first commit

utils/feed_validator.py

@@ -0,0 +1,131 @@
+# utils/feed_validator.py
+"""
+Валидатор YML фида
+"""
+
+import xml.etree.ElementTree as ET
+from pathlib import Path
+
+
+class FeedValidator:
+    """Валидатор YML фида для Prom.ua"""
+
+    def __init__(self):
+        self.errors = []
+        self.warnings = []
+
+    def validate_feed(self, feed_path):
+        """Валидирует YML фид"""
+        self.errors = []
+        self.warnings = []
+
+        try:
+            tree = ET.parse(feed_path)
+            root = tree.getroot()
+
+            # Проверяем структуру
+            self._validate_structure(root)
+
+            # Проверяем offers
+            offers = root.find('.//offers')
+            if offers is not None:
+                self._validate_offers(offers)
+
+            # Проверяем категории
+            categories = root.find('.//categories')
+            if categories is not None:
+                self._validate_categories(categories)
+
+            return len(self.errors) == 0
+
+        except ET.ParseError as e:
+            self.errors.append(f"XML parsing error: {e}")
+            return False
+        except Exception as e:
+            self.errors.append(f"Validation error: {e}")
+            return False
+
+    def _validate_structure(self, root):
+        """Проверяет основную структуру"""
+        if root.tag != 'yml_catalog':
+            self.errors.append("Root element must be 'yml_catalog'")
+
+        shop = root.find('shop')
+        if shop is None:
+            self.errors.append("Missing 'shop' element")
+            return
+
+        required_elements = ['name', 'company', 'currencies', 'categories', 'offers']
+        for element in required_elements:
+            if shop.find(element) is None:
+                self.errors.append(f"Missing required element: {element}")
+
+    def _validate_offers(self, offers):
+        """Проверяет offers"""
+        offer_count = 0
+
+        for offer in offers.findall('offer'):
+            offer_count += 1
+            offer_id = offer.get('id')
+
+            if not offer_id:
+                self.errors.append(f"Offer {offer_count} missing id attribute")
+
+            # Проверяем обязательные поля
+            required_fields = ['name', 'price', 'currencyId']
+            for field in required_fields:
+                if offer.find(field) is None:
+                    self.errors.append(f"Offer {offer_id} missing required field: {field}")
+
+            # Проверяем цену
+            price_elem = offer.find('price')
+            if price_elem is not None:
+                try:
+                    price = float(price_elem.text)
+                    if price <= 0:
+                        self.errors.append(f"Offer {offer_id} has invalid price: {price}")
+                except ValueError:
+                    self.errors.append(f"Offer {offer_id} has non-numeric price")
+
+            # Проверяем изображения
+            pictures = offer.findall('picture')
+            if not pictures:
+                self.warnings.append(f"Offer {offer_id} has no images")
+
+    def _validate_categories(self, categories):
+        """Проверяет категории"""
+        category_ids = set()
+
+        for category in categories.findall('category'):
+            cat_id = category.get('id')
+
+            if not cat_id:
+                self.errors.append("Category missing id attribute")
+                continue
+
+            if cat_id in category_ids:
+                self.errors.append(f"Duplicate category id: {cat_id}")
+
+            category_ids.add(cat_id)
+
+            if not category.text or not category.text.strip():
+                self.errors.append(f"Category {cat_id} has empty name")
+
+    def get_report(self):
+        """Возвращает отчёт валидации"""
+        report = []
+
+        if self.errors:
+            report.append("ERRORS:")
+            for error in self.errors:
+                report.append(f"  - {error}")
+
+        if self.warnings:
+            report.append("WARNINGS:")
+            for warning in self.warnings:
+                report.append(f"  - {warning}")
+
+        if not self.errors and not self.warnings:
+            report.append("Feed is valid!")
+
+        return '\n'.join(report)