sertansenturk / coding-challenge-template

Template repository to kickstart a Python/Docker environment for coding/take-home challenges.

challenge_template

Template repository to kickstart a Python/Docker environment for coding/take-home challenges.

Technology Stack

• Language: Python 3.8
• Containerization: Docker
• Serving: flask
• Continous Integration: Github Actions

Prerequisites

• Docker: Please refer to the documentation for installation instructions.

How to run

Starting the API

To build & start the application, execute in the terminal:

make api

The above command will start a Docker-compose stack. The stack has a single container with the challenge_template package installed behind flask.

Once the stack is up and running, the template API will be exposed at localhost:5000/template.

Posting queries

Open a new terminal window and enter a query, e.g.:

curl -X POST localhost:5000/template -d 'some message'

In the above case, the (dummy) template API returns the message back with additional information as the response.

If you are lazy to remember the full curl command, you could run:

make post QUERY='some message'
# Received POST request: 'some message'

Running the batch app in Docker

For batch use cases, you could run the workflow built on the functionality in challenge_template package as:

make batch

Currently, the (dummy) template driver prints out the command line arguments passed to the driver.

Running notebooks

It is typical to carry out exploratory data analysis (EDA) on the data and test different approaches alongside implementing a solution. Jupyter is a good ecosystem for such work.

If you want to run the notebooks locally, you can start Jupyter Lab by running in the terminal:

make jupyter

Above command builds a docker-compose stack and runs a container with Jupyter installed. Once the Jupyter is up and running, you should see a URL looking like http://127.0.0.1:8888/lab?token=9f59... in the terminal screen.

Clicking the URL should take you the Jupyter Lab on your browser of choice.

Note the first build could take around 6-7 minutes, depending on your machine and internet connection.

Running unittests

Run in the terminal:

make unit-tests 

which runs unit tests using pytest.

The unittests are also run on Github actions after each push: link.

Code formatting

In the terminal run:

make format

which formats the Python code in the repo using black.

Code linting

In the terminal run:

make lint

which shows Python code style / linting issues using flake8 and pylint.