HomeHarvest/homeharvest/core/scrapers/__init__.py

76 lines
2.4 KiB
Python
Raw Normal View History

2023-09-15 15:17:37 -07:00
from dataclasses import dataclass
import requests
2024-04-16 12:55:44 -07:00
import uuid
2023-09-17 16:30:37 -07:00
from .models import Property, ListingType, SiteName
2023-09-15 15:17:37 -07:00
@dataclass
class ScraperInput:
location: str
listing_type: ListingType
2023-10-02 13:58:47 -07:00
radius: float | None = None
mls_only: bool | None = None
2023-09-19 13:43:24 -07:00
proxy: str | None = None
last_x_days: int | None = None
2023-11-03 16:35:41 -07:00
date_from: str | None = None
date_to: str | None = None
2024-03-03 09:45:28 -08:00
foreclosure: bool | None = None
2023-09-15 15:17:37 -07:00
class Scraper:
def __init__(
self,
scraper_input: ScraperInput,
2023-10-05 08:10:48 -07:00
session: requests.Session = None,
):
2023-09-15 15:17:37 -07:00
self.location = scraper_input.location
2023-09-18 08:16:59 -07:00
self.listing_type = scraper_input.listing_type
2023-09-28 18:40:16 -07:00
if not session:
self.session = requests.Session()
2024-04-16 12:55:44 -07:00
self.session.headers.update(
{
"auth": f"Bearer {self.get_access_token()}",
"apollographql-client-name": "com.move.Realtor-apollo-ios",
}
)
2023-09-28 18:40:16 -07:00
else:
self.session = session
2023-09-19 14:05:14 -07:00
if scraper_input.proxy:
proxy_url = scraper_input.proxy
2023-09-19 19:13:20 -07:00
proxies = {"http": proxy_url, "https": proxy_url}
2023-09-19 14:05:14 -07:00
self.session.proxies.update(proxies)
2023-09-28 18:40:16 -07:00
2023-09-17 16:30:37 -07:00
self.listing_type = scraper_input.listing_type
2023-10-02 13:58:47 -07:00
self.radius = scraper_input.radius
self.last_x_days = scraper_input.last_x_days
self.mls_only = scraper_input.mls_only
2023-11-03 16:35:41 -07:00
self.date_from = scraper_input.date_from
self.date_to = scraper_input.date_to
2024-03-03 09:45:28 -08:00
self.foreclosure = scraper_input.foreclosure
2023-09-15 15:17:37 -07:00
2024-04-16 12:55:44 -07:00
def search(self) -> list[Property]: ...
2023-09-15 15:42:47 -07:00
@staticmethod
2024-04-16 12:55:44 -07:00
def _parse_home(home) -> Property: ...
2023-09-15 20:58:54 -07:00
2024-04-16 12:55:44 -07:00
def handle_location(self): ...
def get_access_token(self):
url = "https://graph.realtor.com/auth/token"
payload = f'{{"client_app_id":"rdc_mobile_native,24.20.4.149916,iphone","device_id":"{str(uuid.uuid4()).upper()}","grant_type":"device_mobile"}}'
headers = {
"Host": "graph.realtor.com",
"x-client-version": "24.20.4.149916",
"accept": "*/*",
"content-type": "Application/json",
"user-agent": "Realtor.com/24.20.4.149916 CFNetwork/1410.0.3 Darwin/22.6.0",
"accept-language": "en-US,en;q=0.9",
}
response = requests.post(url, headers=headers, data=payload)
data = response.json()
return data["access_token"]