From ba249ca20d4c7bc531c9e023a11672087eaa994f Mon Sep 17 00:00:00 2001 From: Zachary Hampton <69336300+ZacharyHampton@users.noreply.github.com> Date: Mon, 18 Sep 2023 08:26:35 -0700 Subject: [PATCH] - redfin buildings support --- homeharvest/__init__.py | 2 +- homeharvest/core/scrapers/redfin/__init__.py | 35 ++++++++++++++++++-- 2 files changed, 34 insertions(+), 3 deletions(-) diff --git a/homeharvest/__init__.py b/homeharvest/__init__.py index 4afbbc5..2b53b13 100644 --- a/homeharvest/__init__.py +++ b/homeharvest/__init__.py @@ -71,7 +71,7 @@ def process_result(result: Union[Building, Property]) -> pd.DataFrame: address_data = prop_data["address"] prop_data["site_name"] = prop_data["site_name"] prop_data["listing_type"] = prop_data["listing_type"].value - prop_data["property_type"] = prop_data["property_type"].value.lower() if prop_data["property_type"] else None + prop_data["property_type"] = prop_data["property_type"].value.lower() if prop_data.get("property_type") else None prop_data["address_one"] = address_data.address_one prop_data["city"] = address_data.city prop_data["state"] = address_data.state diff --git a/homeharvest/core/scrapers/redfin/__init__.py b/homeharvest/core/scrapers/redfin/__init__.py index 29855a7..778a49f 100644 --- a/homeharvest/core/scrapers/redfin/__init__.py +++ b/homeharvest/core/scrapers/redfin/__init__.py @@ -1,5 +1,5 @@ import json -from ..models import Property, Address, PropertyType +from ..models import Property, Address, PropertyType, Building from .. import Scraper from typing import Any @@ -86,6 +86,34 @@ def get_value(key: str) -> Any | None: mls_id=get_value("mlsId"), ) + def _parse_building(self, building: dict) -> Building: + return Building( + address=Address( + address_one=" ".join( + [ + building['address']['streetNumber'], + building['address']['directionalPrefix'], + building['address']['streetName'], + building['address']['streetType'], + ] + ), + city=building['address']['city'], + state=building['address']['stateOrProvinceCode'], + zip_code=building['address']['postalCode'], + address_two=" ".join( + [ + building['address']['unitType'], + building['address']['unitValue'], + ] + ) + ), + site_name=self.site_name, + url="https://www.redfin.com{}".format(building["url"]), + listing_type=self.listing_type, + num_units=building["numUnitsForSale"], + ) + + def handle_address(self, home_id: str): """ EPs: @@ -123,5 +151,8 @@ def search(self): homes = [ self._parse_home(home) for home in response_json["payload"]["homes"] - ] #: support buildings + ] + [ + self._parse_building(building) for building in response_json["payload"]["buildings"].values() + ] + return homes