Learning Software Testing: Unit, Integration, E2E, and API Testing

This blog explores key concepts in software testing, from isolated unit tests to comprehensive API validations. We’ll cover the characteristics, best practices, and examples for each type.

Unit Testing

A unit testing verifies one small piece of logic (a function/class) in isolation. Here, unit means smallest testable behaviour, not the whole system.

Features of Good Unit Testing

• Isolated (test only one unit. No DB, API, file system)
• Deterministic (same result every run)
• Fast (runs in millisecond)
• Clear intent (anyone reading can understand what behavior is verified)
• Test behaviour, not implementation (focuses on result, not internal variables)

Features of Bad Unit Testing

• Tests on multiple things (hard to know what failed)
• Depends on DB/API
• Random values (fails sometimes)
• No assertion (verifies nothing)

from math_utils import calculate_discount
import pytest
def test_calculate_correct_discount():
    assert calculate_discount(100, 10) == 90

def calculate_discount(price, percent) -> float:
    if percent < 0 or percent > 100:
        raise ValueError("Invalid percent")
    return price - (price * percent) /100

How to test with mock

import pytest
from utils.user_service import get_username
@pytest.mark.asyncio
async def test_get_username():
    async def fake_fetch(uid):
        return {"name": "Bibe"}

    name = await get_username(fake_fetch, "1")
    assert name == "Bibek"

async def get_username(fetch_user, user_id: str):
    user = await fetch_user(user_id)
    return user["name"]

Integration Testing

• In this, we test interactions between components.
• Its scope is API + database | API + external service
• Mocking is limited (e.g, real DB, mocked external services)
• Example: Test that API endpoint correctly saves to database

Characteristics

• Verifies data flow between components
• Test business logic with real dependencies
• Isolate external systems with mocks/subs
• These are run in CI (Run tests on every PR and deployment) pipeline

End-to-End (E2E) Tests

• Its purpose is to test complete user flows
• Its scope is UI -> API -> DB -> External services -> UI
• Its slower

Characteristics

• Simulate real user scenarios
• Test entire system in production like environment
• Catch integration issues between all components

API Testing

Test categories for apis:

Functional

In this we test:

• Status codes
• Response structure
• Data structure
• Data accuracy

Performance

In this, we test:

• Response time
• Throughput (how many request can an API handle per unit time)
• Load handling

Security

In this, we test:

• Authentication
• Authorization
• Injection attacks

Reliability

In this, we test:

• Error handling
• Rate limiting
• Retry logic

Key API Testing Strategies

Positive Testing

Valid inputs with expected 2xx responses

Negative Testing

Invalid inputs, mal-informed data (4xx / 5xx)

Boundary Testing

Minimum/maximum values, pagination limits

Contract Testing

Ensuring API meets Open API specification

State Transition Testing

Multi-step workflows (e.g, create -> update -> delete)