tutorial_in_60_mins.propagate_docker_changes

Overview

• <SRC_DIR> = class_project/project_template

This skill supports two modes:

• Full sync: Make all Docker files across projects match project_template
• Last-commit sync: Propagate only the changes from the most recent commit in project_template

Mode A: Full sync

• Find each target directory <DST_DIR>: every directory under class_project/ (other than <SRC_DIR> itself) that already has a Dockerfile

  > find class_project -mindepth 2 -maxdepth 2 -name Dockerfile \
      -not -path "class_project/project_template/*"
  

• For each <DST_DIR> found this way
• Make the Docker files <DST_DIR>/docker_*.sh, <DST_DIR>/Dockerfile, <DST_DIR>/run_jupyter.sh as similar as possible to the corresponding ones in <SRC_DIR>
  • The goal is to reuse functions and factored out code (e.g., from class_project/project_template/utils.sh)
• If you are not sure about what the changes are
  • Summarize the changes for the user in bullet points
  • Ask for the user to confirm

Mode B: Last-commit sync

• Use this mode when you only need to propagate the most recent changes

Check the Last Commit

• Look at the changes from the last git commit in <SRC_DIR>
• Summarize the changes for the user in bullet points
• If you are not sure about what the changes are, stop and ask for the user to confirm

Propagate the Changes

• Propagate the changes in the Docker system from <SRC_DIR> to every <DST_DIR> found via the find command in Mode A

Preserve Behavior

• Make sure not to change the behavior unless needed

Update the Docker Docs

• Update class_project/project_template/docker_scripts.README.md explaining how this project uses Docker to build and execute containers given the new changes

Verification (both modes)

• [ ] All the docker_*.sh scripts work, run either directly or through the relevant pytest unit tests