selenium_web_scrape / extract.py
jonathanjordan21's picture
Update extract.py
9787462 verified
raw
history blame
1.88 kB
from selenium import webdriver
from selenium.common.exceptions import WebDriverException
from PIL import Image
from io import BytesIO
def take_webdata(url):
options = webdriver.ChromeOptions()
options.add_argument('--headless')
options.add_argument('--no-sandbox')
options.add_argument('--disable-dev-shm-usage')
try:
wd = webdriver.Chrome(options=options)
wd.set_window_size(1080, 720) # Adjust the window size here
wd.get(url)
wd.implicitly_wait(5)
# Get the page title
page_title = wd.title
screenshot = wd.get_screenshot_as_png()
except WebDriverException as e:
return Image.new('RGB', (1, 1)), page_title
finally:
if wd:
wd.quit()
return Image.open(BytesIO(screenshot)) , page_title
def get_vehicle_info(plate_number: str):
# Configure headless Chrome
options = Options()
options.add_argument("--headless")
options.add_argument("--disable-gpu")
options.add_argument("--no-sandbox")
# Path to chromedriver (adjust if needed)
driver = webdriver.Chrome(options=options)
try:
# Step 1: Open the site
driver.get("https://www.jambisamsat.net/infopkb.html")
time.sleep(2)
# Step 2: Find the input and enter plate number
input_element = driver.find_element(By.NAME, "nopol")
input_element.send_keys(plate_number)
# Step 3: Submit the form
submit_button = driver.find_element(By.CSS_SELECTOR, 'input[type="submit"]')
submit_button.click()
time.sleep(2)
driver.implicitly_wait(5)
page_title = driver.title
screenshot = driver.get_screenshot_as_png()
return Image.open(BytesIO(screenshot)) , page_title
except Exception as e:
return f"Error occurred: {e}"
finally:
driver.quit()