Computational Research Workflows Homework

Doctoral Programme in Science and Engineering (DSSE)

Practicalities

By email to jack.hale@uni.lu, Subject: Computational Research Workflows Homework.

Please ask questions on the Etherpad.

Deadline: Monday 19th December 2022 1800

You must have attended the course and complete the homework to receive the ECTS credits.

The email should contain a link to a GitHub repository with:

• a modified version of this file README.md file containing the specific commands that you used to complete the exercises. You do not have to include everything, just the key commands. I have left placeholders where these commands should go:

  # Add your commands here
  

• A wallet.py file that passes the unit tests.

• A Dockerfile.

• A working GitHub Actions file .github/workflows/test.yml that runs the unit tests inside the built image.

High level overview

At the end of this homework, you will have a computational workflow consisting of:

1. A git repository hosted at https://github.com.

2. A Dockerfile describing an environment containing python and the Python module pytest.

3. A Docker image built from the Dockerfile and uploaded to the Dockerhub.

4. A simple Python program describing a simple Wallet class and a set of unit tests.

5. A continuous integration setup using GitHub Actions.

For reference, you can use my notes that I used during the course here and the references listed at the bottom of the main course page.

Setting up a git version control repository

1. Go to GitHub. Click on the plus icon in the top right hand corner of the screen. Then click on New Repository.

2. Create a public repository called e.g. jhale/computational-workflows-homework.

3. Clone your repository. In a terminal run, e.g.:

   git clone git@github.com:jhale/computational-workflows-homework.git

   substituting with the correct location of your repository.

4. Copy this README.md text file to your new repository. git add, git commit and git push it to your repository.

Dockerfile

1. Create a file Dockerfile in the repository containing the following text.

FROM ubuntu:21.04

RUN apt-get -y update && \
    apt-get install -y python3-minimal python3-ipython python3-pytest python3-numpy && \
    apt-get clean && \
    rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*

2. git add and git push the file Dockerfile to the repository.

# Add your commands here

Build and push Docker image

1. docker build, docker login and docker push the image described by the Dockerfile to the Dockerhub. Tag your Docker image <yourdockerhubusername>/computational-workflows.

# Add your commands here

Run a container, and share in files from the host.

1. docker run the image. Use the -v flag to share the root of the git repository to /root/shared inside the container. Use the -ti flag to get an interactive prompt inside the running container.

# Add your commands here

Setup a simple Python test suite

1. In a terminal running on the host (outside the container), copy across the files wallet.py and test_wallet.py to the root of your homework repository. git add, git commit and git push them.

# Add your commands here

2. Start a Docker container using your image and share your repository into a directory /root/shared into the container.

# Add your commands here

3. Run the tests inside the container by going to /root/shared and running the command py.test-3. The tests should fail.

4. In a terminal on the host modify wallet.py until the tests in test_wallet.py all pass.

5. git add, git commit and git push the working wallet.py file.

GitHub Actions for Continuous Integration

1. Using the example in the class notes make a .github/workflows/test.yml file that checks out your repository and runs the unit tests inside the Docker image that you pushed to the DockerHub.

2. Push the .github/workflows/test.yml file to GitHub. Check that you get the green tick showing that your tests pass.