Adventures in not writing tests
summary
Claim: Hypothesis ghostwriter generates runnable tests from type hints so many data-analysis checks get created automatically. Demonstrations run on a 10,000-sample scenario and find real failures — ceiling() needs NaN and Infinity handling, empty DataFrame paths break aggregation, infinities produce NaN math issues. Shows fixes and workflow tools like Hypothesis extras pandas, assume, target and register_type_strategy, plus CI reproducibility settings. Net result is measurable time saved and more edge cases caught, with concrete examples that actually fail and then get fixed.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 2:35Why test writing focus
- 5:47Property-based testing with Hypothesis
- 11:27Edge cases in standard library
- 13:36Data frame testing with Hypothesis
- 18:35Hypothesis Extras for Pandas
- 20:22Refining tests with assume
- 21:50Advanced Hypothesis features
- 25:05Conclusions and when to test
