What problem does it solve? Building Vision AI pipelines with DX-Stream GStreamer elements is error-prone: mismatched preprocess-id/inference-id values, missing queue elements, wrong postprocess libraries, and incorrect paths cause deadlocks, segfaults, and silent failures that are hard to diagnose. ## Core Features & Use Cases - Static Analysis Without Hardware: Check shell/Python syntax, GStreamer element registration, postprocess library exports, and path resolution without needing an NPU. - Property Validation: Verify preprocess-id and inference-id consistency across DxPreprocess, DxInfer, and DxPostprocess elements, confirm queue placement between dx elements, and ensure model and library paths are absolute. - Smoke and Performance Testing: Run limited-frame pipeline tests on NPU hardware, validate headless/DISPLAY fallback behavior, and measure FPS and latency with GStreamer tracers. - Use Case: After writing a new object detection pipeline script, run the validation checks to catch a missing queue between dxinfer and dxpostprocess before it deadlocks in production. ## Quick Start Validate my dx_stream pipeline script run_yolov5s.sh for correctness and runtime readiness.
