mirror of
https://github.com/Hestia-Homes/Model.git
synced 2026-06-08 11:17:27 +00:00
358 lines
14 KiB
Python
358 lines
14 KiB
Python
import pytest
|
|
from utils.s3 import read_dataframe_from_s3_parquet
|
|
from etl.epc.Record import EPCRecord
|
|
from etl.epc.settings import DATA_ANOMALY_MATCHES
|
|
import random
|
|
|
|
|
|
class TestEpcRecord:
|
|
|
|
@pytest.fixture()
|
|
def cleaning_data(self):
|
|
cleaning_data = read_dataframe_from_s3_parquet(
|
|
bucket_name="retrofit-data-dev", file_key="sap_change_model/cleaning_dataset.parquet",
|
|
)
|
|
|
|
return cleaning_data
|
|
|
|
@pytest.fixture()
|
|
def epc_records_1(self):
|
|
epc_records_1 = {
|
|
'original_epc': {
|
|
'low-energy-fixed-light-count': '', 'address': '139 School Road, Hall Green',
|
|
'uprn-source': 'Energy Assessor', 'floor-height': '2.6', 'heating-cost-potential': '1138',
|
|
'unheated-corridor-length': '', 'hot-water-cost-potential': '175',
|
|
'construction-age-band': 'England and Wales: 1900-1929', 'potential-energy-rating': 'B',
|
|
'mainheat-energy-eff': 'Good', 'windows-env-eff': 'Average', 'lighting-energy-eff': 'Very Good',
|
|
'environment-impact-potential': '82', 'glazed-type': 'double glazing, unknown install date',
|
|
'heating-cost-current': '2711', 'address3': '',
|
|
'mainheatcont-description': 'Programmer, TRVs and bypass',
|
|
'sheating-energy-eff': 'N/A', 'property-type': 'House', 'local-authority-label': 'Birmingham',
|
|
'fixed-lighting-outlets-count': '11', 'energy-tariff': 'Single', 'mechanical-ventilation': 'natural',
|
|
'hot-water-cost-current': '310', 'county': '', 'postcode': 'B28 8JF', 'solar-water-heating-flag': 'N',
|
|
'constituency': 'E14000562', 'co2-emissions-potential': '2.0', 'number-heated-rooms': '4',
|
|
'floor-description': 'Suspended, no insulation (assumed)', 'energy-consumption-potential': '107',
|
|
'local-authority': 'E08000025', 'built-form': 'Semi-Detached', 'number-open-fireplaces': '0',
|
|
'windows-description': 'Fully double glazed', 'glazed-area': 'Normal', 'inspection-date': '2023-07-05',
|
|
'mains-gas-flag': 'Y', 'co2-emiss-curr-per-floor-area': '65', 'address1': '139 School Road',
|
|
'heat-loss-corridor': '', 'flat-storey-count': '', 'constituency-label': 'Birmingham, Hall Green',
|
|
'roof-energy-eff': 'Average', 'total-floor-area': '103.0', 'building-reference-number': '10004697322',
|
|
'environment-impact-current': '43', 'co2-emissions-current': '6.7',
|
|
'roof-description': 'Pitched, 100 mm loft insulation', 'floor-energy-eff': 'N/A',
|
|
'number-habitable-rooms': '4', 'address2': 'Hall Green', 'hot-water-env-eff': 'Good',
|
|
'posttown': 'BIRMINGHAM', 'mainheatc-energy-eff': 'Average', 'main-fuel': 'mains gas (not community)',
|
|
'lighting-env-eff': 'Very Good', 'windows-energy-eff': 'Average', 'floor-env-eff': 'N/A',
|
|
'sheating-env-eff': 'N/A', 'lighting-description': 'Low energy lighting in 82% of fixed outlets',
|
|
'roof-env-eff': 'Average', 'walls-energy-eff': 'Very Poor', 'photo-supply': '0.0',
|
|
'lighting-cost-potential': '182', 'mainheat-env-eff': 'Good', 'multi-glaze-proportion': '100',
|
|
'main-heating-controls': '', 'lodgement-datetime': '2023-07-13 08:23:07', 'flat-top-storey': '',
|
|
'current-energy-rating': 'E', 'secondheat-description': 'None', 'walls-env-eff': 'Very Poor',
|
|
'transaction-type': 'rental', 'uprn': '100070505235', 'current-energy-efficiency': '51',
