Feature/wellness checkin

[kanikaa-3018]

Key Implementations

• Interactive Wellness index card: A compact dashboard widget summarizing the student's safety limits, stress scores, and observation check-ins.
• Animated Breathing Visualizer: An iOS-Relax style breathing module designed to help students reduce stress, featuring smooth breathing-cycle guidance.
• Wellness check-in drawer: A floating Sheet layout prompting users to log check-ins, record study/meal logs, and set automated wellness safety limits.
• Backend wellness services: API routes (/api/wellness) integrating database cascades, auto-nudge notification checks, and stress analytics.
• Hardening: Accompanying account-deletion cascades, counseling accordion dropdown layouts, and fallback grounding checks.

[nishantharkut]

Requesting changes before merge.

Wellness is important because the original problem statement explicitly includes emotional well-being, burnout patterns, meals, sleep, and routines. But
this PR currently expands PocketBuddy into a broad wellness app. That risks bloating the product and distracting from the core: money behavior -> student
routine signals -> small corrective nudges.

What should stay:

• Deterministic wellness signals tied to actual PocketBuddy data:
  • food gaps
  • late-night transaction activity
  • exam window
  • runway pressure
  • spending velocity
• A lightweight check-in/nudge.
• Bedrock-generated supportive text only if grounded in those deterministic signals.

Blockers:

1. Do not infer “social withdrawal” from cart pool inactivity.
   Not joining a cart pool for 7 days is not enough to imply social isolation. Please rename this to something neutral like “shared-order activity” or
   remove it from the score.

2. Avoid clinical or quasi-clinical framing.
   “Pattern suggests high stress” is acceptable only if the UI makes clear this is a spending/routine signal, not diagnosis. Avoid words or presentation
   that imply mental-health detection from financial logs.

3. Remove aura/emoji/therapy-style language.
   Labels like “Balanced Aura” and soothing emoji quotes make the product feel less serious. Use clear product language:

   • “Stable routine”
   • “Needs attention”
   • “Reset suggested”

4. Keep wellness connected to finance, not as a standalone module.
   Guided breathing, care-plan dialogs, and counseling accordions are too much for the current product surface. If kept, they should be behind a small
   “Support resources” drawer, not a major dashboard module.

5. Check-in notes are sensitive data.
   Free-text stress notes are stored in checkin_logs. That needs explicit privacy handling:

   • user-visible retention policy
   • delete-account coverage
   • no cross-user aggregation
   • avoid showing this in demo unless necessary.

6. Make meal language inclusive.
   Current copy assumes mess/canteen. Day scholars, PG students, and students cooking for themselves may not have mess access. Use “meal” language
   instead:

   • “Have a proper meal”
   • “Mess, home food, or a simple canteen plate works.”

7. Do not make wellness prompts blocking.
   If there is a check-in dialog, it must be dismissible. The product should nudge, not lock the dashboard.

Recommended scope for merge:

• Merge only the deterministic signal engine and a compact dashboard card.
• Keep Bedrock message short and grounded.
• Drop or defer breathing/care-plan/counseling UI until after the core demo is stable.

This should feel like “PocketBuddy notices money/routine stress early,” not “PocketBuddy is a therapy product.”

[nishantharkut]

Thanks, this is much closer to the right direction. The deterministic backend signals are useful and aligned with the original problem statement.

I still do not want to merge this as-is because the UI surface is still too much like a standalone wellness product. PocketBuddy should frame this as
money/routine awareness, not therapy or clinical wellness.

Required changes before merge:

1. Keep the deterministic signals:

   • meal gap
   • late-night activity signal
   • exam window
   • runway pressure
   • spend velocity

2. Keep the compact dashboard nudge/check-in.

3. Remove or hide from the main dashboard:

   • breathing exercise widget
   • soothing quote content
   • “AI Care Plan” framing
   • counseling/support accordion as a major UI surface

4. Replace clinical or therapy-like wording:

   • avoid “stress points”, “care plan”, “wellness index” as the primary product language
   • prefer “routine check”, “meal gap”, “late-night activity”, “reset suggested”, “needs attention”

5. Check-in must remain dismissible. The product should nudge, not block.

6. Free-text check-in notes are sensitive. Keep 30-day retention and make sure delete-account coverage remains intact.

Recommended merge shape:

• deterministic backend signal engine
• compact dashboard routine/meal card
• short grounded Bedrock message
• dismissible check-in

Decision: not merge-ready in current UI shape, but close if reduced.