Report summary totals by migraine episode, not by day#2268

Bug: Migraine Medication Days are undercounted for multi-day attacks

Expected behavior:
“Migraine Medication Days” should count every unique calendar day within the selected reporting period on which any medication was logged as taken.

For each individual medication, the report should count the number of unique dates within the reporting period on which that medication was taken.

This calculation should be based on the medication administration date/time itself, regardless of when the parent migraine attack started or ended.

Actual behavior:
The Impact Report undercounts medication days when medications are logged during a long multi-day attack.

Medication use attached to a multi-day attack is not attributed to the actual calendar date on which the medication was taken, but instead summarized into a single “day” linked to the start day of the attack. This is factually inaccurate reporting.

Expected calculation:
Overall Migraine Medication Days = count of DISTINCT calendar dates within the report period where at least one medication was logged.

Per-medication count = count of DISTINCT calendar dates within the report period where that specific medication was logged.

The medication-day count should NOT depend on the start date of the migraine attack containing that medication entry.

To note, in general the reports seem like they were set up for generation pre-variability tracking and have bugs/missing features around accurately reporting on multi day attacks with variability recorded. See: https://suggestions.migrainebuddy.com/2269