hypothesis: Find Bugs with Property-Based Testing

Overview

Hypothesis helps Python developers test behavior across a range of inputs instead of writing a separate test case for each example. You describe the kinds of inputs to explore and the property that should hold, and the library generates test data, including edge cases.

When a generated case fails, Hypothesis can reduce it to a simpler example. This makes the failure easier to inspect and is useful both for finding bugs and understanding them.

Key Features

• Generates test inputs from strategies that describe the values a test should examine.
• Checks a declared property against randomly selected inputs.
• Explores edge cases that may not be covered by manually chosen examples.
• Shrinks failing inputs to simpler cases that help clarify the failure.
• Integrates with Python tests through the @given decorator.
• Offers optional extras for additional functionality.

Use Cases

• Python developers can test functions that handle many possible inputs without maintaining a large collection of hand-written examples.
• Library authors can check that APIs behave consistently across varied data, including boundary cases.
• Developers working with collections or transformations can test general properties such as sorting behavior across generated lists.
• Teams investigating a hard-to-reproduce bug can use a minimized failing input to make the problem easier to understand.

Getting Started

Install the package with:

pip install hypothesis

See the README and documentation for usage details and optional extras.

Considerations

Hypothesis tests properties you specify, so useful coverage depends on choosing meaningful properties and input strategies. Generated tests complement, rather than replace, examples and other testing methods. The repository input reports its license as NOASSERTION, so check the project’s licensing details before adopting it where a confirmed license is required.