Fix procurement data structure and add price trends

Fixed critical structural issues in procurement fixture:

1. **Removed duplicate nested items** (32 items):
   - Previous enhancement script incorrectly added 'items' arrays
     inside purchase_orders
   - Procurement service uses separate purchase_order_items table
   - Removed all nested 'items' to match PurchaseOrderItem model

2. **Added price trends to existing PO items** (10 items updated):
   - Harina T55: +8% (€0.85 → €0.92)
   - Harina T65: +6% (€0.95 → €1.01)
   - Mantequilla: +12% (€6.50 → €7.28) - highest increase
   - Leche: -3% (€0.95 → €0.92) - seasonal surplus
   - Levadura: +4% (€4.20 → €4.37)
   - Azúcar: +2% (€1.10 → €1.12) - stable

3. **Recalculated PO totals** based on updated item prices

This enables procurement AI insights:
- Price trend analysis and alerts
- Supplier performance comparison
- Cost optimization recommendations

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

shared/demo/fixtures/professional/fix_procurement_structure.py

@@ -0,0 +1,208 @@
+#!/usr/bin/env python3
+"""
+Fix Procurement Data Structure and Add Realistic Price Trends
+
+Issues to fix:
+1. Remove nested 'items' arrays from purchase_orders (wrong structure)
+2. Use existing purchase_order_items table structure at root level
+3. Add price trends to existing PO items
+4. Align PO items with actual inventory stock conditions
+"""
+
+import json
+from pathlib import Path
+import random
+
+# Set seed for reproducibility
+random.seed(42)
+
+# Price trend data (realistic price movements over time)
+# These match the 6 ingredients we track in inventory
+PRICE_TRENDS = {
+    "10000000-0000-0000-0000-000000000001": {  # Harina T55
+        "name": "Harina de Trigo T55",
+        "base_price": 0.85,
+        "current_price": 0.92,  # +8% over 90 days
+        "trend": 0.08
+    },
+    "10000000-0000-0000-0000-000000000002": {  # Harina T65
+        "name": "Harina de Trigo T65",
+        "base_price": 0.95,
+        "current_price": 1.01,  # +6%
+        "trend": 0.06
+    },
+    "10000000-0000-0000-0000-000000000011": {  # Mantequilla
+        "name": "Mantequilla sin Sal",
+        "base_price": 6.50,
+        "current_price": 7.28,  # +12% (highest increase)
+        "trend": 0.12
+    },
+    "10000000-0000-0000-0000-000000000012": {  # Leche
+        "name": "Leche Entera Fresca",
+        "base_price": 0.95,
+        "current_price": 0.92,  # -3% (seasonal surplus)
+        "trend": -0.03
+    },
+    "10000000-0000-0000-0000-000000000021": {  # Levadura
+        "name": "Levadura Fresca",
+        "base_price": 4.20,
+        "current_price": 4.37,  # +4%
+        "trend": 0.04
+    },
+    "10000000-0000-0000-0000-000000000032": {  # Azúcar
+        "name": "Azúcar Blanco",
+        "base_price": 1.10,
+        "current_price": 1.12,  # +2% (stable)
+        "trend": 0.02
+    }
+}
+
+
+def calculate_price_for_date(ingredient_id: str, days_ago: int) -> float:
+    """Calculate historical price based on trend"""
+    if ingredient_id not in PRICE_TRENDS:
+        return None
+
+    trend_data = PRICE_TRENDS[ingredient_id]
+    base = trend_data["base_price"]
+    total_trend = trend_data["trend"]
+
+    # Apply trend proportionally
+    # If 90 days trend is +8%, then 45 days ago had +4% from base
+    trend_factor = 1 + (total_trend * (90 - days_ago) / 90)
+
+    # Add small variability (±2%)
+    variability = random.uniform(-0.02, 0.02)
+
+    price = base * trend_factor * (1 + variability)
+    return round(price, 2)
+
+
+def parse_days_ago(date_str: str) -> int:
+    """Parse BASE_TS marker to extract days ago"""
+    if not date_str or 'BASE_TS' not in date_str:
+        return 30
+
+    if '- ' in date_str:
+        parts = date_str.split('- ')[1].strip()
+        if 'd' in parts:
+            try:
+                return int(parts.split('d')[0])
+            except:
+                pass
+        elif 'h' in parts:
+            return 0  # Same day
+    elif '+ ' in date_str:
+        return 0  # Future order, use current price
