junte/Model
1
0
Fork 0
mirror of https://github.com/Hestia-Homes/Model.git synced 2026-06-30 13:10:47 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

updated class definition

Browse source
This commit is contained in:
Khalim Conn-Kowlessar 2023-06-14 18:30:00 +01:00
parent 6e26b941b7
commit d6b4403cfe
2 changed files with 23 additions and 36 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
epc_data/app.py
View file

@ -40,13 +40,13 @@ def handler():
cleaner.clean() cleaner.clean()
# For testing: # For testing:
from epc_data.attributes.MainheatControlAttributes import MainheatControlAttributes from epc_data.attributes.MainFuelAttributes import MainFuelAttributes
from collections import Counter from collections import Counter
count = Counter([x["main-fuel"] for x in data]) count = Counter([x["main-fuel"] for x in data])
descriptions = {x["main-fuel"] for x in data} descriptions = {x["main-fuel"] for x in data}
out = [] out = []
for description in descriptions: for description in descriptions:
res = MainheatControlAttributes(description).process() res = MainFuelAttributes(description).process()
out.append( out.append(
{ {
"original_description": description, "original_description": description,

55
epc_data/attributes/MainFuelAttributes.py
View file

@ -1,37 +1,25 @@
from typing import Dict, Union from typing import Dict, Union
from epc_data.attributes.attribute_utils import clean_description from epc_data.attributes.attribute_utils import clean_description, remove_punctuation
class MainFuelAttributes: class MainFuelAttributes:
COMMUNITY_FUEL_KEYWORDS = [ FUEL_KEYWORDS = [
'from heat network data (community)', 'heat network',
'mains gas (community)', 'mains gas',
'electricity (community)', 'electricity',
'oil (community)', 'oil',
'biomass (community)', 'biomass',
'heat from boilers using biodiesel from any biomass source (community)', 'biodiesel',
'LPG (community)', # Note: there is als a category called 'bottled LPG', but only 2/50k entries had this
'waste combustion (community)' 'LPG',
] 'waste combustion',
'biogas',
NOT_COMMUNITY_FUEL_KEYWORDS = [
'mains gas (not community)',
'electricity (not community)',
'oil (not community)',
'house coal (not community)',
'LPG (not community)',
'biogas (not community)'
]
SPECIFIC_FUEL_KEYWORDS = [
'waste combustion - this is for backwards compatibility only and should not be used',
'bottled LPG',
'wood logs', 'wood logs',
'dual fuel - mineral + wood', 'dual fuel - mineral + wood',
'Gas: mains gas', 'gas',
'Electricity: electricity, unspecified tariff',
'anthracite', 'anthracite',
'smokeless coal' 'smokeless coal',
'house coal'
] ]
NO_FUEL_KEYWORDS = [ NO_FUEL_KEYWORDS = [
@ -39,16 +27,16 @@ class MainFuelAttributes:
] ]
def __init__(self, description: str): def __init__(self, description: str):
self.description: str = clean_description(description.lower()) self.description: str = remove_punctuation(clean_description(description.lower()))
self.nodata = not description or any(keyword in self.description for keyword in self.NO_FUEL_KEYWORDS) self.nodata = not description or any(keyword in self.description for keyword in self.NO_FUEL_KEYWORDS)
self.is_community = False if 'not community' in self.description else 'community' in self.description
if not self.nodata and not any( if not self.nodata and not any(
self._keyword_in_description(keywords) self._keyword_in_description(keywords)
for keywords in [ for keywords in [
self.COMMUNITY_FUEL_KEYWORDS, self.FUEL_KEYWORDS
self.NOT_COMMUNITY_FUEL_KEYWORDS,
self.SPECIFIC_FUEL_KEYWORDS
] ]
): ):
raise ValueError('Invalid description') raise ValueError('Invalid description')
@ -58,10 +46,9 @@ class MainFuelAttributes:
def process(self) -> Dict[str, Union[str, bool]]: def process(self) -> Dict[str, Union[str, bool]]:
result: Dict[str, Union[str, bool]] = { result: Dict[str, Union[str, bool]] = {
"community_fuel": self._find_keyword(self.COMMUNITY_FUEL_KEYWORDS), "fuel_type": self._find_keyword(self.FUEL_KEYWORDS),
"not_community_fuel": self._find_keyword(self.NOT_COMMUNITY_FUEL_KEYWORDS),
"specific_fuel": self._find_keyword(self.SPECIFIC_FUEL_KEYWORDS),
"no_fuel": self._find_keyword(self.NO_FUEL_KEYWORDS), "no_fuel": self._find_keyword(self.NO_FUEL_KEYWORDS),
"is_community": self.is_community
} }
return result return result