feat(products): improve replenishment-aware shopping suggestions

Replace product weight and repurchase intent fields with per-package remaining levels and inventory-first restock signals. Enrich shopping suggestions with usage-aware replenishment scoring so the frontend and LLM can prioritize real gaps and near-empty staples more reliably.

backend/tests/test_products.py

@@ -187,11 +187,15 @@ def test_list_inventory_product_not_found(client):
 
 def test_create_inventory(client, created_product):
     pid = created_product["id"]
-    r = client.post(f"/products/{pid}/inventory", json={"is_opened": False})
+    r = client.post(
+        f"/products/{pid}/inventory",
+        json={"is_opened": True, "remaining_level": "medium"},
+    )
     assert r.status_code == 201
     data = r.json()
     assert data["product_id"] == pid
-    assert data["is_opened"] is False
+    assert data["is_opened"] is True
+    assert data["remaining_level"] == "medium"
 
 
 def test_create_inventory_product_not_found(client):
@@ -204,11 +208,16 @@ def test_parse_text_accepts_numeric_strength_levels(client, monkeypatch):
 
     class _FakeResponse:
         text = (
-            '{"name":"Test Serum","actives":[{"name":"Niacinamide","percent":10,'
+            '{"name":"Test Serum","category":"serum","recommended_time":"both",'
+            '"leave_on":true,"actives":[{"name":"Niacinamide","percent":10,'
             '"functions":["niacinamide"],"strength_level":2,"irritation_potential":1}]}'
         )
 
-    monkeypatch.setattr(products_api, "call_gemini", lambda **kwargs: _FakeResponse())
+    monkeypatch.setattr(
+        products_api,
+        "call_gemini",
+        lambda **kwargs: (_FakeResponse(), None),
+    )
 
     r = client.post("/products/parse-text", json={"text": "dummy input"})
     assert r.status_code == 200