Python Blueprint

This is a template for Python development on Docker with VSCode.

You can customize the details according to your own preference.

Features

• Docker
  • Python 3.12
  • Support for CPU
• Package management with uv
• VScode
  • Optimized setup for Python development

QuickStart with GCP VM

1. Install VSCode Locally

• Install Remote Development Extension (including Remote-SSH and Dev Containers)
• Install gcloud CLI. Log in to your account.

2. Prepare a VM

Using a Marketplace option is convenient (note that there is a license fee).

1. For CPU (docker and docker-compose preinstalled)

Note: This template currently supports CPU only. Instead of launching VMs, I now use serverless services for GPU tasks.

You may need to SSH in and perform initial setup, so connect with gcloud compute ssh beforehand.

3. Connect via SSH in VSCode

• Run gcloud compute config-ssh locally

  • Adds the VM connection settings to ~/.ssh/config
  • Set the user to your own account (Tip: it's a good idea to add it to the end of the config)

  Host *
      User your_name
  

• Select "Open a Remote Window" in the lower left corner. Select "Connect to Host". A list of VMs will appear, so select your own VM.

4. Prepare the Dockerfile and .devcontainer on the VM

Clone this repository:

git clone https://github.com/shunyooo/python-blueprint.git

5. Project Setup

Run the following commands:

cd python-blueprint
. setup.sh

• The project name in the related files will be replaced.

• Login and configure gcloud and git on the host ( settings passed to container)

5. Access via devcontainer

• If there is a .devcontainer in the current directory, a pop-up will appear.
• Press the pop-up to start and attach the container. Alternatively, build from "Command Palette > Dev Containers: Rebuild…".
• Once the build is complete, execute inside the container.