Introduction ¶

The piece emphasizes that “Testing is an important part of software development, be it manual or automated.” The author argues untested code poses significant risks and advocates for developers implementing automated testing, particularly unit tests.

Use Case Scenario ¶

The article presents a practical example: developing a weather subscription feature for a GIS farm application that integrates with a paid third-party weather service. The challenge involves testing code that depends on external APIs without incurring costs from repeated real API calls.

The Problem Function ¶

The article demonstrates a create_weather_subscription function that makes POST requests to a weather service:

import requests

def create_weather_subscription(user_id, package_id, start_date, end_date):
    payload = {
        'user_id': user_id,
        'package_id': package_id,
        'start_date': start_date,
        'end_date': end_date,
    }
    response = requests.post(
        'https://real-weather-service.com/weather/subscribe/', data=payload
    )
    return response

Solution: requests-mock Library ¶

The author recommends using requests-mock to intercept HTTP requests during testing. The library “allows preloading requests with responses” without making actual API calls, enabling cost-effective and reliable unit testing.

Installation ¶

pip install requests-mock

Initial Test Implementation ¶

The article walks through creating test_create_weather_subscription.py with unittest:

import requests_mock
import unittest

from create_weather_subscription import create_weather_subscription

class TestSubscribeWeather(unittest.TestCase):
    @classmethod
    def setUpClass(cls):
        cls.user_id = 10
        cls.package_id = 'dummy-package-id'
        cls.start_date = '2020-10-01'
        cls.end_date = '2020-12-31'

    def test_create_weather_subscription_success(self):
        with requests_mock.Mocker() as rm:
            response = create_weather_subscription(
                self.user_id, self.package_id, self.start_date, self.end_date
            )

            self.assertEqual(response, 'Weather data subscribed successfully!')

Key Feature: Safety Guard ¶

Running this test initially fails with: “No mock address: POST https://real-weather-service.com/weather/subscribe/" This prevents accidental real API calls—a protective feature of the library.

Adding Mock Responses ¶

The solution involves configuring mock responses before executing the function:

return_value = {
    'id': 101,
    "user_id": self.user_id,
    "package_id": self.package_id,
    "start_date": self.start_date,
    "end_date": self.end_date,
}
rm.post('https://real-weather-service.com/weather/subscribe/',
        json=return_value, status_code=201)

The json parameter defines the response payload, while status_code specifies HTTP status codes.

Comprehensive Test Suite ¶

The final implementation includes three test cases:

1. Success scenario (201 status code)
2. Authorization error (401 status code with “Unauthorized” message)
3. Overlapping subscription (400 status code with conflict message)

All tests pass successfully, demonstrating the library’s effectiveness.

About the Author ¶

Zulfikar Akbar Muzakki is an Indonesian software developer based in Purworejo, Central Java. His professional journey includes web development with Python and Django. Beyond coding, he volunteers with a kids’ learning center and enjoys motorcycle travel, cultural exploration, and performing arts.