feat(catalog): supplier catalog cleanup, fuzzy matching, and navigation fixes

- Cleaned 137+ fake engine-displacement models from supplier imports
  (v3/v4 scripts: Chevrolet, Ford, Chrysler, Dodge, Jeep, Nissan, etc.)
- Removed 1,251+ corrupted models (INT. prefixes, year-suffix, torque specs,
  empty names, trailing-year variants)
- Migrated supplier tables to master DB (supplier_catalog,
  supplier_catalog_compat, supplier_catalog_interchange)
- Fixed _get_mye_ids_with_parts() to query supplier_catalog_compat from
  master DB so supplier-only vehicles appear for all tenants
- Added fuzzy model matcher with parenthesis stripping, noise suffix removal,
  compact matching, prefix/substring fallback, model aliases, and ±3 year
  proximity
- Matched compat rows: KEEP GREEN +14,152, KNADIAN +3,021, VAZLO +127,500,
  LUK +477, RAYBESTOS +1,743
- Added KNADIAN catalog importer with year-range expansion and future-year
  filtering
- Added VAZLO catalog importer with position parsing and SKU-in-model cleanup
- Added Keep Green, LUK, Yokomitsu, Raybestos catalog importers
- Cache clearing after cleanups (_classify_cache_*, nexus:mye_ids:*,
  nexus:brand_mye_counts:*)

Final match rates:
- KEEP GREEN: 90.3%
- VAZLO: 93.6%
- YOKOMITSU: 100.0%
- KNADIAN: 57.4%
- LUK: 51.0%
- RAYBESTOS: 55.9%

pos/tests/test_bulk_import.py

@@ -0,0 +1,109 @@
+#!/usr/bin/env python3
+"""Test bulk import endpoint — CSV parsing, column normalisation, upsert logic."""
+
+import os
+import sys
+import io
+import csv
+
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+os.environ.setdefault('MASTER_DB_URL', 'postgresql://postgres@/nexus_autopartes')
+os.environ.setdefault('TENANT_DB_URL_TEMPLATE', 'postgresql://postgres@/{db_name}')
+os.environ.setdefault('POS_JWT_SECRET', 'test-secret-12345678901234567890123456789012')
+
+from blueprints.inventory_bp import _to_decimal, _to_int
+
+RED = '\033[91m'
+GREEN = '\033[92m'
+YELLOW = '\033[93m'
+RESET = '\033[0m'
+
+
+def print_result(name, passed, detail=""):
+    status = f"{GREEN}PASS{RESET}" if passed else f"{RED}FAIL{RESET}"
+    print(f"  [{status}] {name}" + (f" — {detail}" if detail else ""))
+
+
+def test_to_decimal():
+    assert _to_decimal("10.5") == 10.5
+    assert _to_decimal("1,000.50") == 1000.50
+    assert _to_decimal("") == 0
+    assert _to_decimal(None, 5) == 5
+    assert _to_decimal("abc", 99) == 99
+    print_result("_to_decimal parsing", True)
+
+
+def test_to_int():
+    assert _to_int("42") == 42
+    assert _to_int("1,000") == 1000
+    assert _to_int("") == 0
+    assert _to_int(None, 7) == 7
+    assert _to_int("abc", 99) == 99
+    print_result("_to_int parsing", True)
+
+
+def test_csv_column_normalisation():
+    """Simulate the column normalisation done in bulk_import_items."""
+    raw_rows = [
+        {"SKU": "ABC123", "Nombre": "Filtro de aceite", "Marca": "Bosch", "Precio": "150", "Cantidad": "10"},
+        {"sku": "DEF456", "name": "Pastillas de freno", "brand": "TRW", "price": "450.50", "stock": "5"},
+    ]
+
+    # Normalise keys
+    for r in raw_rows:
+        normalised = {}
+        for k, v in r.items():
+            nk = str(k).strip().lower().replace(' ', '_')
+            normalised[nk] = v
+        r.clear()
+        r.update(normalised)
+
+    col_map = {
+        'sku': 'part_number', 'numero_de_parte': 'part_number', 'parte': 'part_number',
+        'nombre': 'name', 'producto': 'name', 'descripcion': 'name',
+        'marca': 'brand', 'precio': 'price', 'costo': 'cost',
+        'cantidad': 'stock', 'existencia': 'stock', 'inventario': 'stock',
+        'ubicacion': 'location', 'categoria': 'category',
+        'fabricante': 'make', 'vehiculo': 'make', 'auto': 'make',
+        'modelo': 'model', 'anio': 'year', 'ano': 'year',
+        'motor': 'engine', 'codigo_motor': 'engine_code',
+    }
+    for r in raw_rows:
+        for old_k, new_k in col_map.items():
+            if old_k in r and new_k not in r:
+                r[new_k] = r.pop(old_k)
+
+    assert raw_rows[0]['part_number'] == 'ABC123'
+    assert raw_rows[0]['name'] == 'Filtro de aceite'
+    assert raw_rows[0]['stock'] == '10'
+    assert raw_rows[1]['part_number'] == 'DEF456'
+    assert raw_rows[1]['price'] == '450.50'
+    print_result("CSV column normalisation", True)
+
+
+def test_csv_dict_reader():
+    """Verify csv.DictReader produces the expected structure."""
+    csv_text = "sku,name,brand,price,stock\nABC123,Filtro,Bosch,150,10\nDEF456,Pastillas,TRW,450,5"
+    f = io.StringIO(csv_text)
+    reader = csv.DictReader(f)
+    rows = list(reader)
+    assert len(rows) == 2
+    assert rows[0]['sku'] == 'ABC123'
+    assert rows[1]['price'] == '450'
+    print_result("csv.DictReader parsing", True)
+
+
+def run_all():
+    print("\nBulk Import Tests")
+    print("=" * 40)
+    test_to_decimal()
+    test_to_int()
+    test_csv_column_normalisation()
+    test_csv_dict_reader()
+    print("=" * 40)
+    print("Done.\n")
+
+
+if __name__ == '__main__':
+    run_all()