replay-auto-aim

name	replay-auto-aim
description	Reproduce and isolate auto-aim detection, tracking, solving, aiming, and planning issues from recorded video in this repository. Use for false positives, missed armors, tracker state changes, aim jitter, fire decisions, or latency regressions.

Replay Auto-Aim

Use recorded inputs before using a camera or gimbal. The goal is a tight, repeatable loop that can go red on the reported symptom.

Workflow

1. State the symptom as an observable assertion: frame range, expected armor/target, tracker state, command tolerance, or timing threshold.
2. Verify assets/demo/demo.avi and assets/demo/demo.txt, select a small frame range, and run from the repository root.
3. Choose the narrowest executable:
   • detector_video_test for YOLO versus traditional detection and corner geometry;
   • auto_aim_test for YOLO, Solver, Tracker, and Aimer together;
   • planner_test_offline for trajectory and fire planning without camera or gimbal.
4. If OpenVINO GPU is unavailable, use a temporary copied YAML with device: CPU; never overwrite the robot config during diagnosis.
5. Compare one variable at a time: model, detector mode, config threshold, frame interval, or calibration. Record frame index, detection count, tracker state, command, and stage timings.
6. Minimise the frame range until removing any remaining input makes the symptom disappear. Turn that replay into a regression test at a public seam before changing implementation.

Evidence and boundaries

• auto_aim_test and detector_video_test use OpenCV windows and require a display; report headless limitations explicitly.
• Plotter sends JSON to UDP 127.0.0.1:9870; capture or disable that observer only when it is part of the repro.
• Do not infer planner correctness from a process that merely stays alive. Check target yaw/pitch, trajectory values, and fire flag.
• Remove temporary debug logging and rerun the original frame range after the fix. Use $diagnosing-bugs for difficult reproductions and $tdd for the retained regression test.

Read replay-matrix.md for command syntax and pipeline facts.