Merge pull request #356 from Hestia-Homes/gla-proposal

Gla proposal

backend/Property.py

@@ -1204,7 +1204,7 @@ class Property:
             return False
 
         suitable_house = self.data["property-type"] == "House" and self.data["built-form"] in [
-            "Detached", "Semi-Detached",
+            "Detached", "Semi-Detached", "End-Terrace",
         ]
 
         suitable_bungalow = self.data["property-type"] == "Bungalow" and self.data["built-form"] in [

backend/app/plan/router.py

@@ -543,7 +543,11 @@ async def trigger_plan(body: PlanTriggerRequest):
         representative_recommendations = {}
         for p in tqdm(input_properties):
             recommender = Recommendations(
-                property_instance=p, materials=materials, exclusions=body.exclusions, inclusions=body.inclusions
+                property_instance=p,
+                materials=materials,
+                exclusions=body.exclusions,
+                inclusions=body.inclusions,
+                default_u_values=body.default_u_values
             )
             property_recommendations, property_representative_recommendations = recommender.recommend()
 

backend/app/plan/schemas.py

@@ -89,6 +89,9 @@ class PlanTriggerRequest(BaseModel):
     # if False, allows optimisation to be switched off
     optimise: Optional[bool] = True
 
+    # If True, uses default u-values for models
+    default_u_values: Optional[bool] = True
+
     _allowed_goals = {"Increasing EPC"}
 
     _allowed_housing_types = {"Social", "Private"}

backend/ml_models/Valuation.py

@@ -103,6 +103,8 @@ class PropertyValuation:
         # Vander Elliot Intrusive surveys
         12103116: 1_537_000,
         12103117: 1_404_000,
+        # GLA Proposal
+        100020606627: 409_000
     }
 
     # We base our valuation uplifts on a number of sources

etl/customers/gla/example_model_outputs.py

@@ -0,0 +1,38 @@
+import pandas as pd
+from utils.s3 import save_csv_to_s3
+
+asset_list = [
+    {
+        "address": "4, King Henrys Drive",
+        "postcode": "CR0 0PA"
+    },
+]
+portfolio_id = 110
+user_id = 8
+
+asset_list = pd.DataFrame(asset_list)
+
+filename = f"{user_id}/{portfolio_id}/asset_list.csv"
+save_csv_to_s3(
+    dataframe=asset_list,
+    bucket_name="retrofit-plan-inputs-dev",
+    file_name=filename
+)
+
+body1 = {
+    "portfolio_id": str(portfolio_id),
+    "housing_type": "Private",
+    "goal": "Increasing EPC",
+    "goal_value": "A",
+    "trigger_file_path": filename,
+    "already_installed_file_path": "",
+    "patches_file_path": "",
+    "non_invasive_recommendations_file_path": "",
+    "inclusions": [
+        "cavity_wall_insulation", "loft_insulation", "air_source_heat_pump", "solar_pv"
+    ],
+    "budget": None,
+    "scenario_name": "Whole House",
+    "multi_plan": False,
+}
+print(body1)

