Summary and Setup

Welcome to learn about Machine learning using CMS open data during the Midsummer QCD school.

Docker containers

You will be using the Python container as instructed in the Docker lesson.

You are expected to have worked though the CMS open data example in that python container as instructed in the Dataset scouting lesson. if you have not done it yet, do it now.

Get the code

Clone the code from the example repository in a working directory that you have shared with the container. If you followed the instructions in the Docker tutorial, the working directory in cms_open_data_python:

BASH

cd cms_open_data_python
git clone git@github.com:thaarres/qcd_school_ml.git
cd qcd_school_ml

Add this code to your personal GitHub account.

In the GitHub Web UI, create a new repository as instructed in the Git tutorial.

Go to your GitHub area (https://github.com/[yourgithubname]), choose the “Repositories” tab and click on New.

Choose qcd_school_ml as the repository name, choose Public and leave other options as they are. This will create the repository and generate an instruction page.

On the terminal, check the remote:

BASH

git remote -v

OUTPUT

origin  git@github.com:thaarres/qcd_school_ml.git (fetch)
origin  git@github.com:thaarres/qcd_school_ml.git (push)

Change remote’s URL in order to be able to add the code to your personal GitHub account.

BASH

git remote set-url origin git@github.com:<GitHub username>/qcd_school_ml.git

Check again the name of the current remote:

BASH

git remote -v

Push the code to your new repository with

BASH

git push -u origin main

Whoops!

We noticed that there are changes in the original repository. Let us define it as upstream:

BASH

git remote add upstream git@github.com:thaarres/qcd_school_ml.git

Now you can pull the changes to the local repository:

BASH

git pull upstream main

and push them to your remote GitHub repository with

BASH

git push origin main

Check the container

Open the my_python container and check that you see the code repository:

BASH

docker start -i my_python
ls

The code should be visible under qcd_school_ml.

About the file list

Take note the file names of the CMS open data files used in the tutorial can be downloaded with the cernopendata-client as explained in the Dataset scounting tutorial

BASH

docker run -i -t --rm docker.io/cernopendata/cernopendata-client  get-file-locations --recid 63168 --protocol xrootd 
docker run -i -t --rm docker.io/cernopendata/cernopendata-client  get-file-locations --recid 33703 --protocol xrootd

A subset of them is already in the notebook, but this is how you would get them.

Callout

That’s it, you are ready to (machine) learn!