A repo that contains all the code and guides for doing basic spike sorting in a containerized terminal app using docker.
- Install Docker desktop
- Launch Docker. Note, that all containers and images can be deleted from docker to reset it if necessary, no need to worry about deleting docker images as everything is backed up and pulled from Docker Hub automatically.
- Run the
spikesort.batfile by double clicking it or run a shortcut of this file (Only works for windows machines, but can be expanded for Unix systems in the future) - A command line terminal will run and show the following prompt:
FULL PATH TO THE PARENT FOLDER OF THE SPIKESORT DATA FOLDER:
Copy and paste the path into the terminal of the folder that contains the data folder where all the spikesort input data is located.
e.g. C:\Users\Padilla-Coreano\Desktop\GITHUB_REPOS\diff_fam_social_memory_ephys is a correct example. C:\Users\Padilla-Coreano\Desktop\GITHUB_REPOS\diff_fam_social_memory_ephys\data is not.
5. Press Enter and let it run.
app.py- This file contains the python script that handles the process and spike sorting.Dockerfile- This file controls the building process of building a new docker image.nancyprobe_linearprobelargespace.prb or .txt- This file represents a default example of the required .prb file that is required to run the spike sorting in the current setup.spikesort.bat- This file is a script that pulls the docker image and controls the process of spinning up a container from the image. This file also launches the actual python script/app and is what is run when clicking on the shortcut of spike sorting. Technically this is also the only file you need to run the app on a windows machine with docker installed.synapse-spike.ico- This is just an icon that can be used when creating a shortcut for thespikesort.batfile.
Currently this Repo contains the necessary code to build a docker image. The docker image contains a Conda environment wherein a python script is run. The python script that is run is the app.py file. This is the file that controls all the processing. When this file is changed a new version of the docker image ahs to be built and pushed to docker Hub before the app works on a local PC. To create the docker image run the following commands from within the spikesort_docker folder in the terminal:
# Build the docker Image (--no-cache can be excluded to cache previous builds to speed up the process)
docker build --no-cache -t spikesort:latest .
# Tag the Docker Iamge as the latest version
docker tag spikesort padillacoreanolab/spikesort:latest
# Push the Docker image to DockerHub
docker push padillacoreanolab/spikesort:latest
When running the spikesort.bat file the docker image is pulled/updated from docker Hub. Thereafter a container is created running the app.py file inside the Conda environment from within the container. Note, that the docker image copies all the files in the root folder of this repo over into it when creating the image. When the container of the docker is spun up from the image it links a specified folder on the local computer to a path in the container named /spikesort. This should be kept in mind when modifying paths in the app.py file.
- Use the PHY2_Shortcut as an example to run this app outside docker.
- Use a Params.py file to give more various inputs to the processing functions.
- Change it to be a command line run function with additional parameters.
- Update the
app.pyfile so that it is not as dependent on the /data/ folder and needs it as a requirement. - Add the default .prb file parameters directly into the code as a default for a .prb parameter.
- It should have a parameter to disable batch processing.
- It should not reprocess files when their results already exist in the output folder.
Code was written by @ChristopherMarais so contact him for any questions