bakery-ia/services/demo_session/scripts/README.md

Dashboard Demo Seed Scripts

Comprehensive demo data seeding scripts for the JTBD-aligned dashboard.

🎯 Purpose

These scripts create realistic demo data to showcase all dashboard features and user flows:

• Time-based Action Queue (URGENT/TODAY/WEEK grouping)
• AI Prevented Issues (showcasing AI value)
• Execution Progress Tracking (production/deliveries/approvals)
• Stock Receipt Modal workflows
• Health Status tri-state checklist
• All Alert Types with full enrichment

📋 Available Scripts

1. seed_dashboard_comprehensive.py ⭐ RECOMMENDED

Comprehensive dashboard demo covering ALL scenarios

What it seeds:

• 🔴 URGENT actions (<6h deadline): 3 alerts

  • PO approval escalation (72h aged, 2h deadline)
  • Delivery overdue (4h late, supplier contact needed)
  • Batch at risk (missing ingredients, 5h window)

• 🟡 TODAY actions (<24h deadline): 3 alerts

  • PO approval needed (dairy products, 20h deadline)
  • Delivery arriving soon (8h, prep required)
  • Low stock warning (yeast, order today recommended)

• 🟢 THIS WEEK actions (<7d deadline): 2 alerts

  • Weekend demand surge prediction
  • Stock receipt incomplete (2 days old)

• ✅ AI PREVENTED ISSUES: 3 alerts

  • Prevented stockout (PO created, €250 saved)
  • Prevented waste (production adjusted, €120 saved)
  • Prevented delay (batches rescheduled, €85 saved)

Expected Dashboard State:

Health Status: YELLOW (actions needed)
├─ ⚡ AI Handled: 3 issues (€455 saved)
└─ ⚠️  Needs You: 8 actions

Action Queue: 8 total actions
├─ 🔴 URGENT: 3
├─ 🟡 TODAY: 3
└─ 🟢 WEEK: 2

AI Impact: €455 in prevented costs

Usage:

# Quick start
python services/demo_session/scripts/seed_dashboard_comprehensive.py

# With custom tenant
DEMO_TENANT_ID=your-tenant-id python services/demo_session/scripts/seed_dashboard_comprehensive.py

# With custom RabbitMQ
RABBITMQ_URL=amqp://user:pass@host:5672/ python services/demo_session/scripts/seed_dashboard_comprehensive.py

---

2. seed_enriched_alert_demo.py

Legacy enriched alert demo (basic scenarios)

Seeds 5 basic alert types with automatic enrichment:

• Low stock (AI handled)
• Supplier delay (critical)
• Waste trend (standard)
• Forecast anomaly (info)
• Equipment maintenance (medium)

Usage:

python services/demo_session/scripts/seed_enriched_alert_demo.py

Note: For full dashboard testing, use seed_dashboard_comprehensive.py instead.

---

🚀 Quick Start

Prerequisites

1. RabbitMQ running:

   kubectl get pods | grep rabbitmq
   # Should show: rabbitmq-0  1/1  Running
   

2. Alert Processor service running:

   kubectl get pods -l app.kubernetes.io/name=alert-processor-service
   # Should show: alert-processor-service-xxx  1/1  Running
   

3. Python dependencies:

   pip install -r requirements.txt
   

Running the Demo

# 1. Navigate to project root
cd /path/to/bakery-ia

# 2. Load environment variables (if needed)
source .env

# 3. Run comprehensive dashboard seeder
python services/demo_session/scripts/seed_dashboard_comprehensive.py

Expected Output

================================================================================
🚀 SEEDING COMPREHENSIVE DASHBOARD DEMO DATA
================================================================================

📋 Configuration:
   Tenant ID: demo-tenant-bakery-ia
   RabbitMQ: amqp://guest:guest@localhost:5672/

📊 Dashboard Scenarios to Seed:
   🔴 URGENT actions (<6h): 3
   🟡 TODAY actions (<24h): 3
   🟢 WEEK actions (<7d): 2
   ✅ AI Prevented Issues: 3
   📦 Total Alerts: 11