|
|
'energy-consumption-current': '366', 'mainheat-description': 'Boiler and radiators, mains gas',
|
|
'lighting-cost-current': '182', 'lodgement-date': '2023-07-13', 'extension-count': '0',
|
|
'mainheatc-env-eff': 'Average',
|
|
'lmk-key': 'c1d137711da433fb3cced74b1a6848da8bbc1159d076455d26d7b4668982601e',
|
|
'wind-turbine-count': '0',
|
|
'tenure': 'Rented (social)', 'floor-level': '', 'potential-energy-efficiency': '84',
|
|
'hot-water-energy-eff': 'Good', 'low-energy-lighting': '82',
|
|
'walls-description': 'Solid brick, as built, no insulation (assumed)',
|
|
'hotwater-description': 'From main system'}, 'full_sap_epc': {}, 'old_data': []
|
|
}
|
|
return epc_records_1
|
|
|
|
def test_clean_mechanical_ventilation(self, cleaning_data, epc_records_1):
|
|
# We have an epc with Natural ventilation - the resulting epc should also have natural ventulation
|
|
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
record.prepared_epc = {
|
|
"mechanical-ventilation": "natural"
|
|
}
|
|
record._clean_ventilation()
|
|
|
|
assert record.prepared_epc["mechanical-ventilation"] == "natural"
|
|
|
|
record2 = EPCRecord(cleaning_data=cleaning_data)
|
|
record2.prepared_epc = {
|
|
"mechanical-ventilation": ""
|
|
}
|
|
|
|
record2._clean_ventilation()
|
|
|
|
assert record2.prepared_epc["mechanical-ventilation"] is None
|
|
|
|
record3 = EPCRecord(cleaning_data=cleaning_data)
|
|
record3.prepared_epc = {
|
|
"mechanical-ventilation": None
|
|
}
|
|
|
|
record3._clean_ventilation()
|
|
|
|
assert record3.prepared_epc["mechanical-ventilation"] is None
|
|
|
|
record4 = EPCRecord(cleaning_data=cleaning_data)
|
|
record4.prepared_epc = {
|
|
"mechanical-ventilation": "INVALID"
|
|
}
|
|
|
|
record4._clean_ventilation()
|
|
|
|
assert record4.prepared_epc["mechanical-ventilation"] is None
|
|
|
|
def test_clean_energy_valid_values(self, cleaning_data, epc_records_1):
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
record.prepared_epc = {
|
|
"energy-consumption-current": "200",
|
|
"co2-emissions-current": "5.5"
|
|
}
|
|
record._clean_energy()
|
|
|
|
assert record.prepared_epc["energy-consumption-current"] == 200.0
|
|
assert record.prepared_epc["co2-emissions-current"] == 5.5
|
|
|
|
def test_clean_energy_empty_values(self, cleaning_data):
|
|
# We cannot have invalid values so this should raise an exception
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
record.prepared_epc = {
|
|
"energy-consumption-current": "",
|
|
"co2-emissions-current": ""
|
|
}
|
|
|
|
with pytest.raises(ValueError):
|
|
record._clean_energy()
|
|
|
|
def test_clean_built_form_valid_remap(self, cleaning_data, epc_records_1):
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
# Assuming "Semi" should be remapped to "Semi-Detached"
|
|
record.prepared_epc = {
|
|
"built-form": "Semi-Detached",
|
|
"property-type": "Flat" # Assuming this affects the remapping
|
|
}
|
|
record._clean_built_form()
|
|
|
|
assert record.prepared_epc["built-form"] == "Semi-Detached"
|
|
|
|
def test_clean_built_form_anomaly(self, cleaning_data, epc_records_1):
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
|
|
record.prepared_epc = {
|
|
"built-form": "",
|
|
"property-type": "Flat"
|
|
}
|
|
record._clean_built_form()
|
|
|
|
assert record.prepared_epc["built-form"] == "End-Terrace"