+
+    return 0  # BASE_TS alone = today
+
+
+def main():
+    fixture_path = Path(__file__).parent / "07-procurement.json"
+
+    print("🔧 Fixing Procurement Data Structure...")
+    print()
+
+    # Load existing data
+    with open(fixture_path, 'r', encoding='utf-8') as f:
+        data = json.load(f)
+
+    purchase_orders = data.get('purchase_orders', [])
+    po_items = data.get('purchase_order_items', [])
+
+    print(f"📦 Found {len(purchase_orders)} purchase orders")
+    print(f"📋 Found {len(po_items)} PO items")
+    print()
+
+    # Step 1: Remove nested 'items' arrays from POs (wrong structure)
+    items_removed = 0
+    for po in purchase_orders:
+        if 'items' in po:
+            items_removed += len(po['items'])
+            del po['items']
+
+    if items_removed > 0:
+        print(f"✓ Removed {items_removed} nested items arrays (wrong structure)")
+        print()
+
+    # Step 2: Update existing PO items with realistic price trends
+    items_updated = 0
+
+    for item in po_items:
+        ingredient_id = item.get('inventory_product_id')
+
+        if ingredient_id in PRICE_TRENDS:
+            # Find the PO to get order date
+            po_id = item.get('purchase_order_id')
+            po = next((p for p in purchase_orders if p['id'] == po_id), None)
+
+            if po:
+                order_date = po.get('order_date', 'BASE_TS')
+                days_ago = parse_days_ago(order_date)
+
+                # Calculate price for that date
+                historical_price = calculate_price_for_date(ingredient_id, days_ago)
+
+                if historical_price:
+                    # Update item with historical price
+                    ordered_qty = float(item.get('ordered_quantity', 0))
+                    item['unit_price'] = historical_price
+                    item['line_total'] = round(ordered_qty * historical_price, 2)
+
+                    items_updated += 1
+
+    print(f"✓ Updated {items_updated} PO items with price trends")
+    print()
+
+    # Step 3: Recalculate PO totals based on updated items
+    for po in purchase_orders:
+        po_id = po['id']
+        po_items_for_this_po = [item for item in po_items if item.get('purchase_order_id') == po_id]
+
+        if po_items_for_this_po:
+            # Calculate subtotal from items
+            subtotal = sum(float(item.get('line_total', 0)) for item in po_items_for_this_po)
+            tax_rate = 0.21  # 21% IVA in Spain
+            tax = subtotal * tax_rate
+
+            # Keep existing shipping cost or default
+            shipping = float(po.get('shipping_cost', 15.0 if subtotal < 500 else 20.0))
+            discount = float(po.get('discount_amount', 0.0))
+
+            total = subtotal + tax + shipping - discount
+
+            po['subtotal'] = round(subtotal, 2)
+            po['tax_amount'] = round(tax, 2)
+            po['shipping_cost'] = round(shipping, 2)
+            po['discount_amount'] = round(discount, 2)
+            po['total_amount'] = round(total, 2)
+
+    print(f"✓ Recalculated totals for {len(purchase_orders)} purchase orders")
+    print()
+
+    # Save fixed data
+    with open(fixture_path, 'w', encoding='utf-8') as f:
+        json.dump(data, f, indent=2, ensure_ascii=False)
+
+    print("=" * 60)
+    print("✅ PROCUREMENT STRUCTURE FIXED")
+    print("=" * 60)
+    print()
+    print("🎯 Changes Applied:")
+    print(f"   • Removed {items_removed} incorrectly nested items")
+    print(f"   • Updated {items_updated} PO items with price trends")
+    print(f"   • Recalculated {len(purchase_orders)} PO totals")
+    print()
+    print("📊 Price Trends Applied:")
+    for ing_id, data in PRICE_TRENDS.items():
+        direction = "↑" if data["trend"] > 0 else "↓"
+        print(f"   {direction} {data['name']}: {data['trend']*100:+.1f}%")
+    print()
+    print("✅ Data structure now matches PurchaseOrderItem model")
+    print("✅ Price trends enable procurement AI insights")
+    print()
+
+
+if __name__ == "__main__":
+    main()