Implementation:Scikit learn Scikit learn Conftest

Knowledge Sources	Scikit_learn Scikit-learn Docs
Domains	Machine Learning, Testing, Configuration
Last Updated	2026-02-08 15:00 GMT

Overview

Concrete pytest configuration module for the scikit-learn test suite provided by scikit-learn.

Description

The conftest module configures the pytest test environment for scikit-learn. It defines fixtures for dataset fetchers (e.g., fetch_20newsgroups, fetch_california_housing), handles warning filters, manages thread pool limits, configures BLAS settings, and provides skip conditions for 32-bit platforms and network-dependent tests. It also supports parallel test execution via pytest-run-parallel.

Usage

This module is automatically loaded by pytest when running scikit-learn tests. It provides fixtures and hooks that control test execution behavior, dataset caching, and platform-specific test skipping.

Code Reference

Source Location

• Repository: scikit-learn
• File: sklearn/conftest.py

Signature

def raccoon_face_or_skip():
    ...

dataset_fetchers = {
    "fetch_20newsgroups_fxt": fetch_20newsgroups,
    "fetch_california_housing_fxt": fetch_california_housing,
    ...
}

@pytest.fixture
def pyplot():
    ...

def pytest_collection_modifyitems(config, items):
    ...

def pytest_configure(config):
    ...

Import

# Automatically loaded by pytest; not imported directly
# Fixtures are available to all tests in the sklearn package

I/O Contract

Inputs

Name	Type	Required	Description
config	pytest.Config	Yes	Pytest configuration object (provided by pytest)
items	list	Yes	List of collected test items (provided by pytest)

Outputs

Name	Type	Description
fixtures	various	Dataset fixtures, pyplot fixture, and global configuration for tests

Usage Examples

Basic Usage

# In a test file within sklearn/tests/
def test_with_dataset(fetch_california_housing_fxt):
    data = fetch_california_housing_fxt
    assert data is not None

def test_plotting(pyplot):
    # pyplot fixture ensures matplotlib is properly cleaned up
    import matplotlib.pyplot as plt
    fig, ax = plt.subplots()

Related Pages

• Principle:Scikit_learn_Scikit_learn_Metric_Evaluation