This skill provides specialized guidance for writing, debugging, and optimizing tests for BayesFlow extension packages. It ensures critical environment setups, such as Keras 3 backend initialization in conftest.py, are handled correctly to prevent runtime errors. Developers can leverage standardized patterns for simulator shape verification, configuration roundtrip testing, and advanced symmetry testing for invariance and equivariance properties. By implementing mock-based approaches and deterministic RNG seeding, this skill helps build fast, reliable, and reproducible test suites for complex simulation-based inference models.
Características Principales
01Critical Keras 3 backend configuration for PyTorch compatibility
02Template-based shape testing for simulator outputs
03Standardized test directory mirroring for src layout
040 GitHub stars
05Advanced symmetry testing for network invariance and equivariance
06Mock-based inference network patterns for fast unit testing
