Problem
The test suite utilizes random.choice(), random.sample(), and random.randint() approximately 400 times across various test files, yet random.seed() is never called globally. This omission leads to intermittent CI failures due to non-deterministic random draws.
Evidence
grep -rn "random\.choice\|random\.sample\|random\.randint" tests/ | wc -l: 399grep -rn "random\.seed" tests/ | wc -l: 1 (limited scope)
Proposed Fix
Implement a global seeding mechanism in a top-level tests/conftest.py to seed both the random and numpy.random modules with a fixed value (e.g., 42). This will ensure that all test runs are reproducible and eliminate false CI failures.
Impact
- Eliminates intermittent CI failures across all Pull Requests.
- Improves developer experience by providing consistent test results.
- Facilitates easier debugging of test failures.
Problem
The test suite utilizes
random.choice(),random.sample(), andrandom.randint()approximately 400 times across various test files, yetrandom.seed()is never called globally. This omission leads to intermittent CI failures due to non-deterministic random draws.Evidence
grep -rn "random\.choice\|random\.sample\|random\.randint" tests/ | wc -l: 399grep -rn "random\.seed" tests/ | wc -l: 1 (limited scope)Proposed Fix
Implement a global seeding mechanism in a top-level
tests/conftest.pyto seed both therandomandnumpy.randommodules with a fixed value (e.g., 42). This will ensure that all test runs are reproducible and eliminate false CI failures.Impact