modular parser framework

2026-05-21 08:21:49 +00:00
parent 1223791074
commit cf5348a0c8
35 changed files with 452 additions and 166 deletions
@@ -0,0 +1,129 @@
+import logging
+import re
+from datetime import datetime, date
+from openmensa_parsers.xml_types.canteen_xml import CanteenMeta, CanteenXML
+from openmensa_parsers.xml_types.times_xml import CanteenOpenTimespec, TimesXML
+from openmensa_parsers.xml_types.meal_xml import MealXML
+
+
+EURO_PRICE_PATTERN = re.compile(r"(\d+(?:[,.]\d{1,2})?)\s*€")
+
+
+class WebspeiseplanParser:
+    """Parser for Webspeiseplan menu and outlet data."""
+
+    def __init__(self) -> None:
+        """Init WebspeiseplanParser object."""
+        logging.basicConfig()
+        self.logger = logging.getLogger(__name__)
+
+    def parse_canteen_meta_times(self, outlet: dict):
+        """Parse the outlet data from outlet.
+
+        Args:
+            outlet (dict): [description]
+        """
+        self.logger.debug("parse_canteen_meta_times(): %s", outlet["name"])
+        addr_info = outlet["addressInfo"]
+        meta = {
+            "name": outlet["name"],
+            "address": f'{addr_info["street"]}, {addr_info["postalCode"]} '
+            + f'{addr_info["city"]}',
+            "city": addr_info["city"],
+            "phone": outlet["contactInfo"][0]["phone"],
+            "email": outlet["contactInfo"][0]["email"],
+            "availability": outlet["isPublic"]
+        }
+
+        if outlet["positionInfo"]:
+            meta["location"] = (
+                outlet["positionInfo"]["longitude"],
+                outlet["positionInfo"]["latitude"],
+            )
+        canteen_meta = CanteenMeta(**meta)
+        weekday_dict = {
+            # this approach only lists the first (valid) opening time,
+            # since OpenMensa does not support multiple time specs
+            # (yet).
+            "monday": outlet['moZeit1'] or outlet['moZeit2'],
+            "tuesday": outlet['diZeit1'] or outlet['diZeit2'],
+            "wednesday": outlet['miZeit1'] or outlet['miZeit2'],
+            "thursday": outlet['doZeit1'] or outlet['doZeit2'],
+            "friday": outlet['frZeit1'] or outlet['frZeit2'],
+            "saturday": outlet['saZeit1'] or outlet['saZeit2'],
+            "sunday": outlet['soZeit1'] or outlet['soZeit2'],
+        }
+
+        canteen_times = TimesXML({
+            k: CanteenOpenTimespec(v) for k, v in weekday_dict.items()
+        })
+        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]]:
+        """Parse the menu and adds it to the builder."""
+        meals = []
+        for menu in menu_data:
+            for meal_data in menu["speiseplanGerichtData"]:
+                info = meal_data["speiseplanAdvancedGericht"]
+                additional_info = meal_data["zusatzinformationen"]
+                price = {
+                    "student": self._parse_price(
+                        additional_info["mitarbeiterpreisDecimal2"]
+                    ),
+                    "employee": self._parse_price(
+                        additional_info["price3Decimal2"]
+                    ),
+                    "other": self._parse_price(
+                        additional_info["gaestepreisDecimal2"]
+                    ),
+                }
+                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(
+                    {"day": day, "category": category, "meal": meal}
+                )
+        self.logger.debug("parse_meals(): %s meals parsed", len(meals))
+        return meals