Miniconda Container

Simple non-root user Miniconda environment, with support for multiple architectures.

Features

• Installs the latest version of Miniconda.
• Sets up a non-root user (default: app) with sudo privileges (for debug purposes).
• The base Conda environment is automatically activated upon login.
• Supports linux/amd64 and linux/arm64 architectures.
• Includes git, curl, sudo, and gosu.

Build

To build the multi-platform Docker image, you can use:

docker buildx build \
  --platform linux/amd64 \
  -t c127/miniconda:latest \
  .

This script uses docker buildx and will tag the image as c127/miniconda:latest.

How to Run

You can run the container using the following command:

docker run -it --rm \
  -e PUID=$(id -u) \
  -e PGID=$(id -g) \
  -e APP_USER=myuser \
  -e APP_PASS=mypassword \
  -v $HOME:/app \
  c127/miniconda:latest

This will start an interactive bash session within the container, with the Miniconda base environment activated for the specified user.

Environment Variables

The following environment variables can be configured when running the container:

• PUID: Process User ID for the in-container user. Defaults to 1000.
• PGID: Process Group ID for the in-container user. Defaults to 1000.
• APP_USER: The username for the in-container user. Defaults to app (as set in the Dockerfile).
• APP_PASS: The password for the APP_USER. Warning: The default password (RT2fEDayXh4zVQBtr7WwCHgd) set in the Dockerfile is for debugging purposes only. Please override this with a secure password for any production or sensitive use.

Volumes

Use volumes to persist data on host computer and keep fast your environment (Recommended)

• Main Volume: For conda environments & packages and home for user data /app.

Inside the Container

Once the container is running, you will be logged in as the user defined by APP_USER. The Miniconda base environment is automatically sourced via the source/app-start script, which is called by the source/entrypoint script.

You can immediately start using conda commands:

conda info
conda install numpy
conda create -n workspace python=3.9
conda activate workspace

The user also has sudo privileges, using the password provided via the APP_PASS environment variable.

Dockerfile Breakdown

The Dockerfile performs these key steps:

1. Uses debian:bullseye-slim as the base image.
2. Sets essential environment variables, including CONDA_INSTALL_PATH, APP_USER, and APP_PASS.
3. Installs necessary packages like curl, bash, sudo, git, and gosu.
4. Copies the source/entrypoint and source/app-start scripts into the image.
5. Sets /usr/local/bin/entrypoint as the ENTRYPOINT.

The source/entrypoint script handles:

• Creation of the user and group based on PUID, PGID, and APP_USER.
• Installation of Miniconda for the appropriate architecture (TARGETARCH).
• Initialization of conda for the bash shell.
• Setting the user's password and granting sudo access.
• Execution of the source/app-start script.

The source/app-start script:

• Activates the base Conda environment.
• Starts an interactive bash shell.

License

This project is licensed under the MIT License. See the LICENSE file for details.