Validate product attributes and decide match status

app/catalog.py

@@ -5,6 +5,10 @@ from dataclasses import dataclass
 from decimal import Decimal, InvalidOperation
 from pathlib import Path
 
+from app.aliases import ProductAliasIndex
+from app.attributes import BrandIndex, ParsedAttributes, parse_attributes
+from app.normalization import meaningful_tokens, normalize_text, tokenize
+
 
 @dataclass(frozen=True, slots=True)
 class RawCatalogItem:
@@ -14,6 +18,18 @@ class RawCatalogItem:
     price: Decimal
 
 
+@dataclass(frozen=True, slots=True)
+class CatalogItem:
+    sku: str
+    name: str
+    unit: str
+    price: Decimal
+    normalized_name: str
+    search_tokens: tuple[str, ...]
+    product_type: str
+    attributes: ParsedAttributes
+
+
 def load_raw_catalog(path: Path) -> tuple[RawCatalogItem, ...]:
     if not path.is_file():
         raise FileNotFoundError(f"Catalog file does not exist: {path}")
@@ -52,3 +68,36 @@ def load_raw_catalog(path: Path) -> tuple[RawCatalogItem, ...]:
     if not items:
         raise ValueError("Catalog is empty")
     return tuple(items)
+
+
+def build_catalog(
+    raw_items: tuple[RawCatalogItem, ...],
+    alias_index: ProductAliasIndex,
+    brand_index: BrandIndex,
+) -> tuple[CatalogItem, ...]:
+    result: list[CatalogItem] = []
+    for raw in raw_items:
+        normalized = normalize_text(raw.name)
+        product_type = alias_index.detect_catalog_type(raw.name)
+        attributes = parse_attributes(
+            raw.name,
+            product_type,
+            unit=raw.unit,
+            catalog_item=True,
+            brand_index=brand_index,
+        )
+        result.append(
+            CatalogItem(
+                sku=raw.sku,
+                name=raw.name,
+                unit=raw.unit,
+                price=raw.price,
+                normalized_name=normalized,
+                search_tokens=meaningful_tokens(
+                    tokenize(normalized, already_normalized=True)
+                ),
+                product_type=product_type,
+                attributes=attributes,
+            )
+        )
+    return tuple(result)