etl/customers/gla/proposal_investigation.py

@@ -0,0 +1,173 @@
+"""
+This script performs some basic analysis to identify EPC data for postcodes specified in the Warmer Homes Local Grant
+"""
+
+import inspect
+import requests
+import json
+import pandas as pd
+from pathlib import Path
+from etl.ownership.Ownership import Ownership
+
+postcodes = pd.read_excel(
+    "/Users/khalimconn-kowlessar/Downloads/WHLG-eligible-postcodes_RP edit.xlsx", sheet_name='Eligible postcodes'
+)
+# Take just the first three columns
+postcodes = postcodes[
+    ['List of eligible postcodes via the IMD Income Decile 1-2 pathway', 'Unnamed: 1', 'Unnamed: 2']
+]
+
+postcodes.columns = ['postcode', 'Local Authority', 'London Borough?']
+# Drop the first row
+postcodes = postcodes.drop([0, 1])
+# Take just the London Boroughs
+postcodes = postcodes[postcodes["London Borough?"] == "Yes"]
+# Since there are a large number of potcodes (425k), let's just take a few examples
+# Take postcodes that begin with "BN15"
+# postcodes = postcodes[postcodes["postcode"].str.startswith("BN15")]
+
+# The Local Authority is Adur, so let's get the EPC data for this area
+# epc_data = pd.read_csv(
+#     "/Users/khalimconn-kowlessar/Documents/hestia/Model/local_data/all-domestic-certificates/domestic-E07000223-Adur"
+#     "/certificates.csv", low_memory=False
+# )
+# # Filter on these postcodes
+# epc_data = epc_data[epc_data["POSTCODE"].str.lower().isin(postcodes["postcode"].str.lower())]
+# epc_data = epc_data[~pd.isnull(epc_data["UPRN"])]
+# # Take the newest EPC for each UPRN, based on LODGEMENT_DATE
+# epc_data["LODGEMENT_DATE"] = pd.to_datetime(epc_data["LODGEMENT_DATE"])
+# epc_data = epc_data.sort_values("LODGEMENT_DATE", ascending=False).drop_duplicates("UPRN")
+#
+# # Let's look at the breakdown of EPC ratings. We want the count and the % of the total
+# ratings_distribution = epc_data.groupby("CURRENT_ENERGY_RATING").size().reset_index()
+# ratings_distribution.columns = ["Rating", "Count"]
+# ratings_distribution["Percentage"] = ratings_distribution["Count"] / ratings_distribution["Count"].sum() * 100
+
+# Can we identify the owners of these units so we can contact them?
+
+file_src = inspect.getfile(lambda x: None)
+DATA_DIRECTORY = Path(file_src).parent / "local_data" / "all-domestic-certificates"
+epc_paths = [entry for entry in DATA_DIRECTORY.iterdir() if entry.is_dir()]
+epc_paths = [str(entry / "certificates.csv") for entry in epc_paths]
+
+ownership = Ownership(
+    epc_paths=epc_paths,
+    domestic_ownership_path="/Users/khalimconn-kowlessar/Downloads/CCOD_FULL_2024_07.csv",
+    overseas_ownership_path="/Users/khalimconn-kowlessar/Downloads/OCOD_FULL_2024_07.csv",
+    land_registry_path="/Users/khalimconn-kowlessar/Downloads/pp-complete.csv",
+    project_name="gla-proposal",
+    bucket="retrofit-data-dev",
+    average_property_value=0,
+    portfolio_value=0,
+    excluded_owners=[],
+    excluded_uprns=[],
+    save=True
+)
+
+# Data will be found at ownership/gla-proposal
+ownership.source_epc_properties(column_filters={}, postcodes=postcodes["postcode"].str.lower().tolist())
+
+# Step 2: Get company ownership data
+ownership.load_company_ownership()
+
+# Step 3: Prepare data for matching
+ownership.prepare_for_matching()
+
+# Step 4: Match EPC data to ownership data
+ownership.match()
+
+from utils.s3 import save_excel_to_s3, read_excel_from_s3
+
+# Save the data to S3
+# save_excel_to_s3(
+#     df=ownership.matched_addresses,
+#     bucket_name=ownership.bucket,
+#     file_key=ownership.matched_addresses_pre_filter_filepath
+# )
+
+# Read in matches
+matches = read_excel_from_s3(
+    bucket_name=ownership.bucket,
+    file_key="ownership/gla-proposal/2024-10-10 19:02:34.131365/matched_addresses_pre_filter.xlsx",
+    header_row=0
+)
+
+# We have the matches, which we now need to match to the postcodes
+matches = ownership.matched_addresses.copy()
+# filter matches on the postcodes we're interested in
+matches = matches[matches["epc_postcode"].str.lower().isin(postcodes["postcode"].str.lower())]
+# Remove any social transactions
+matches = matches[~matches["TENURE"].isin(
+    ["Rented (social)", "rental (social)",
+     "Not defined - use in the case of a new dwelling for which the intended tenure in not known. It is not to be "
+     "used for an existing dwelling", "NO DATA!"])
+]
+matches["is_prs"] = matches["TENURE"].isin(["rental (private)", "Rented (private)"])
+# Look at the EPC ratings
+epc_ratings = matches.groupby(["CURRENT_ENERGY_RATING"]).size().reset_index()
+epc_ratings.columns = ["EPC Rating", "Count"]
+epc_ratings["Percentage"] = epc_ratings["Count"] / epc_ratings["Count"].sum() * 100
+
+# Take properties that are below an EPC C rating, as defined by the guidance and remove any new builds
+matches = matches[matches["CURRENT_ENERGY_RATING"].isin(["D", "E", "F", "G"])]
+# 11,694 properties
+matches["epc_postcode"].nunique()
+# 6899
+
+owners_count = matches.groupby(['Proprietor Name (1)', 'Company Registration No. (1)']).size().reset_index()
+owners_count.columns = ['Owner', 'Owner Registration #', 'Count']
+owners_count = owners_count.sort_values('Count', ascending=False)
+owners_count["Percentage"] = owners_count["Count"] / owners_count["Count"].sum() * 100
+
+# Take an example postal region
+matches = matches.sort_values("epc_postcode", ascending=True)
+# BR1, BR5
+example = matches[matches["epc_postcode"].str.startswith("CR0 ")].copy()
+example = example[example["TENURE"].isin(["rental (private)", "Rented (private)"])]
+
+pd.set_option('display.max_rows', 500)
+pd.set_option('display.max_columns', 500)
+pd.set_option('display.width', 1000)
+example[
+    ["epc_address", "epc_postcode", "CURRENT_ENERGY_RATING", "CURRENT_ENERGY_EFFICIENCY", "Proprietor Name (1)",
+     "Company Registration No. (1)"]
+].head(4)
+
+ownership.epc_data["UPRN"] = ownership.epc_data["UPRN"].astype(int)
+example = example.merge(
+    ownership.epc_data[["UPRN", "BUILT_FORM", "PROPERTY_TYPE", "WALLS_DESCRIPTION", "ROOF_DESCRIPTION"]],
+    on="UPRN",
+    how="left"
+)
+z = example[example["CURRENT_ENERGY_RATING"] == "E"]
+z = z[z["TENURE"].isin(["rental (private)", "Rented (private)"])]
+
+companies_house_api_key = "1d9c2877-3271-4642-80ed-a6170971653f"
+
+company_number = example.head(1)["Company Registration No. (1)"].values[0]
+url = f'https://api.company-information.service.gov.uk/company/{company_number}'
+
+# Make the API request
+response = requests.get(url, auth=(companies_house_api_key, ''))
+
+# Check if the request was successful
+if response.status_code == 200:
+    company_data = response.json()
+    # Pretty-print the fetched data
+    print(json.dumps(company_data, indent=4))
+else:
+    print(f"Failed to fetch data. Status code: {response.status_code}")
+    # Try appending a zero the beginning of the company number
+    company_number = f"0{company_number}"
+    url = f'https://api.company-information.service.gov.uk/company/{company_number}'
+    response = requests.get(url, auth=(companies_house_api_key, ''))
+    company_data = response.json()
+
+from pprint import pprint
+
+pprint(company_data)
+
+psc_url = f'https://api.company-information.service.gov.uk/company/{company_number}/persons-with-significant-control'
+psc_response = requests.get(psc_url, auth=(companies_house_api_key, ''))
+psc_data = psc_response.json()
+pprint(psc_data)

