Skip to the content.

Back to the projects list

Dockerize spike sorters

Key Investigators

Project Description

One of the biggest usability challenges of SpikeInterface is installing the different spike sorters. This is particularly challenging for sorters that run on MATLAB and those that have specific requirements like GPU. Creating docker containers for each sorter will make them easier to install locally and deploy remotely.

Objectives

The goal of this project is to finalize and clean up the container images for all sorters in SpikeInterface.

Approach and Plan

The plan is to proritize KS and PyKilosort implementations, since they are the most widely used sorters. If we have time, we’ll move on to WaveClus first and then HDsort. A parallel effort will involve setting up a testing suite on in github actions to test the built images on some toy data using docker and/or singularity.

Progress and Next Steps

Materials

Background and References