Objective
Implement a lightweight C++ automated testing harness (or integrate a header-only framework like Catch2/Google Test) to verify the internal logic of the state-machine pre-filter, telemetry normalization, and action-clamping boundaries.
Scope & Key Tasks
- Telemetry Scaling Verification: Assert that simulated packet sizes and F2 sliding window sum-of-squares are normalized correctly to the
[0.0, 1.0] float interval before calling ONNX inference.
- Safety Boundary Enforcement: Assert that extreme values output by the neural network (e.g., SQ threshold > 700, penalty multiplier < 10) are successfully clamped to the configured safety parameters (e.g., 650 and 20.0).
- Action Space Dynamic Routing: Test that both 3D and 4D output tensors from dummy Ort models are correctly parsed, and S0 sampling rate is calculated accurately for the 3D case.
- FSM State Transitions: Validate that mock virtual CPU budgets trigger state transitions (S0 -> S1 -> S2 -> S3) in a deterministic fashion.
Definition of Done (DoD)
- Running tests (e.g., via
./manage_build.sh --test or ./run_tests.sh) successfully evaluates all assertions.
- Test coverage covers 100% of the boundary clamping conditions and telemetry scaling equations inside
rl_bridge.cpp and pre_filter.cpp.
Objective
Implement a lightweight C++ automated testing harness (or integrate a header-only framework like Catch2/Google Test) to verify the internal logic of the state-machine pre-filter, telemetry normalization, and action-clamping boundaries.
Scope & Key Tasks
[0.0, 1.0]float interval before calling ONNX inference.Definition of Done (DoD)
./manage_build.sh --testor./run_tests.sh) successfully evaluates all assertions.rl_bridge.cppandpre_filter.cpp.