Template for dockerizing your python application for local development

Disclaimer: 
* This is work in progress.
* Contributions very welcome. 
* This is highly opinionated. 
* Tested and optimized for Mac. 
* Heavily inspired by https://github.com/fjetter and stackoverflow.

Getting started

You need to export some environment variables (make sure $USER is set correctly)

export UID=$(id -u)
export GID=$(id -g)

A simple

docker-compose up -d

Should start the docker compose setup You can login for ineractive work via

docker-compose exec python_local_dev bash

To execute the tests, you need to activate the virtual environment

source /tmp/venv/bin/activate

Then you need to install pytest (sorry, WIP)

pip install pytest

Then you should be able to develop your package (==execute tests ;-)

pytest
================================================= test session starts ==================================================
platform linux -- Python 3.7.4, pytest-5.2.1, py-1.8.0, pluggy-0.13.0
rootdir: /home/sneubauer
collected 1 item

tests/test_funniest_app.py .                                                                                     [100%]

================================================== 1 passed in 0.21s ===================================================

Get pip caching working

One of the quirks is, that the mounted volume once needs to be chown'ed in order to get it running with a non-root user inside docker compose (docker/compose#3270)

Change the "user" to "root" in the docker-compose.yaml and restart docker compose. Then log into a shell (now as "root") and chown the mounted volume to your user (YOURUSERNAME).

chown -R $(id -u YOURUSERNAME):$(id -g YOURUSERNAME) /home/YOURUSERNAME/.cache/pip

Then log out and change the "user" back in the docker-compose.yml.