Model/etl/epc_clean/epc_attributes/FloorAttributes.py

import re
from typing import Dict, Union
from BaseUtility import Definitions
from etl.epc_clean.epc_attributes.attribute_utils import (
    extract_thermal_transmittance,
    extract_component_types,
    handle_mixed_translation
)


class FloorAttributes(Definitions):
    DWELLING_BELOW = ["another dwelling below", "other premises below"]
    FLOOR_TYPES = [
        "assumed",
        "to unheated space",
        "to external air",
        "suspended",
        "solid",
    ]

    # For the short term, while we are still exploring the data, we maintain a list of error cases which
    # we want to ignore and consider as no data.

    OBSERVED_ERRORS = ["Conservatory", "insulated", "Basement"]

    WELSH_TEXT = {
        "(anheddiad arall islaw)": "(another dwelling below)",
        "solet, dim inswleiddio (rhagdybiaeth)": "solid, no insulation (assumed)",
        "solet, dim inswleiddio": "solid, no insulation)",
        "crog, dim inswleiddio (rhagdybiaeth)": "suspended, no insulation (assumed)",
        "crog, dim inswleiddio": "suspended, no insulation",
        "(eiddo arall islaw)": "(other premises below)",
        "solet, inswleiddio cyfyngedig (rhagdybiaeth)": "solid, limited insulation (assumed)",
        "solet, inswleiddio cyfyngedig": "solid, limited insulation",
        "crog, wedigçöi inswleiddio (rhagdybiaeth)": "suspended, insulated (assumed)",
        "crog, wedigçöi inswleiddio": "suspended, insulated",
        "igçör awyr y tu allan, dim inswleiddio (rhagdybiaeth)": "to external air, no insulation (assumed)",
        "igçör awyr y tu allan, dim inswleiddio": "to external air, no insulation",
        "i ofod heb ei wresogi, wedigçöi inswleiddio (rhagdybiaeth)": "to unheated space, insulated (assumed)",
        "i ofod heb ei wresogi, wedigçöi inswleiddio": "to unheated space, insulated",
        "solet, wedigçöi inswleiddio (rhagdybiaeth)": "solid, insulated (assumed)",
        "solet, wedigçöi inswleiddio": "solid, insulated",
        "solet, wedi???i inswleiddio (rhagdybiaeth)": "solid, insulated (assumed)",
        "i ofod heb ei wresogi, dim inswleiddio (rhagdybiaeth)": "to unheated space, no insulation (assumed)",
        "i ofod heb ei wresogi, heb ei inswleiddio (rhagdybiaeth)": "to unheated space, no insulation (assumed)",
        "i ofod heb ei wresogi, dim inswleiddio": "to unheated space, no insulation",
        "igçör awyr y tu allan, wedigçöi inswleiddio (rhagdybiaeth)": "to external air, insulated (assumed)",
        "crog, inswleiddio cyfyngedig (rhagdybiaeth)": "suspended, limited insulation (assumed)",
    }

    REMAP = {
        # Have only seen this once - though perhaps need to investigate older EPCs in the production of EPC clean.
        # When looking at a newer EPC, which had been re-assessed as another dwelling below
        "above unheated space or full exposed": "(another dwelling below)",
    }

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

        self.nodata = (
            (not description)
            or (description in self.DATA_ANOMALY_MATCHES)
            or (description in self.OBSERVED_ERRORS)
            or (self.description == "sap05:floor")
            or not self.description
        )

        # Try and perform a translation, incase it's in welsh
        self.translate_welsh_text()

        # Remap known issues
        if self.description in self.REMAP:
            self.description = self.REMAP[self.description]

        # We handle seemind occurances of mixed translations
        self.description = handle_mixed_translation(self.description)

        if not self.nodata and not any(
            rt in self.description
            for rt in self.FLOOR_TYPES
                      + self.DWELLING_BELOW
                      + ["average thermal transmittance"]
        ):
            raise ValueError("Invalid description")

    def translate_welsh_text(self):

        uvalue_match = re.search(
            r"trawsyriannedd thermol cyfartalog (\d+(\.\d+)?)\s*w/m-¦k",
            self.description,
        )

        uvalue_match2 = re.search(
            r"trawsyriannedd thermol cyfartalog (\d+(\.\d+)?)\s*w/m.+k",
            self.description,
        )

        # Step 2: Generalized translation with placeholder
        if uvalue_match is not None or uvalue_match2 is not None:
            if uvalue_match is not None:
                uvalue = uvalue_match.group(1)
            else:
                uvalue = uvalue_match2.group(1)
            self.description = f"average thermal transmittance {uvalue} w/m-¦k"

        else:
            translation = self.WELSH_TEXT.get(self.description)
            if translation:
                self.nodata = False
                self.description = translation

    def process(self) -> Dict[str, Union[str, bool, int, None]]:

        if self.nodata:
            return {
                'thermal_transmittance': None, 'thermal_transmittance_unit': None, 'is_assumed': False,
                'is_to_unheated_space': False, 'is_to_external_air': False, 'is_suspended': False, 'is_solid': False,
                'another_property_below': False, 'insulation_thickness': 'none', 'no_data': True
            }

        result: Dict[str, Union[float, str, bool, None]] = {}
        description = self.description

        # thermal transmittance
        result, description = extract_thermal_transmittance(result, description)

        # floor type
        result, description = extract_component_types(
            result, description, list_of_components=self.FLOOR_TYPES
        )

        # check if there is another dwelling below
        result["another_property_below"] = (
            "(another dwelling below)" in description
            or "(other premises below)" in description
            or "another dwelling below" in description
            or "other premises below" in description
        )

        thickness_map = {
            "external insulation": "average",
            "internal insulation": "average",
            "limited insulation": "below average",
            "partial insulation": "below average",
            "no insulation": "none",
            "additional insulation": "above average",
            "insulated": "average",
        }
        for key, value in thickness_map.items():
            if key in description:
                result["insulation_thickness"] = value
                break
        else:
            result["insulation_thickness"] = None

        if result["another_property_below"]:
            result["thermal_transmittance"] = 0
            result["thermal_transmittance_unit"] = "w/m-¦k"

        return result