from typing import Dict, Union
from model_data.BaseUtility import Definitions
from model_data.epc_attributes.attribute_utils import clean_description, remove_punctuation, find_keyword


class MainFuelAttributes(Definitions):
    FUEL_KEYWORDS = [
        'heat network',
        'mains gas',
        'electricity',
        'oil',
        'biomass',
        'biodiesel',
        # Note: there is als a category called 'bottled LPG', but only 2/50k entries had this
        'lpg',
        'waste combustion',
        'biogas',
        'wood logs',
        'dual fuel mineral wood',
        'gas',
        'anthracite',
        'smokeless coal',
        'house coal',
        'wood chips',
        # We don't treat wood chips and wood pelles as the same.
        # Wood pellets have a higher energy density than wood chips. This is due to their manufacturing process,
        # which compresses the wood and removes most of the moisture, making them more efficient as a fuel
        'wood pellets',
        'b30k',
        'dual fuel appliance mineral and wood',
        'coal',
        'b30d',
        'bioethanol',
    ]

    COMPLEX_FUEL_KEYWORDS = [
        'heat from boilers using biodiesel from any biomass source'
    ]

    TARIFF_KEYWORDS = [
        'unspecified tariff'
        # We may come across more later but this is all observed for now
    ]

    UNKNOWN_FUEL = "unknown"

    NO_INDIVIDUAL_HEATING_OR_COMMUNITY_NETWORK = [
        'to be used only when there is no heatinghotwater system or data is from a community network',
        'to be used only when there is no heatinghotwater system'
    ]

    def __init__(self, description: str):
        self.description: str = remove_punctuation(clean_description(description.lower()))

        self.is_community = 'community' in self.description and 'not community' not in self.description
        self.is_unknown = False
        self.nodata = not description or description in self.DATA_ANOMALY_MATCHES

        if not self.nodata and not any(
            self._keyword_in_description(keywords)
            for keywords in [
                self.FUEL_KEYWORDS,
                self.NO_INDIVIDUAL_HEATING_OR_COMMUNITY_NETWORK,
                self.TARIFF_KEYWORDS,
                self.COMPLEX_FUEL_KEYWORDS
            ]
        ):
            raise ValueError('Invalid description')

    def _keyword_in_description(self, keywords):
        return any(keyword in self.description for keyword in keywords)

    def process(self) -> Dict[str, Union[str, bool]]:

        if self.nodata:
            result = {
                "fuel_type": self.UNKNOWN_FUEL,
                "tariff_type": None,
                "is_community": False,
                "no_individual_heating_or_community_network": False,
                "complex_fuel_type": None
            }
            return result

        result: Dict[str, Union[str, bool]] = {
            "fuel_type": find_keyword(self.description, self.FUEL_KEYWORDS),
            "tariff_type": find_keyword(self.description, self.TARIFF_KEYWORDS),
            "is_community": self.is_community,
            "no_individual_heating_or_community_network": find_keyword(
                self.description, self.NO_INDIVIDUAL_HEATING_OR_COMMUNITY_NETWORK
            ),
            "complex_fuel_type": find_keyword(self.description, self.COMPLEX_FUEL_KEYWORDS),

        }

        # to make this field palettable, if no_individual_heating_or_community_network is populated, we'll
        # just set it to true
        result["no_individual_heating_or_community_network"] = bool(
            result["no_individual_heating_or_community_network"]
        )

        if not result["fuel_type"]:
            result["fuel_type"] = self.UNKNOWN_FUEL
            # We'll do checks on unknown fuel data_types to ensure we don't miss anything
            self.is_unknown = True

        return result