Skip to content

step1.md

Latest commit

 

History

History
169 lines (111 loc) · 3.28 KB

step1.md

File metadata and controls

169 lines (111 loc) · 3.28 KB

Step 1: Creating a Docker image for data analysis [DL] [PL]

Install the Docker Desktop/Engine

https://docs.docker.com/engine/install/

https://docs.docker.com/desktop/setup/install/mac-install/

Open Docker Desktop/Engine

Open the application and log in. The container can't exceed the Docker VM's ceiling.

So use Docker Desktop → Settings → Resources → Memory and swap → set to maximum GB → Apply & Restart → then rebuild the container.

Open the terminal, try:

docker login 
Authenticating with existing credentials... [Username: enesn]

Also try:

docker images

Which should return

IMAGE   ID             DISK USAGE   CONTENT SIZE   EXTRA
(base) enes@vl965-172-31-78-189 ~ %

Find an image on DockerHub to get started

Find an image in which preferred environments and packages are already installed. Good candidates are:

Hardened Images catalog | Python | Docker Hub

rocker/tidyverse - Docker Image

Pull the image

Using the terminal,

docker pull rocker/tidyverse:latest
docker pull python:3.12-slim
  1. Downloads the rocker/tidyverse image (tag latest) from Docker Hub to your machine.
  2. You now have a local copy you can run/modify.

or

docker pull quay.io/jupyter/scipy-notebook

Run and improve the container from the image

Run the image to start a container:

docker run -it rocker/tidyverse:latest bash

Then install anything you need inside the bash to use it later always:

root@7f16ee033970:/# apt update && apt install -y curl
curl -fsSL https://claude.ai/install.sh | bash
apt update && apt install -y \
    python3 \
    python3-pip \
    python3-venv
apt update && apt install -y python3-venv
python3 -m venv /opt/venv
source /opt/venv/bin/activate
pip install pandas seaborn matplotlib numpy ipython jupyter

Commit the container

Learn the container name: (you can also use the Docker Desktop)

docker ps

If you modified a running container interactively and want to save it:

docker commit <container_id> <image_name>

Example:

docker commit infallible_heyrovsky enesn/rtidyverse-python-ds:052026

Tagging means duplicating a container, changing its name, and version.

Docker registries require:

username/repository:tag

Format:

docker tag <local_image> <dockerhub_username>/<repo>:<tag>

Example:

docker tag my-analysis-env enesn/rtidyverse-python-ds:052027

Push to DockerHub

docker push enesn/rtidyverse-python-ds:052027

Push also to GitHub Container Registry (GHCR)

DockerHub can delete inactive containers. To be on the safe side, push also to GHCR. It is easier to link a container to a GitHub repository.

Generate a Git token if you haven’t done yet. Then authenticate:

echo token| docker login ghcr.io -u enesn --password-stdin
docker tag enesn/rtidyverse-python-ds:052027 ghcr.io/enesn/rtidyverse-python-ds:052027
docker push ghcr.io/enesn/rtidyverse-python-ds:052027
  1. Uploads the image to GitHub Container Registry.
  2. Needs authentication (a GitHub token with write:packages) and docker login ghcr.io.

Pull the image

docker pull ghcr.io/enesn/rtidyverse-python-ds:052027