willian-m/hep-analysis

A docker image with jupyterlab and tools for high-energy physics phenomenology analysis. Can connect with google colab.

HEP Analysis toolbox

This repository is my personal toolbox for analysis of high-energy physics data using jupyter-notebook. It contains a Dockerfile to build a Docker container with the necessary tools, includiong the jupyter-notebook. It is tailored for use along Google's collab, allowing easy integration Google's drive and thus easy sharing of you analysis notebook with colleagues.

Workflow

Assuming your data is in a folder /data/path on your local computer, you do the following:

1. Clone this repository
2. From within the cloned repo, build the docker image: docker build . -t hep-analysis
3. Run the container with: docker run --mount type=bind,source=/data/path,target=/path/desired/on/container -p 8888:8888 -l container-name
4. The output you give a URL similar to: http://127.0.0.1:8888/?token=<random token string>. Copy it.
5. You can either paste the URL into a browser, or open a new Google collab notebook, click on the arrow at the side of the resource meter/reconnect button at the top left corner and select Connect to a local runtime. On the modal opened, past in the backend URL the copied URL.

Tip 1: If you want to keep all data oon Google Drive, you can either use rclone to manually sync data back and forth between your hard drive and Google Drive or use Google's client.

Tip 2: It is possible to have the docker container running in a remote machine and use your local computer to edit a collab notebook. You will need to forward the remote port 8888 to a local port (that may be the same 8888) port, even if there is a gateway between your remote and local computer (the most common architecture in universities). The easiest way is via SSH with the command ssh -L 8888:remotehost:8888 gatehost.