# TASK: Process All Seeding + Spraying TAP Data **Priority:** HIGH **Agent:** data_pipeline **Filed:** 2026-02-18 **Filed by:** Doug (via Claude Code session) --- ## Context The sync manifest has been reclassified with expanded equipment keywords. Junk fields and garbage dates are now filtered. Only seeding and spraying matter. **DO NOT process:** harvest, tillage, haying, soil, or mapping files. The Topcon yield data is unreliable, and implement tracks have no agronomic value. ## Step 1: Wait for seeding batch to finish A seeding batch is already running (PID check: `ps aux | grep tap_batch`). Wait for it to complete before starting spray. ```bash # Check if seeding is still running ps aux | grep tap_batch | grep -v grep # Check results tail -5 /data/agentpi/logs/tap_processing.log ``` ## Step 2: Run spraying batch Once seeding is done: ```bash python3 /data/Sandbox/AgentPi/scripts/tap_batch_process.py --customer doug --operation spraying ``` This is ~1,125 files. It will take hours. That's fine — just let it run. ## Step 3: Report results After both complete, post a summary: - How many seeding files processed OK vs NO_DATA - How many spraying files processed OK vs NO_DATA - List of fields with successful seeding data - List of fields with successful spraying data ## DO NOT - Do NOT process harvest, tillage, haying, soil, or mapping files - Do NOT retry files that got NO_DATA — those are old BIN files without GPS data - Do NOT modify the sync manifest or processing manifest manually - Do NOT run both batches simultaneously — wait for seeding to finish first