Race Timing System: Review and export

Review and export

python3 zebra_timing.py --config my_meet.json --db my_meet.sqlite3 status
python3 zebra_timing.py --config my_meet.json --db my_meet.sqlite3 decide 104 dnf
python3 zebra_timing.py --config my_meet.json --db my_meet.sqlite3 decide 104 clear
python3 zebra_timing.py --config my_meet.json --db my_meet.sqlite3 export reviewed_results.csv --basis gun

decide supports DNS/DNF/DQ or clear, which restores calculated state. Decisions remain audited. Exports include calculated finishers and manually reviewed non-finishers; pending runners are omitted. --basis chip requires a captured chip start for every finisher; there is no silent fallback to gun time. Exports retain one selected elapsed time for the current Hub API. Local status retains both time types and lap elapsed values.

Before publishing, compare crossings against backup observations/video and review unknown tags and pending runners. All computed finishers are provisional. A tag re-entering a read zone after the gap/minimum interval may represent a genuine lap or a false crossing; this first version cannot distinguish them by direction. Placement/RSSI rules alone do not prove the exact physical finish-line crossing.