from typing import Dict, List, Union class MainHeatAttributes: HEAT_SYSTEMS = ["boiler", "air source heat pump", "room heaters", "electric storage heaters", "warm air", "electric underfloor heating", "electric ceiling heating", "community scheme", "ground source heat pump", "no system present", "portable electric heaters", "water source heat pump"] FUEL_TYPES = ["electric", "mains gas", "wood logs", "LPG", "coal", "oil", "wood pellets", "anthracite", "dual fuel mineral and wood", "smokeless fuel", "lpg"] DISTRIBUTION_SYSTEMS = ["radiators", "fan coil units", "pipes in screed above insulation", "pipes in insulated timber floor", "pipes in concrete slab"] OTHERS = ["assumed", "electricaire", "assumed for most rooms"] def __init__(self, description: str): self.description: str = self._clean_description(description.lower()) # Remove special characters if not description or not any( rt in self.description for rt in self.HEAT_SYSTEMS + self.FUEL_TYPES + self.DISTRIBUTION_SYSTEMS + self.OTHERS ): raise ValueError('Invalid description') @staticmethod def _clean_description(description: str) -> str: """ Clean the description by replacing any special characters with a space. """ special_chars = [":", ";", "*", "@", "?", "!", "(", ")"] for char in special_chars: description = description.replace(char, " ") return description def process(self) -> Dict[str, Union[str, bool]]: result: Dict[str, Union[str, bool]] = {f'has_{ds.replace(" ", "_")}': False for ds in self.DISTRIBUTION_SYSTEMS} result.update({f'has_{hs.replace(" ", "_")}': False for hs in self.HEAT_SYSTEMS}) result.update({f'has_{ft.replace(" ", "_")}': False for ft in self.FUEL_TYPES}) result.update({f'has_{ot.replace(" ", "_")}': False for ot in self.OTHERS}) description = self.description.split(',') # Process each part separately for part in description: part = part.strip() # remove leading/trailing white spaces # Heating Systems self._process_part(result, part, self.HEAT_SYSTEMS, 'has_') # Fuel Types self._process_part(result, part, self.FUEL_TYPES, 'has_') # Distribution Systems self._process_part(result, part, self.DISTRIBUTION_SYSTEMS, 'has_') # Other attributes self._process_part(result, part, self.OTHERS, 'has_') # Check for "underfloor" separately in the entire description if "underfloor" in self.description: result['has_underfloor_heating'] = True return result @staticmethod def _process_part(result: Dict[str, Union[str, bool]], part: str, attr_list: List[str], prefix: str): """ Process a part of the description with a given list of attributes and update the result dictionary. """ part_words = part.split() for attr in attr_list: attr_words = attr.split() if set(attr_words).issubset(set(part_words)): result[f'{prefix}{attr.replace(" ", "_")}'] = True return result