Unmatched food terms aggregated across all patients. Multi-word phrases (2–3 word) are prioritized — they are more likely to be real food names missing from the dictionary.

Trigger Backfill

Re-tags all logs with empty triggers and saves them back to Supabase. Run this after updating the food dictionary.

Cross-referencing food entries against symptom severity, timing, stress level, and menstrual cycle phase across all logged episodes. Each trigger shows percentage of high-severity episodes where that category was present.