Add iAUC (incremental area under the curve) metric to health dashboard

Summary

Add iAUC (incremental area under the curve) computation to the health dashboard. iAUC measures total glucose exposure above baseline over a meal window, which is a better proxy for insulin demand than peak glucose (spike). A sharp spike that resolves quickly has a small iAUC; a moderate rise that lingers has a large one.

Context

Ben's CGM data shows fruit causes high peak spikes but he had his lowest insulin (20) during a fruititarian + high-running period. The hypothesis is that fruit spikes are sharp but brief (low iAUC = low insulin demand), especially with GLUT4 uptime from running. The dashboard currently only shows peak-based spike metrics, which may misrepresent the metabolic impact.

What to change

Analyze.hs

1. Add an iAUC computation function. The algorithm:

• For each meal, identify the baseline BG (pre-meal reading, same as current mealSpike baseline logic)
• Compute the window (until next meal or 3h, same as mealSpike windowing)
• For each CGM reading in the window where glucose > baseline, sum up (glucose - baseline) * interval_minutes
• This gives incremental AUC in mg/dL·min
• Return the iAUC alongside existing spike data

2. Add MealIauc or extend MealSpikeRecord with an msrIauc :: Double field (mg/dL·min).

3. Add computeMealIaucRankings that ranks meals/categories by average iAUC (similar to computeFoodRankings but using iAUC instead of peak spike).

4. Add iAUC to DailyCgmStats — total daily iAUC across all meals would be a useful aggregate metric.

Web.hs

1. Add an iAUC column to the existing food rankings table and meal rankings table. 2. Optionally add a new section or integrate into the existing food rankings to show the iAUC perspective — meals ranked by iAUC vs by peak spike, so the user can see which foods are truly high-burden vs just high-peak.

Score.hs

No changes needed for now — scoring still uses spike thresholds. iAUC is informational.

CGM data format

CSV in /var/health/cgm.csv:

Date/Time,Blood Glucose (mg/dL),Meal Time,Sources
2026-02-19 22:21:00,122,,

5-minute intervals. Meals are in /var/health/meals.jsonl.

Acceptance criteria

• computeMealIaucRankings exists and computes iAUC per meal from CGM+meal data
• iAUC values appear on the /health dashboard (at minimum in the food rankings and meal rankings tables)
• Existing tests pass, new tests cover iAUC computation
• iAUC unit is mg/dL·min (standard in glycemic research)