Race Timing System: Configure your real meet

Configure your real meet

1. Copy race_config.json to my_meet.json; do not use the demonstration timestamps/race keys for a real event.
2. Set each race's exact UTC/offset-aware gun_start timestamp and Sports Hub race_key. Separate waves can have separate race entries and gun starts.
3. Configure reader names and antenna ports. For example, antenna 1 = start, antenna 2 = finish, antenna 3 = lap. Multiple antennas at one line should map to the same checkpoint. This is a configurable example, not a required physical antenna layout.
4. point_to_point uses finish_checkpoint; laps uses lap_checkpoint plus a positive laps target. Chip mode also requires start_checkpoint. With a shared chip start/lap line, the first qualified crossing starts the chip clock; subsequent eligible crossings count laps. Gun mode uses the configured gun timestamp without requiring a start read.
5. Set min_seconds to the minimum plausible race duration or lap interval; cluster_gap_ms to the inactivity gap that separates crossing clusters; min_reads to the minimum number of distinct-timestamp reads confirming a cluster. Optional min_rssi rejects weaker reads from calculations but still retains them in the archive. Sample settings are demonstration values and require field calibration.
6. Create a tags CSV with exact headers epc,bib,race_key,runner_name. EPC values must be hexadecimal idHex values, not decimal tag numbers. Each bib belongs to one runner/race in a meet session. Existing EPC assignments cannot be silently changed.
7. Use a fresh database for this meet:

python3 zebra_timing.py --config my_meet.json --db my_meet.sqlite3 tags my_tags.csv

Configuration is frozen to the database to avoid silently changing recorded timing rules. Preserve the old database and start a reviewed new session if rules change. Unknown tags can be assigned later, but their earlier excluded reads remain excluded. Enroll before timing begins. Do not import demonstration reads into the real meet database.