|
|
|
|
def test_clean_floor_area_valid(self, cleaning_data):
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
record.prepared_epc = {
|
|
"total-floor-area": "120.5"
|
|
}
|
|
record._clean_floor_area()
|
|
|
|
assert record.prepared_epc["total-floor-area"] == 120.5
|
|
|
|
def test_clean_floor_area_empty(self, cleaning_data):
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
record.prepared_epc = {
|
|
"total-floor-area": ""
|
|
}
|
|
# We have no known case of missing floor area
|
|
with pytest.raises(ValueError):
|
|
record._clean_floor_area()
|
|
|
|
def test_clean_heat_loss_corridor_valid(self, cleaning_data):
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
record.prepared_epc = {
|
|
"heat-loss-corridor": "unheated corridor",
|
|
"unheated-corridor-length": ""
|
|
}
|
|
record._clean_heat_loss_corridor()
|
|
|
|
assert record.prepared_epc["heat-loss-corridor"] == "unheated corridor"
|
|
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
record.prepared_epc = {
|
|
"heat-loss-corridor": "unheated corridor",
|
|
"unheated-corridor-length": None
|
|
}
|
|
record._clean_heat_loss_corridor()
|
|
|
|
assert record.prepared_epc["heat-loss-corridor"] == "unheated corridor"
|
|
assert record.prepared_epc["unheated-corridor-length"] is None
|
|
|
|
def test_clean_heat_loss_corridor_anomaly(self, cleaning_data):
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
# Assuming "InvalidCorridor" is an anomaly
|
|
record.prepared_epc = {
|
|
"heat-loss-corridor": "InvalidCorridor",
|
|
"unheated-corridor-length": ""
|
|
}
|
|
record._clean_heat_loss_corridor()
|
|
|
|
assert record.prepared_epc["heat-loss-corridor"] == "no corridor"
|
|
|
|
def test_clean_mains_gas_valid(self, cleaning_data):
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
record.prepared_epc = {
|
|
"mains-gas-flag": "Y"
|
|
}
|
|
record._clean_mains_gas()
|
|
|
|
assert record.prepared_epc["mains-gas-flag"] is True
|
|
|
|
def test_clean_mains_gas_anomaly(self, cleaning_data):
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
record.prepared_epc = {
|
|
"mains-gas-flag": "InvalidValue"
|
|
}
|
|
# It should always be Y or N or an anomally value
|
|
with pytest.raises(KeyError):
|
|
record._clean_mains_gas()
|
|
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
record.prepared_epc = {
|
|
"mains-gas-flag": random.choice(list(DATA_ANOMALY_MATCHES))
|
|
}
|
|
record._clean_mains_gas()
|
|
|
|
assert record.prepared_epc["mains-gas-flag"] is None
|
|
|
|
def test_clean_solar_hot_water_valid(self, cleaning_data):
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
record.prepared_epc = {
|
|
"solar-water-heating-flag": "Y"
|
|
}
|
|
record._clean_solar_hot_water()
|
|
|
|
assert record.prepared_epc["solar-water-heating-flag"] == "Y"
|
|
assert record.solar_water_heating_flag_bool is True
|
|
|
|
def test_clean_solar_hot_water_empty(self, cleaning_data):
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
record.prepared_epc = {
|
|
"solar-water-heating-flag": ""
|
|
}
|
|
record._clean_solar_hot_water()
|
|
|
|
assert record.prepared_epc["solar-water-heating-flag"] == "N"
|
|
assert record.solar_water_heating_flag_bool is False
|
|
|
|
def test_clean_number_lighting_outlets_valid(self, cleaning_data, epc_records_1):
|
|
record = EPCRecord(cleaning_data=cleaning_data, epc_records=epc_records_1)
