Running Tests¶

This page explains how to run the test suite and add new tests to the Phentrieve project.

Testing Framework¶

Phentrieve uses pytest as its testing framework. Tests are organized into:

Unit tests: Test individual components in isolation
Integration tests: Test interactions between components
End-to-end tests: Test complete workflows

Running Tests¶

Running the Full Test Suite¶

Use the repository standard commands from the project root:

make test
make check
make typecheck-fast

Running Specific Test Categories¶

uv run pytest tests/unit/api/ -n 0 -v
uv run pytest tests/ --collect-only -q

Test Coverage¶

Coverage is configured in the default pytest settings. The full suite writes HTML and XML reports automatically when you run the standard test commands.

For targeted local work, prefer focused commands:

uv run pytest tests/unit/api/test_text_processing_router.py -n 0 -v --no-cov

The HTML report is written to the htmlcov directory during the default coverage-enabled runs.

Writing Tests¶

Unit Test Example¶

Here's an example of a unit test for a Phentrieve component:

import pytest
from phentrieve.embeddings import EmbeddingModel

def test_embedding_model_initialization():
    # Arrange
    model_name = "FremyCompany/BioLORD-2023-M"

    # Act
    model = EmbeddingModel(model_name)

    # Assert
    assert model.name == model_name
    assert model.dimension > 0
    assert callable(model.embed)

def test_embedding_generation():
    # Arrange
    model = EmbeddingModel("FremyCompany/BioLORD-2023-M")
    text = "Microcephaly"

    # Act
    embedding = model.embed(text)

    # Assert
    assert len(embedding) == model.dimension

Integration Test Example¶

import pytest
from phentrieve.data_processing import HPOProcessor
from phentrieve.indexing import IndexBuilder

def test_index_building_with_processor():
    # Arrange
    processor = HPOProcessor()
    builder = IndexBuilder("test_model")

    # Act
    terms = processor.process()
    builder.build_index(terms)

    # Assert
    assert builder.index_exists()
    assert builder.index_size() > 0

Test Fixtures¶

Phentrieve uses pytest fixtures for setting up test dependencies:

import pytest
from phentrieve.embeddings import EmbeddingModel

@pytest.fixture
def test_model():
    """Provides a test embedding model."""
    return EmbeddingModel("FremyCompany/BioLORD-2023-M")

def test_with_fixture(test_model):
    # The test_model fixture is automatically injected
    assert test_model.name == "FremyCompany/BioLORD-2023-M"

Test Configuration¶

Tests are configured in the repository root pyproject.toml file:

[tool.pytest.ini_options]
testpaths = ["tests"]
python_files = ["test_*.py"]

Mocking¶

For tests that require external dependencies, use the unittest.mock module or pytest-mock:

def test_with_mock(mocker):
    # Mock the ChromaDB client
    mock_client = mocker.patch("phentrieve.indexing.index_builder.chromadb.Client")
    mock_client.return_value.get_collection.return_value.count.return_value = 42

    # Act
    builder = IndexBuilder("test_model")
    count = builder.get_document_count()

    # Assert
    assert count == 42

Continuous Integration¶

Tests run automatically in GitHub Actions for pull requests and pushes. The main Python workflow is defined in .github/workflows/ci.yml.