📤 Publishing Alerts:
────────────────────────────────────────────────────────────────────────────────
   1. ✅ [🔴 URGENT] URGENT: PO Approval Needed - Yeast Supplier
   2. ✅ [🔴 URGENT] Delivery Overdue: Flour Delivery
   3. ✅ [🔴 URGENT] Batch At Risk: Missing Ingredients
   4. ✅ [🟡 TODAY] PO Approval: Butter & Dairy Products
   5. ✅ [🟡 TODAY] Delivery Arriving in 8 Hours: Sugar & Ingredients
   6. ✅ [🟡 TODAY] Low Stock: Fresh Yeast
   7. ✅ [🟢 WEEK] Weekend Demand Surge Predicted
   8. ✅ [🟢 WEEK] Stock Receipt Pending: Flour Delivery
   9. ✅ [✅ PREVENTED] ✅ AI Prevented Stockout: Flour
  10. ✅ [✅ PREVENTED] ✅ AI Prevented Waste: Reduced Monday Production
  11. ✅ [✅ PREVENTED] ✅ AI Prevented Delay: Rescheduled Conflicting Batches

✅ Published 11/11 alerts successfully

================================================================================
🎉 DASHBOARD DEMO SEEDED SUCCESSFULLY!
================================================================================

---

🔍 Verification

1. Check Alert Processing

# View alert-processor logs (real-time)
kubectl logs -f deployment/alert-processor-service | grep 'enriched_alert'

# Should see:
# alert_enriched  alert_id=xxx  type_class=action_needed  priority_score=92
# alert_enriched  alert_id=xxx  type_class=prevented_issue  priority_score=35

2. Access Dashboard

# Port forward if needed
kubectl port-forward svc/frontend-service 3000:3000

# Open browser
open http://localhost:3000/dashboard

3. Verify Dashboard Sections

✅ Health Status Card:

• Should show YELLOW status
• Tri-state checklist items visible
• AI prevented issues badge showing "3 issues prevented"

✅ Action Queue Card:

• 🔴 URGENT section with 3 items (2h countdown visible)
• 🟡 TODAY section with 3 items
• 🟢 THIS WEEK section with 2 items

✅ Orchestration Summary:

• "User Needed: 8" in yellow (clickable)
• "AI Prevented: 3 issues" in green badge

✅ AI Impact Card:

• Shows €455 total savings
• Lists 3 prevented issues

---

🧪 Testing Scenarios

Scenario 1: Urgent Action with Countdown

Test: PO Approval Escalation (2h deadline)

1. Navigate to dashboard
2. Find "URGENT: PO Approval Needed - Yeast Supplier" in 🔴 URGENT section
3. Verify countdown timer shows ~2 hours
4. Click "Approve" button
5. Verify alert moves to resolved/archived

Expected Smart Actions:

• ✅ Approve PO (primary, green)
• ⚙️ Modify PO (secondary)
• ❌ Reject PO (danger, red)

---

Scenario 2: Stock Receipt Modal

Test: Mark Delivery as Received

1. Find "Delivery Arriving in 8 Hours" in 🟡 TODAY section
2. Click "Mark as Received" button
3. Stock Receipt Modal should open:
   • Shows PO details (supplier, items)
   • Lot input fields for each line item
   • Quantity validation (lots must sum to actual)
   • Mandatory expiration dates
4. Fill in lot details:
   • Lot number (e.g., "LOT-2024-089")
   • Quantity per lot
   • Expiration date (required)
   • Warehouse location
5. Click "Confirm Receipt"
6. Verify inventory is updated

Expected Validation:

• ❌ Error if lot quantities don't sum to actual quantity
• ❌ Error if expiration date missing
• ✅ Success toast on confirmation

---

Scenario 3: AI Prevented Issue Showcase

Test: View AI Value Proposition

1. Find "✅ AI Prevented Stockout: Flour" in alert list

2. Verify prevented issue badge (⚡ lightning bolt)

3. Click to expand reasoning

4. Should show:

   • AI action taken: "Purchase order created automatically"
   • Savings: €250
   • Reasoning: "Detected stock would run out in 1.8 days..."
   • Business impact: "Secured 4 production batches"

5. Navigate to Orchestration Summary Card

6. Verify "AI Prevented: 3 issues" badge shows €455 total

---

Scenario 4: Call Supplier (External Action)

