added foreclose parameter

pull/55/head
Roberto Mora 2024-03-01 22:58:35 -05:00
parent 03198428de
commit 6d987e8fc2
4 changed files with 33 additions and 3 deletions

View File

@ -15,6 +15,7 @@ def scrape_property(
proxy: str = None, proxy: str = None,
date_from: str = None, date_from: str = None,
date_to: str = None, date_to: str = None,
foreclosure: bool = None,
) -> pd.DataFrame: ) -> pd.DataFrame:
""" """
Scrape properties from Realtor.com based on a given location and listing type. Scrape properties from Realtor.com based on a given location and listing type.
@ -38,6 +39,7 @@ def scrape_property(
last_x_days=past_days, last_x_days=past_days,
date_from=date_from, date_from=date_from,
date_to=date_to, date_to=date_to,
foreclosure=foreclosure,
) )
site = RealtorScraper(scraper_input) site = RealtorScraper(scraper_input)

View File

@ -13,6 +13,7 @@ class ScraperInput:
last_x_days: int | None = None last_x_days: int | None = None
date_from: str | None = None date_from: str | None = None
date_to: str | None = None date_to: str | None = None
foreclosure: bool | None = None
class Scraper: class Scraper:
@ -40,6 +41,7 @@ class Scraper:
self.mls_only = scraper_input.mls_only self.mls_only = scraper_input.mls_only
self.date_from = scraper_input.date_from self.date_from = scraper_input.date_from
self.date_to = scraper_input.date_to self.date_to = scraper_input.date_to
self.foreclosure = scraper_input.foreclosure
def search(self) -> list[Property]: def search(self) -> list[Property]:
... ...

View File

@ -383,6 +383,13 @@ class RealtorScraper(Scraper):
) )
listing_type = ListingType.FOR_SALE if self.listing_type == ListingType.PENDING else self.listing_type listing_type = ListingType.FOR_SALE if self.listing_type == ListingType.PENDING else self.listing_type
is_foreclosure = ""
if 'foreclosure' in variables and variables['foreclosure'] == True:
is_foreclosure = "foreclosure: true"
if 'foreclosure' in variables and variables['foreclosure'] == False:
is_foreclosure = "foreclosure: false"
if search_type == "comps": #: comps search, came from an address if search_type == "comps": #: comps search, came from an address
query = """query Property_search( query = """query Property_search(
@ -392,6 +399,7 @@ class RealtorScraper(Scraper):
) { ) {
home_search( home_search(
query: { query: {
%s
nearby: { nearby: {
coordinates: $coordinates coordinates: $coordinates
radius: $radius radius: $radius
@ -404,6 +412,7 @@ class RealtorScraper(Scraper):
limit: 200 limit: 200
offset: $offset offset: $offset
) %s""" % ( ) %s""" % (
is_foreclosure,
listing_type.value.lower(), listing_type.value.lower(),
date_param, date_param,
pending_or_contingent_param, pending_or_contingent_param,
@ -420,6 +429,7 @@ class RealtorScraper(Scraper):
) { ) {
home_search( home_search(
query: { query: {
%s
city: $city city: $city
county: $county county: $county
postal_code: $postal_code postal_code: $postal_code
@ -432,6 +442,7 @@ class RealtorScraper(Scraper):
limit: 200 limit: 200
offset: $offset offset: $offset
) %s""" % ( ) %s""" % (
is_foreclosure,
listing_type.value.lower(), listing_type.value.lower(),
date_param, date_param,
pending_or_contingent_param, pending_or_contingent_param,
@ -586,6 +597,9 @@ class RealtorScraper(Scraper):
"postal_code": location_info.get("postal_code"), "postal_code": location_info.get("postal_code"),
} }
if self.foreclosure:
search_variables['foreclosure'] = self.foreclosure
result = self.general_search(search_variables, search_type=search_type) result = self.general_search(search_variables, search_type=search_type)
total = result["total"] total = result["total"]
homes = result["properties"] homes = result["properties"]

View File

@ -139,3 +139,15 @@ def test_realtor_bad_address():
if len(bad_results) == 0: if len(bad_results) == 0:
assert True assert True
def test_realtor_foreclosed():
foreclosed = scrape_property(
location="Dallas, TX", listing_type="for_sale", past_days=100, foreclosure=True
)
not_foreclosed = scrape_property(
location="Dallas, TX", listing_type="for_sale", past_days=100, foreclosure=False
)
assert len(foreclosed) != len(not_foreclosed)