37 lines
1.2 KiB
Python
37 lines
1.2 KiB
Python
|
from selenium import webdriver
|
||
|
|
from selenium.webdriver.chrome.options import Options
|
||
|
|
import json
|
||
|
|
|
||
|
|
# Configure the Selenium driver options
|
||
|
|
options = Options()
|
||
|
|
options.add_argument('--headless')
|
||
|
|
options.add_argument('--disable-gpu')
|
||
|
|
driver = webdriver.Chrome(options=options)
|
||
|
|
|
||
|
|
# Load the page and extract the __NEXT_DATA__ data
|
||
|
|
url = 'https://www.sayweee.com/en/product/Pete-and-Gerry-s-Organic-Large-Brown-Eggs/65902?category=dairy02'
|
||
|
|
driver.get(url)
|
||
|
|
next_data_element = None
|
||
|
|
for script_element in driver.find_elements_by_xpath('//script'):
|
||
|
|
if '__NEXT_DATA__' in script_element.get_attribute('innerHTML'):
|
||
|
|
next_data_element = script_element
|
||
|
|
break
|
||
|
|
|
||
|
|
if next_data_element is not None:
|
||
|
|
next_data_text = next_data_element.get_attribute('innerHTML')
|
||
|
|
next_data_start_index = next_data_text.index('{')
|
||
|
|
next_data_end_index = next_data_text.rindex('}') + 1
|
||
|
|
next_data_json = next_data_text[next_data_start_index:next_data_end_index]
|
||
|
|
next_data = json.loads(next_data_json)
|
||
|
|
|
||
|
|
# Extract the price from the __NEXT_DATA__ data
|
||
|
|
price = next_data['props']['pageProps']['product']['price']
|
||
|
|
|
||
|
|
# Print the price
|
||
|
|
print(price)
|
||
|
|
else:
|
||
|
|
print('__NEXT_DATA__ element not found')
|
||
|
|
|
||
|
|
# Close the Selenium driver
|
||
|
|
driver.quit()
|