amstee / sdet-c19-test

SDET COVID-19 Coding Test 🦠

The app estimates the chance that at least one COVID-19 positive individual will be present at a event of a given size. For example:

At a concert in England with 2000 people in the audience, what is the probability that at least one of them currently has COVID-19?

The dataset in data.json contains the daily number of new Coronavirus cases for each United Kingdom nation (England, Scotland, Wales, Northern Ireland) published by the UK government.

{
    "length": 1700,
    "body": [
        {
            "date": "2021-03-02",
            "areaType": "nation",
            "areaCode": "N92000002",
            "areaName": "Northern Ireland",
            "newCasesByPublishDate": 149
        },
        // etc.
    ]
}

The API exposes two endpoints:

• GET /areas - returns a list of areas where events can take place.
• POST /compute - given an area and event size, returns the probability of at least one person at the event having Coronavirus.

The user chooses an area from the list returned by the GET /areas API:

curl http://localhost:5000/areas

{
  "items": [
    {
      "code": "E92000001",
      "name": "England"
    },
    {
      "code": "N92000002",
      "name": "Northern Ireland"
    },
    {
      "code": "S92000003",
      "name": "Scotland"
    },
    {
      "code": "W92000004",
      "name": "Wales"
    }
  ]
}

Then uses the POST /compute API to calculate the probability:

curl -X POST http://localhost:5000/compute --data '{"sample": 2000, "area_code": "E92000001"}'

{
  "probability": 0.9999994281491117
}

Note: the probabilities it produces are "just for fun", don't take them too seriously!

Overview 🔭

The app is written in Python using the Flask framework.

covid_app/
  model.py       ← module to compute probabilities
  dataset.py     ← module to read the dataset from the filesystem
  validation.py  ← module to validate incoming requests

tests/
  unit/          ← unit tests
  integration/   ← integration tests
  e2e/           ← end to end tests

app.py           ← API views
data.json        ← number of new cases per day and nation
requirements.txt ← Python dependencies

Goals 🎯

As you will discover, the app is not ready for production quite yet! The goals of this exercise are to:

1. Talk through the code and explain what it does.
2. Run the application, and make a request with curl.
3. Write some tests to give us confidence the app is working (hint: edge cases!). Use your judgement to decide what types of test to focus on: unit, integration, and E2E.

Running the app

In Repl.it, you can simply click the green Run button, the output will be shown in the Console tab.

Create a virtualenv and install the requirements:

python3 -m venv .venv
. .venv/bin/activate
pip3 install -r requirements.txt

Run the app:

flask run

Note: this will run the app.py.

Make an API request using curl:

curl -X POST http://localhost:5000/compute --data '{"sample": 2000, "area_code": "E92000001"}'

Running the tests

In Repl.it, you can run commands from the Shell tab.

You can run all the tests with:

pytest tests

Or just the unit tests (for example) with:

pytest tests/unit

Note: the E2E tests expect the app to be running and listening on localhost:5000.