debugging cleaning class for examples that hadn't been covered previously

etl/epc_clean/app.py

@@ -44,7 +44,8 @@ def app():
         # Rename the columns to the same format as the api returns
         data.columns = [c.replace("_", "-").lower() for c in data.columns]
         # Take just date before the date threshold
-        data = data[data["lodgement-date"] >= EARLIEST_EPC_DATE]
+        # For this cleaning dataset, let's try and use all EPCs
+        # data = data[data["lodgement-date"] >= EARLIEST_EPC_DATE]
 
         # Convert to list of dictioaries as returned by the api
         data = data.to_dict("records")

etl/epc_clean/epc_attributes/MainheatAttributes.py

@@ -66,6 +66,7 @@ class MainHeatAttributes(Definitions):
         "electric heat pumps": "electric heat pump",
         "solar-assisted heat pump": "solar assisted heat pump",
         "portable electric heating": "portable electric heaters",
+        "portable electric heating assumed for most rooms": "portable electric heaters assumed for most rooms",
     }
 
     edge_case_result = {}
@@ -138,6 +139,11 @@ class MainHeatAttributes(Definitions):
             self.is_edge_case = True
             return
 
+        if self.description == ', electric':
+            self.edge_case_result['has_electric'] = True
+            self.is_edge_case = True
+            return
+
     def process(self) -> Dict[str, Union[str, bool]]:
 
         result: Dict[str, Union[str, bool]] = {f'has_{ds.replace(" ", "_")}': False for ds in self.DISTRIBUTION_SYSTEMS}

etl/epc_clean/epc_attributes/MainheatControlAttributes.py

@@ -75,6 +75,7 @@ class MainheatControlAttributes(Definitions):
     TO_REMAP = {
         "celect control": 'celect-type control',
         "celect controls": 'celect-type control',
+        "trv's, program & flow switch": 'trvs, programmer & flow switch',
     }
 
     WELSH_TEXT = {