aritztg/ep-dev-test

EP Dev challenge

EP Dev Test

Installation

This guide assumes that the reader has a minimal experience with python and its virtualenvs.

1. Install gdal $ brew install gdal or (untested) $ apt install gdal.
2. Install a python version (this was tested with 3.11.3): $ pyenv install 3.11.3. If pyenv is not available in your OS, you can use virtualenv (poetry may require different steps).
3. Create a virtualenv: $ pyenv virtualenv 3.11.3 earthpulse-simple-api
4. If not already, now you can start using that environment: $ pyenv activate earthpulse-simple-api
5. Install dependences: $ pip install -r requirements.txt
6. (Optional) Install dev dependences (for linting and such): $ pip install -r requirements-dev.txt
7. Fill credentials in .env file. Do not use double quotes or whitespaces.
8. Execute the cli app. $ python app/simple_api.py (no params/argparse use yet).

How to run

Locally

As a regular python app (in the right environment):

$ python app/simple_api.py

or using makefile entrypoint:

$ make run

Then you should be able to access [http://0.0.0.0:80/docs] and play around with the exposed endpoint. If you prefer, you can also check docs via [http://0.0.0.0:80/redoc] UI.

Docker

First build the image (read the file first to understand what it does):

$ docker build -t earthpulse-simple-api .

Then you can run a container based on that image:

$ docker run -d --name earthpuse-test-container -p 80:80 earthpulse-simple-api

Further development

• Consider the use of pylama for more complete linting.
• Add some FastAPI middlewares, such HTTPSRedirectMiddleware, TrustedHostMiddleware.
• Integrate some logs/app monitor like Sentry or Datadog.
• Using Github actions, check linters and go further (staging/deploy) only if certain threshold has been met (done).

Linters

Please ensure you have installed development dependences first: $ pip install -r requirements-dev.txt.

Ruff

Very fast linter, but still does not have the same amount or rules available in pylint. It would be perfect to be used in a pre-commit Git hook. Uses pyproject.toml as configuration file.

ruff check .

Pylint

It uses the configuration pylintrc settings file. No rules were disabled, AFAIK. It takes more time with larger codebases. It can be executed locally, and tipically in CD/CI pipelines instead of ruff (because of its large amount of rules), making the pipeline failing if a minimum threshold score is not reached.

$ pylint $(find . -name "*.py")

--------------------------------------------------------------------
Your code has been rated at 10.00/10 (previous run: 10.00/10, +0.00)

Isort

Not exactly a linter, but a useful tool anyway. It ensures some guide styling in regards of module imports.

$ isort .

Tests

Ensure you install dev-requirements first. Then simply execute:

$ python -m unittest discover

Makefile

You can run (you need to be in the right virtualenv first):

make run
make tests
make lint

Github actions

A Github action has been added (.github/workflows/code_quality.yml) to check pylint over all py files in the main branch. If linter score is higher than 9, it could continue to the next ci/cd step (not defined yet).