Test: Supplier contact integration

1. Find "Delivery Overdue: Flour Delivery" in 🔴 URGENT section
2. Click "Call Supplier" button
3. Expected behavior:
   • Phone dialer opens with +34-555-5678
   • OR clipboard copies phone number
   • Toast notification confirms action

Metadata displayed:

• Supplier: Harinera San José
• Phone: +34-555-5678
• Email: pedidos@harinerasj.es
• Hours overdue: 4

---

Scenario 5: Navigation to Linked Pages

Test: Smart action navigation

1. Find "Batch At Risk: Missing Ingredients" in 🔴 URGENT
2. Click "View Production" button
3. Should navigate to: /production?batch_id=batch-chocolate-cake-evening
4. Production page shows batch details
5. Missing ingredients highlighted

---

🐛 Troubleshooting

Issue: Alerts not appearing in dashboard

Check:

# 1. Verify RabbitMQ is running
kubectl get pods | grep rabbitmq

# 2. Check alert-processor logs
kubectl logs deployment/alert-processor-service --tail=100

# 3. Verify alerts.exchange exists
# (Check RabbitMQ management UI: localhost:15672)

# 4. Check for errors in seeder output
python services/demo_session/scripts/seed_dashboard_comprehensive.py 2>&1 | grep ERROR

Common Fixes:

• Restart alert-processor: kubectl rollout restart deployment/alert-processor-service
• Re-run seeder with debug: python -u services/demo_session/scripts/seed_dashboard_comprehensive.py
• Check RabbitMQ queue: raw_alerts_queue should have consumers

---

Issue: Countdown timer not working

Check:

# Verify urgency_context.auto_action_countdown_seconds is set
# Should be in alert metadata

Fix: Re-run seeder to ensure urgency_context is populated

---

Issue: Stock Receipt Modal not opening

Check:

# 1. Verify modal component is imported in DashboardPage
grep -r "StockReceiptModal" frontend/src/pages/app/DashboardPage.tsx

# 2. Check browser console for errors
# Look for: "delivery:mark-received event not handled"

# 3. Verify smartActionHandlers.ts is loaded

Fix: Ensure event listener is registered in DashboardPage.tsx

---

📊 Data Reference

Alert Type Classes

• action_needed - Requires user decision (yellow)
• prevented_issue - AI already handled (blue/green)
• trend_warning - Proactive insight (info)
• escalation - Time-sensitive with countdown (red)
• information - Pure informational (gray)

Priority Levels

• critical (90-100) - Needs decision in 2 hours
• important (70-89) - Needs decision today
• standard (50-69) - Review when convenient
• info (0-49) - For awareness

Time Groups

• 🔴 URGENT - Deadline <6 hours
• 🟡 TODAY - Deadline <24 hours
• 🟢 THIS WEEK - Deadline <7 days

---

🔄 Resetting Demo Data

To clear all demo alerts and start fresh:

# 1. Delete all alerts for demo tenant
# (This requires admin access to alert-processor DB)

# 2. Or restart alert-processor (clears in-memory cache)
kubectl rollout restart deployment/alert-processor-service

# 3. Re-run seeder
python services/demo_session/scripts/seed_dashboard_comprehensive.py

---

📝 Notes

• Automatic Enrichment: All alerts are automatically enriched by alert-processor service
• Priority Scoring: Multi-factor algorithm considers urgency, impact, user agency
• Smart Actions: Dynamically generated based on alert type and context
• Real-time Updates: Dashboard subscribes to SSE for live alert updates
• i18n Support: All alerts support EN/ES/EU languages

---

🚀 Next Steps

After seeding:

1. Test all smart actions (approve, reject, call, navigate, etc.)
2. Verify performance (<500ms dashboard load time)
3. Test responsive design (mobile, tablet, desktop)
4. Check translations (switch language in UI)
5. Test SSE updates (create new alert, see real-time update)

---

🤝 Contributing

To add new demo scenarios:

1. Edit seed_dashboard_comprehensive.py
2. Add new alert to appropriate function (create_urgent_actions(), etc.)
3. Include full metadata for enrichment
4. Test enrichment output
5. Update this README with new scenario

---

📚 Related Documentation

• Alert Type Schemas
• Dashboard Service API
• Smart Action Handlers
• JTBD Implementation Status