handling more cases in EpcClean, 14% of data covered

This commit is contained in:
Khalim Conn-Kowlessar 2023-09-08 12:43:07 +03:00
parent 9bca33cf08
commit ddb2ce88e7

View file

@ -8,7 +8,7 @@ class MainHeatAttributes(Definitions):
"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", "electric heat pumps",
"water source heat pump", "electric heat pump",
# "Micro-cogeneration", also known as micro combined heat and power (micro-CHP), is a technology that
# generates heat and electricity simultaneously from the same energy source in residential or commercial
# buildings. The main output of micro-CHP systems is heat, with electricity generation as a secondary output.
@ -25,11 +25,13 @@ class MainHeatAttributes(Definitions):
WELSH_TEXT = {
"bwyler a rheiddiaduron, nwy prif gyflenwad": "boiler and radiators, mains gas",
"st+¦r wresogyddion trydan": "electric storage heaters",
"bwyler a rheiddiaduron, olew": "boiler and radiators, oil"
"bwyler a rheiddiaduron, olew": "boiler and radiators, oil",
"heat pumptrydan": "electric heat pump",
}
REMAP = {
"electric ceiling": "electric ceiling heating"
"electric ceiling": "electric ceiling heating",
"electric heat pumps": "electric heat pump",
}
def __init__(self, description: str):
@ -45,9 +47,16 @@ class MainHeatAttributes(Definitions):
self.nodata = False
self.description = translation
remap = self.REMAP.get(self.description)
if remap:
self.description = remap
remapped = []
for term in self.description.split(", "):
remap = self.REMAP.get(term)
if remap:
remapped.append(remap)
else:
remapped.append(term)
remapped = ", ".join(remapped)
self.description = remapped
if not description or not any(
rt in self.description for rt in