etl/ownership/Ownership.py

@@ -61,6 +61,7 @@ class Ownership:
         portfolio_value: float,
         excluded_owners: List[str] = None,
         excluded_uprns: List[int] = None,
+        save=True
     ):
         """
 
@@ -115,6 +116,8 @@ class Ownership:
             f"ownership/{self.project_name}/{self.run_timestamp}/portfolio_epc_data.xlsx"
         )
 
+        self.save = save
+
         # Data
         self.epc_data = None
         self.ownership_data = None
@@ -158,21 +161,22 @@ class Ownership:
         # Step 5: Match land registry data to existing matches
         self.match_with_land_registry()
         # We store this data in s3 before we perform any filtering
-        save_excel_to_s3(
-            df=self.matched_addresses,
-            bucket_name=self.bucket,
-            file_key=self.matched_addresses_pre_filter_filepath
-        )
-        save_excel_to_s3(
-            df=self.combined_matching_lookup,
-            bucket_name=self.bucket,
-            file_key=self.combined_matching_lookup_pre_filter_filepath
-        )
+        if self.save:
+            save_excel_to_s3(
+                df=self.matched_addresses,
+                bucket_name=self.bucket,
+                file_key=self.matched_addresses_pre_filter_filepath
+            )
+            save_excel_to_s3(
+                df=self.combined_matching_lookup,
+                bucket_name=self.bucket,
+                file_key=self.combined_matching_lookup_pre_filter_filepath
+            )
 
