribute manipulations that I (Phillip) find useful in my own work.

This is very much aa hobby project developed in my free time (in a language I don't use much anymore) and the API is subject to change in a rather volatile fashion as improvements, corrections, etc. are made. The idea is provide a convenient way to redistribute functions that I (Phillip) find useful. The hope is that many of these functions are eventually integrated into packages such as MNE, bambi, etc. At that point, the functions will be changed into thin wrappers for those other packages, deprecated and eventually removed.

The BV-writer functionality will likely be removed in a future release. MNE now has an export module, which takes advantage of [pybv](https://pypi.org/project/pybv/), which in turn took the good parts of the writer here and added some active maintenance.

Installation

Philistine requires a working Python interpreter. As of version 0.2.0, this must be at least Python 3.7 for compatibility with MNE 1.3.

Assuming a standard Python environment is installed on your machine (including pip), Philistine itself can be installed in one line using pip:

python -m pip install --user --upgrade philistine

Alternatively, if you want the bleeding edge version of the package, you can install from GitLab:

python -m pip install --user --upgrade  git+https://gitlab.com/palday/philistine.git

Dependencies should be handled automatically by pip.

Development

The primary hosting for this project is on GitLab, and issues should be raised there. A GitHub mirror is provided for convenience and redundancy. Pull requests can be made on either site.