Testing¶

Pipelit uses pytest for its test suite. All tests are located in platform/tests/ and use an in-memory SQLite database for isolation.

Running the Test Suite¶

cd platform
source ../.venv/bin/activate
export FIELD_ENCRYPTION_KEY=$(python -c "from cryptography.fernet import Fernet; print(Fernet.generate_key().decode())")
python -m pytest tests/ -v

Running Specific Tests¶

# Run a single test file
python -m pytest tests/test_api.py -v

# Run a specific test function
python -m pytest tests/test_api.py::test_create_workflow -v

# Run tests matching a keyword
python -m pytest tests/ -k "scheduler" -v

# Run with short traceback
python -m pytest tests/ -v --tb=short

Test Structure¶

Tests are organized by the subsystem they exercise:

Test Fixtures¶

Shared fixtures are defined in platform/conftest.py:

The _setup_db fixture runs automatically before each test, creating all tables and dropping them afterward for complete isolation.

Authentication in Tests¶

All API tests must use Bearer token authentication. Never use session auth, basic auth, or OAuth in tests.

Pattern: Authenticated Test Client¶

@pytest.fixture
def auth_client(client, api_key):
    client.headers["Authorization"] = f"Bearer {api_key.key}"
    return client


def test_create_workflow(auth_client):
    response = auth_client.post("/api/v1/workflows/", json={
        "name": "My Workflow",
    })
    assert response.status_code == 201

Pattern: Database Override¶

Tests override the get_db dependency to use the in-memory test database:

@pytest.fixture
def app(db):
    from main import app as _app
    from database import get_db

    def _override_get_db():
        try:
            yield db
        finally:
            pass

    _app.dependency_overrides[get_db] = _override_get_db
    yield _app
    _app.dependency_overrides.clear()

Writing New Tests¶

Test File Template¶

"""Tests for <subsystem>."""

from __future__ import annotations

import sys
from pathlib import Path

import pytest

# Ensure platform/ is importable
_platform_dir = str(Path(__file__).resolve().parent.parent)
if _platform_dir not in sys.path:
    sys.path.insert(0, _platform_dir)


def test_my_feature(db, workflow, api_key):
    """Test description."""
    # Arrange
    # ...

    # Act
    # ...

    # Assert
    assert result == expected

Guidelines¶

• Use descriptive test function names (test_switch_routes_to_matching_condition)
• Use the db, workflow, and other shared fixtures from conftest.py
• Always authenticate API requests with Bearer {api_key.key}
• Test both success and error paths
• Mock external services (Redis, LLM providers) where appropriate
• Keep tests independent -- do not rely on test execution order

Coverage¶

Check test coverage with:

python -m pytest tests/ --cov=. --cov-report=term-missing -v

Generate an HTML coverage report:

python -m pytest tests/ --cov=. --cov-report=html -v
# Open htmlcov/index.html in a browser

The project uses Codecov for tracking coverage on pull requests. Coverage badges are displayed on the repository README.