The bayesflow-testing skill provides domain-specific guidance for writing, modifying, and debugging tests within the BayesFlow ecosystem. It enforces critical configuration patterns, such as setting the Keras backend before imports in conftest.py, and provides standardized templates for shape testing simulators, adapter pipelines, and config roundtrips. By leveraging mock-based testing and invariance/equivariance patterns, this skill helps developers build fast, deterministic, and high-coverage test suites for complex simulation-based inference models.
Key Features
01Invariance and equivariance testing for architectural symmetry
02Standardized CI/CD matrix configuration for GitHub Actions
03Mock-based testing patterns for fast inference network validation
04Automated conftest.py setup for Keras 3 backend management
05Template generation for simulator output and shape verification
060 GitHub stars
