Hello World, Autodesk!

This is Tony's submission for the Autodesk DPI(ADP?) take-home challenge. The application is created using the Python Django web framework and features:

Docker container artifact that can be deployed to any environment/OS
Instructions for local development and deployment using Pipenv
Pinned dependencies for reproducibility
Logging configurations for either Docker/Pipenv
Simple PoC CI/CD pipeline leveraging CircleCI and Google Kubernetes Engine

OS-independent Build, Deployment, and Test

For deployment, the application is containerized to be as portable as possible. This only assumes Docker Engine 18.09 or later to be installed.

docker build . which returns an IMAGE_ID
- Uses the production-grade command gunicorn project.asgi:application -k uvicorn.workers.UvicornWorker (later: can try this locally after cd project and pipenv shell)
docker run -it -p 8000:8000 <IMAGE_ID>
Test app in different terminal with curl http://127.0.0.1:8000/ -H "Accept: application/json" (header optional)
For unit tests, run docker run -it <IMAGE_ID> python manage.py test

Local Development/Testing installation instructions

For local development and improving the app, pipenv is used. The only assumption is Python 3.7 (with pip) installed on either a Linux, MacOS, or WSL workstation.

pip install pipenv (can also do this outside of pip e.g. MacOS brew install pipenv) and export PIPENV_VENV_IN_PROJECT="enabled"
In repo root, pipenv install --dev --python 3.7.5 and pipenv shell
cd project and python manage.py runserver to start the local dev server
Test app (separate terminal window) with curl http://127.0.0.1:8000/ or curl http://127.0.0.1:8000/ -H "Accept: application/json" to see locahost server response.
Run unit tests with python manage.py test
Materialize pip requirements pipenv lock -r > requirements.txt to be used with Dockerfile

Logging

Django uses Python’s builtin logging module to perform system logging. By default, Django uses the dictConfig format, configured inside of settings.LOGGING dictionary. Logging level is by default INFO which does not include the request URL log. To turn on request URL logs:

If running in Docker mode, use docker run --env API_LOGLEVEL=DEBUG -it -p 8000:8000 <IMAGE_ID>
If running in pipenv mode, use API_LOGLEVEL=DEBUG python manage.py runserver
If running on kubernetes, add API_LOGLEVEL=DEBUG to deployment spec.template.spec.containers.[0].env

Cloud and CI/CD

Included is a simple k8s/manifest.yaml file for use with kubectl once the CI/CD service account is granted access to a hosted k8s cluster (via GCP service accounts).
The .circleci/config.yml defines a basic pipeline which tests branch commits before merging into main and pushes a new docker image to Container Registry.
A human operator can approve the "Hold" step which then deploys the image tagged with "latest"

Other information