Refactor retailer normalization outputs

enrich_giant.py

@@ -16,6 +16,9 @@ OUTPUT_FIELDS = [
     "retailer",
     "order_id",
     "line_no",
+    "normalized_row_id",
+    "normalized_item_id",
+    "normalization_basis",
     "observed_item_key",
     "order_date",
     "retailer_item_id",
@@ -44,13 +47,21 @@ OUTPUT_FIELDS = [
     "size_unit",
     "pack_qty",
     "measure_type",
+    "normalized_quantity",
+    "normalized_quantity_unit",
     "is_store_brand",
+    "is_item",
     "is_fee",
     "is_discount_line",
     "is_coupon_line",
     "price_per_each",
+    "price_per_each_basis",
+    "price_per_count",
+    "price_per_count_basis",
     "price_per_lb",
+    "price_per_lb_basis",
     "price_per_oz",
+    "price_per_oz_basis",
     "parse_version",
     "parse_notes",
 ]
@@ -329,6 +340,65 @@ def derive_prices(item, measure_type, size_value="", size_unit="", pack_qty=""):
     return price_per_each, price_per_lb, price_per_oz
 
 
+def derive_normalized_quantity(size_value, size_unit, pack_qty, measure_type):
+    parsed_size = to_decimal(size_value)
+    parsed_pack = to_decimal(pack_qty) or Decimal("1")
+
+    if parsed_size not in (None, Decimal("0")) and size_unit:
+        return format_decimal(parsed_size * parsed_pack), size_unit
+    if parsed_pack not in (None, Decimal("0")) and measure_type == "count":
+        return format_decimal(parsed_pack), "count"
+    if measure_type == "each":
+        return "1", "each"
+    return "", ""
+
+
+def derive_price_fields(price_per_each, price_per_lb, price_per_oz, line_total, qty, pack_qty):
+    line_total_decimal = to_decimal(line_total)
+    qty_decimal = to_decimal(qty)
+    pack_decimal = to_decimal(pack_qty)
+    price_per_count = ""
+    price_per_count_basis = ""
+    if line_total_decimal is not None and qty_decimal not in (None, Decimal("0")) and pack_decimal not in (
+        None,
+        Decimal("0"),
+    ):
+        price_per_count = format_decimal(line_total_decimal / (qty_decimal * pack_decimal))
+        price_per_count_basis = "line_total_over_pack_qty"
+
+    return {
+        "price_per_each": price_per_each,
+        "price_per_each_basis": "line_total_over_qty" if price_per_each else "",
+        "price_per_count": price_per_count,
+        "price_per_count_basis": price_per_count_basis,
+        "price_per_lb": price_per_lb,
+        "price_per_lb_basis": "parsed_or_picked_weight" if price_per_lb else "",
+        "price_per_oz": price_per_oz,
+        "price_per_oz_basis": "parsed_or_picked_weight" if price_per_oz else "",
+    }
+
+
+def normalization_identity(row):
+    if row.get("upc"):
+        return f"{row['retailer']}|upc={row['upc']}", "exact_upc"
+    if row.get("retailer_item_id"):
+        return f"{row['retailer']}|retailer_item_id={row['retailer_item_id']}", "exact_retailer_item_id"
+    if row.get("item_name_norm"):
+        return (
+            "|".join(
+                [
+                    row["retailer"],
+                    f"name={row['item_name_norm']}",
+                    f"size={row.get('size_value', '')}",
+                    f"unit={row.get('size_unit', '')}",
+                    f"pack={row.get('pack_qty', '')}",
+                ]
+            ),
+            "exact_name_size_pack",
+        )
+    return row["normalized_row_id"], "row_identity"
+
+
 def parse_item(order_id, order_date, raw_path, line_no, item):
     cleaned_name = clean_item_name(item.get("itemName", ""))
     size_value, size_unit, pack_qty = parse_size_and_pack(cleaned_name)
@@ -352,11 +422,42 @@ def parse_item(order_id, order_date, raw_path, line_no, item):
     if size_value and not size_unit:
         parse_notes.append("size_without_unit")
 
+    normalized_row_id = f"{RETAILER}:{order_id}:{line_no}"
+    normalized_quantity, normalized_quantity_unit = derive_normalized_quantity(
+        size_value,
+        size_unit,
+        pack_qty,
+        measure_type,
+    )
+    identity_key, normalization_basis = normalization_identity(
+        {
+            "retailer": RETAILER,
+            "normalized_row_id": normalized_row_id,
+            "upc": stringify(item.get("primUpcCd")),
+            "retailer_item_id": stringify(item.get("podId")),
+            "item_name_norm": normalized_name,
+            "size_value": size_value,
+            "size_unit": size_unit,
+            "pack_qty": pack_qty,
+        }
+    )
+    price_fields = derive_price_fields(
+        price_per_each,
+        price_per_lb,
+        price_per_oz,
+        stringify(item.get("groceryAmount")),
+        stringify(item.get("shipQy")),
+        pack_qty,
+    )
+
     return {
         "retailer": RETAILER,
         "order_id": str(order_id),
         "line_no": str(line_no),
-        "observed_item_key": f"{RETAILER}:{order_id}:{line_no}",
+        "normalized_row_id": normalized_row_id,
+        "normalized_item_id": f"gnorm:{identity_key}",
+        "normalization_basis": normalization_basis,
+        "observed_item_key": normalized_row_id,
         "order_date": normalize_whitespace(order_date),
         "retailer_item_id": stringify(item.get("podId")),
         "pod_id": stringify(item.get("podId")),
@@ -384,13 +485,14 @@ def parse_item(order_id, order_date, raw_path, line_no, item):
         "size_unit": size_unit,
         "pack_qty": pack_qty,
         "measure_type": measure_type,
+        "normalized_quantity": normalized_quantity,
+        "normalized_quantity_unit": normalized_quantity_unit,
         "is_store_brand": "true" if bool(prefix) else "false",
+        "is_item": "false" if is_fee else "true",
         "is_fee": "true" if is_fee else "false",
         "is_discount_line": "false",
         "is_coupon_line": "false",
-        "price_per_each": price_per_each,
-        "price_per_lb": price_per_lb,
-        "price_per_oz": price_per_oz,
+        **price_fields,
         "parse_version": PARSER_VERSION,
         "parse_notes": ";".join(parse_notes),
     }
@@ -443,6 +545,7 @@ def write_csv(path, rows):
     help="CSV path for enriched Giant item rows.",
 )
 def main(input_dir, output_csv):
+    click.echo("legacy entrypoint: prefer normalize_giant_web.py for data-model outputs")
     raw_dir = Path(input_dir)
     output_path = Path(output_csv)