         # Prepare the final outputs:
         self.create_final_matches()
 
-    def source_epc_properties(self, column_filters=None):
+    def source_epc_properties(self, column_filters=None, postcodes=None):
         """
         This function will filter the epc data as specified by column filters, searching across all of the EPC tables
         :param column_filters: Dictionary with column names as keys and list of acceptable values as values. This
@@ -180,6 +184,7 @@ class Ownership:
                                 {"column_name": ["value1", "value2", ...]}, where column_name is the name of the column
                                 in the EPC data and ["value1", "value2", ...] is a list of acceptable values for that
                                 column. If a column is not found in the EPC data, an exception is raised.
+        :param postcodes: A list of postcodes to filter the data on
         """
 
         column_filters = {} if column_filters is None else column_filters
@@ -203,6 +208,11 @@ class Ownership:
                 else:
                     raise Exception(f"Column {column} not found in data. column_filters is malformed")
 
+            if postcodes is not None:
+                epc_data = epc_data[epc_data["POSTCODE"].str.lower().isin(postcodes)]
+            if epc_data.empty:
+                continue
+
             data.append(epc_data)
 
         self.epc_data = pd.concat(data, ignore_index=True)
@@ -210,12 +220,13 @@ class Ownership:
         if self.excluded_uprns:
             self.epc_data = self.epc_data[~self.epc_data["UPRN"].astype(float).isin(self.excluded_uprns)]
 
-        # We now store the data in s3
-        save_excel_to_s3(
-            df=self.epc_data,
-            bucket_name=self.bucket,
-            file_key=self.epc_data_filepath
-        )
+        if self.save:
+            # We now store the data in s3
+            save_excel_to_s3(
+                df=self.epc_data,
+                bucket_name=self.bucket,
+                file_key=self.epc_data_filepath
+            )
 
     def load_company_ownership(self):
         """
@@ -484,11 +495,11 @@ class Ownership:
                 house_no = house_no.replace(",", "")
 
             if house_no is None:
-                # It's hard for us to get a reliable match
-                # filtered = filtered[filtered["Property Address"].str.contains(address["ADDRESS1"])]
-                # if filtered.shape[0] > 1:
-                #     raise Exception("No valid - maybe we should do levenstein?")
-                continue
+                # If the house number is missing, it means that we usually have a named property so we look for an
+                # exact match on that name
+                filtered = filtered[filtered["Property Address"].str.lower().str.contains(address["ADDRESS"].lower())]
+                if filtered.shape[0] != 1:
+                    continue
 
             else:
 
@@ -590,7 +601,8 @@ class Ownership:
                     "CURRENT_ENERGY_RATING",
                     "POSTCODE",
                     "LODGEMENT_DATE",
-                    "TRANSACTION_TYPE"
+                    "TRANSACTION_TYPE",
+                    "TENURE",
                 ]
             ].rename(
                 columns={
@@ -1002,25 +1014,26 @@ class Ownership:
         if self.portfolio_properties["UPRN"].nunique() != self.portfolio_epc_data["UPRN"].nunique():
             raise ValueError("Portfolio properties and epc data don't match")
 
-        logger.info("Storing final outpus")
-        # Store data
-        save_excel_to_s3(
-            df=self.portfolio_owners,
-            bucket_name=self.bucket,
-            file_key=self.portfolio_owners_filepath,
-        )
+        if self.save:
+            logger.info("Storing final outpus")
+            # Store data
+            save_excel_to_s3(
+                df=self.portfolio_owners,
+                bucket_name=self.bucket,
+                file_key=self.portfolio_owners_filepath,
+            )
 
-        save_excel_to_s3(
-            df=self.portfolio_properties,
-            bucket_name=self.bucket,
-            file_key=self.portfolio_properties_filepath,
-        )
+            save_excel_to_s3(
+                df=self.portfolio_properties,
+                bucket_name=self.bucket,
+                file_key=self.portfolio_properties_filepath,
+            )
 
-        save_excel_to_s3(
-            df=self.portfolio_epc_data,
-            bucket_name=self.bucket,
-            file_key=self.portfolio_epc_data_filepath,
-        )
+            save_excel_to_s3(
+                df=self.portfolio_epc_data,
+                bucket_name=self.bucket,
+                file_key=self.portfolio_epc_data_filepath,
+            )
 
     def get_asset_list(self):
         """