fixed salattheke prices

stw_potsdam/swp_webspeiseplan_parser.py

@@ -1,10 +1,14 @@
 import logging
+import re
 from datetime import datetime, date
 from stw_potsdam.xml_types.canteen_xml import CanteenMeta, CanteenXML
 from stw_potsdam.xml_types.times_xml import CanteenOpenTimespec, TimesXML
 from stw_potsdam.xml_types.meal_xml import MealXML
 
 
+EURO_PRICE_PATTERN = re.compile(r"(\d+(?:[,.]\d{1,2})?)\s*€")
+
+
 class SWPWebspeiseplanParser:
     """Class method to parse SWP_Webspeiseplan."""
 
@@ -56,6 +60,42 @@ class SWPWebspeiseplanParser:
         canteen = CanteenXML(canteen_meta, canteen_times)
         return canteen
 
+    def _parse_price(self, value):
+        if value in (None, "", {}):
+            return None
+        return float(str(value).replace(",", "."))
+
+    def _parse_embedded_prices(
+        self, name: str, price: dict[str, float | None]
+    ) -> tuple[str, dict[str, float | None]]:
+        if any(price.values()):
+            return name, price
+
+        matches = EURO_PRICE_PATTERN.findall(name)
+        if len(matches) < 2:
+            return name, price
+
+        parsed = [self._parse_price(match) for match in matches]
+        if len(parsed) >= 3:
+            price = {
+                "student": parsed[0],
+                "employee": parsed[1],
+                "other": parsed[2],
+            }
+        elif "Stud" in name and ("Gäste" in name or "Gaeste" in name):
+            price = {
+                "student": parsed[0],
+                "employee": price["employee"],
+                "other": parsed[1],
+            }
+        else:
+            return name, price
+
+        name = EURO_PRICE_PATTERN.sub("", name)
+        name = re.sub(r"\s*/\s*", " ", name)
+        name = re.sub(r"\s+", " ", name).strip()
+        return name, price
+
     def parse_meals(
         self, menu_data, meal_categories
     ) -> list[tuple[date, str, MealXML]]:
@@ -66,11 +106,20 @@ class SWPWebspeiseplanParser:
                 info = meal_data["speiseplanAdvancedGericht"]
                 additional_info = meal_data["zusatzinformationen"]
                 price = {
-                    "student": additional_info["mitarbeiterpreisDecimal2"],
-                    "employee": additional_info["price3Decimal2"],
-                    "other": additional_info["gaestepreisDecimal2"],
+                    "student": self._parse_price(
+                        additional_info["mitarbeiterpreisDecimal2"]
+                    ),
+                    "employee": self._parse_price(
+                        additional_info["price3Decimal2"]
+                    ),
+                    "other": self._parse_price(
+                        additional_info["gaestepreisDecimal2"]
+                    ),
                 }
-                meal = MealXML(name=info["gerichtname"], price=price)
+                name, price = self._parse_embedded_prices(
+                    info["gerichtname"], price
+                )
+                meal = MealXML(name=name, price=price)
                 day = datetime.fromisoformat(info["datum"]).date()
                 category = meal_categories[info["gerichtkategorieID"]]["name"]
                 meals.append(