Update scrape.py
Browse files
scrape.py
CHANGED
|
@@ -2,25 +2,20 @@ from selenium import webdriver # type: ignore
|
|
| 2 |
from selenium.webdriver.chrome.service import Service # type: ignore
|
| 3 |
from selenium.webdriver.chrome.options import Options # type: ignore
|
| 4 |
from bs4 import BeautifulSoup # type: ignore
|
| 5 |
-
from dotenv import load_dotenv
|
| 6 |
import os
|
| 7 |
import time
|
| 8 |
|
| 9 |
-
# Load environment variables
|
| 10 |
load_dotenv()
|
| 11 |
|
| 12 |
-
CHROME_DRIVER_PATH = os.getenv("
|
| 13 |
|
| 14 |
def scrape_website(website):
|
| 15 |
print("Connecting to Chrome Browser...")
|
| 16 |
|
| 17 |
-
# Setup ChromeDriver service and options
|
| 18 |
service = Service(CHROME_DRIVER_PATH)
|
| 19 |
options = Options()
|
| 20 |
-
options.add_argument("--headless") # Run Chrome in headless mode
|
| 21 |
-
options.add_argument("--no-sandbox")
|
| 22 |
-
options.add_argument("--disable-dev-shm-usage")
|
| 23 |
-
|
| 24 |
driver = webdriver.Chrome(service=service, options=options)
|
| 25 |
|
| 26 |
try:
|
|
|
|
| 2 |
from selenium.webdriver.chrome.service import Service # type: ignore
|
| 3 |
from selenium.webdriver.chrome.options import Options # type: ignore
|
| 4 |
from bs4 import BeautifulSoup # type: ignore
|
| 5 |
+
from dotenv import load_dotenv # type: ignore
|
| 6 |
import os
|
| 7 |
import time
|
| 8 |
|
|
|
|
| 9 |
load_dotenv()
|
| 10 |
|
| 11 |
+
CHROME_DRIVER_PATH = os.getenv("./chrome")
|
| 12 |
|
| 13 |
def scrape_website(website):
|
| 14 |
print("Connecting to Chrome Browser...")
|
| 15 |
|
| 16 |
+
# Setup ChromeDriver service and options
|
| 17 |
service = Service(CHROME_DRIVER_PATH)
|
| 18 |
options = Options()
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
driver = webdriver.Chrome(service=service, options=options)
|
| 20 |
|
| 21 |
try:
|