|
|
record.prepared_epc = {
|
|
"fixed-lighting-outlets-count": "5"
|
|
}
|
|
record._clean_number_lighting_outlets()
|
|
|
|
assert record.prepared_epc["fixed-lighting-outlets-count"] == 5.0
|
|
|
|
def test_clean_number_lighting_outlets_empty(self, cleaning_data, epc_records_1):
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
record.run_mode = "newdata"
|
|
record.prepared_epc = {
|
|
"fixed-lighting-outlets-count": "",
|
|
"property-type": "Flat",
|
|
"built-form": "Semi-Detached",
|
|
"construction-age-band": "England and Wales: 1900-1929",
|
|
"local-authority": "E08000025",
|
|
"number-habitable-rooms": "4",
|
|
"number-heated-rooms": "4",
|
|
}
|
|
record.old_data = []
|
|
record.full_sap_epc = []
|
|
record._clean_number_lighting_outlets()
|
|
|
|
assert record.prepared_epc["fixed-lighting-outlets-count"] == 8.0
|
|
|
|
def test_clean_count_variables(self, cleaning_data):
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
|
|
record.prepared_epc = {
|
|
"number-open-fireplaces": "1",
|
|
"extension-count": None,
|
|
"flat-storey-count": "",
|
|
"number-habitable-rooms": "INVALID!",
|
|
}
|
|
|
|
record._clean_count_variables()
|
|
|
|
assert record.prepared_epc["number-open-fireplaces"] == 1.0
|
|
assert record.prepared_epc["extension-count"] == 0
|
|
assert record.prepared_epc["flat-storey-count"] is None
|
|
assert record.prepared_epc["number-habitable-rooms"] is None
|
|
|
|
def test_clean_floor_level(self, cleaning_data):
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
|
|
record.prepared_epc = {
|
|
"floor-level": "1",
|
|
}
|
|
|
|
record._clean_floor_level()
|
|
|
|
assert record.prepared_epc["floor-level"] == 1.0
|
|
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
|
|
record.prepared_epc = {
|
|
"floor-level": "",
|
|
}
|
|
|
|
record._clean_floor_level()
|
|
|
|
assert record.prepared_epc["floor-level"] is None
|
|
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
|
|
record.prepared_epc = {
|
|
"floor-level": None,
|
|
}
|
|
|
|
record._clean_floor_level()
|
|
|
|
assert record.prepared_epc["floor-level"] is None
|
|
|
|
def test_clean_solar_hot_water(self, cleaning_data):
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
|
|
record.prepared_epc = {
|
|
"solar-water-heating-flag": "Y",
|
|
}
|
|
|
|
record._clean_solar_hot_water()
|
|
|
|
assert record.prepared_epc["solar-water-heating-flag"] == "Y"
|
|
assert record.solar_water_heating_flag_bool is True
|
|
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
|
|
record.prepared_epc = {
|
|
"solar-water-heating-flag": "N",
|
|
}
|
|
|
|
record._clean_solar_hot_water()
|
|
|
|
assert record.prepared_epc["solar-water-heating-flag"] == "N"
|
|
assert record.solar_water_heating_flag_bool is False
|
|
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
|
|
record.prepared_epc = {
|
|
"solar-water-heating-flag": "",
|
|
}
|
|
|
|
record._clean_solar_hot_water()
|
|
|
|
assert record.prepared_epc["solar-water-heating-flag"] == "N"
|
|
assert record.solar_water_heating_flag_bool is False
|
|
|
|
record = EPCRecord(cleaning_data=cleaning_data)
|
|
|
|
record.prepared_epc = {
|
|
"solar-water-heating-flag": None,
|
|
}
|
|
|
|
record._clean_solar_hot_water()
|
|
|
|
assert record.prepared_epc["solar-water-heating-flag"] == "N"
|
|
assert record.solar_water_heating_flag_bool is False
|