Spaces:
Sleeping
Sleeping
Download scripts/experiments/check_farzone_gates.bat from masterdanh/poolcoach-analyzer: direct link, hf CLI and curl.
- Browser
- Download file 2.12 kB
-
https://huggingface.co/spaces/masterdanh/poolcoach-analyzer/resolve/main/scripts/experiments/check_farzone_gates.bat
- Command line
-
hf download hf://spaces/masterdanh/poolcoach-analyzer/scripts/experiments/check_farzone_gates.bat
-
curl -L -o check_farzone_gates.bat https://huggingface.co/spaces/masterdanh/poolcoach-analyzer/resolve/main/scripts/experiments/check_farzone_gates.bat
2.12 kB
| @echo off | |
| REM ============================================================ | |
| REM FAR-ZONE HAND-OFF 2 -- the clip-measured gates (28/08/2026) | |
| REM Steps 1, 2 and 3 of design doc 7_FarZone_Detection_BallID.md | |
| REM section 5, BRIEF 28/08 (hand-off 2). | |
| REM | |
| REM Why this run happens: | |
| REM Unit tests can prove the two-band association and the rack | |
| REM geometry on synthetic input, but three acceptance numbers | |
| REM only exist in the footage: ball 5 wedged in the far right | |
| REM pocket of shot_02, the 0.5 s hole in the cue track of | |
| REM shot_05 against the far cushion, and which shots are break | |
| REM racks. This driver measures those three on the SHIPPED | |
| REM analyze front end and reports PASS/FAIL against the | |
| REM thresholds the design doc fixed. It tunes nothing. | |
| REM | |
| REM Input clips: the ones the probe already cut on 28/08 with the | |
| REM worker's own cut_clip, so these frames are the frames the | |
| REM analyze job read: | |
| REM D:\poolcoach\out_probe_farzone_20260828\clips\shot_*.mp4 | |
| REM Same corners as run A2b (14/08) and every regeneration since: | |
| REM [[527.5,869.7],[1380.1,864.0],[1232.3,278.5],[673.0,272.8]] | |
| REM | |
| REM CPU environment: poolcoach-cv-env carries torch 2.12.1+cpu, | |
| REM so there is no CUDA to fall back to. Expect a couple of | |
| REM minutes for the five clips at imgsz 960. | |
| REM | |
| REM Nothing is written except farzone_gates.json next to the | |
| REM clips. D:\Khoa luan is never opened. | |
| REM | |
| REM No app and no Redis has to be running: the driver loads YOLO | |
| REM itself and imports scripts\cv_worker.py only for the shipped | |
| REM constants, clip_frame_wh and comp_homography_cls. | |
| REM ============================================================ | |
| setlocal | |
| set PYTHONUTF8=1 | |
| set PYTHONDONTWRITEBYTECODE=1 | |
| set POOLCOACH_CV_DEVICE=cpu | |
| set POOLCOACH_APP_ROOT=D:\poolcoach\poolcoach-app | |
| set POOLCOACH_GATE_CLIPS=D:\poolcoach\out_probe_farzone_20260828\clips | |
| set POOLCOACH_GATE_OUT=D:\poolcoach\out_probe_farzone_20260828 | |
| set PY=D:\poolcoach\poolcoach-cv-env\Scripts\python.exe | |
| "%PY%" "D:\poolcoach\launchers\check_farzone_gates.py" | |
| echo [check-farzone-gates] exit code %ERRORLEVEL% | |
| endlocal | |