53 lines
2.1 KiB
Python
53 lines
2.1 KiB
Python
|
import logging
|
||
|
|
from playwright.sync_api import sync_playwright
|
||
|
|
from fake_useragent import UserAgent
|
||
|
|
import time
|
||
|
|
import playwright
|
||
|
|
from playwright_stealth import stealth_sync
|
||
|
|
def get_raw_product_data(url):
|
||
|
|
retries = 1
|
||
|
|
for _ in range(retries):
|
||
|
|
try:
|
||
|
|
with sync_playwright() as p:
|
||
|
|
browser = p.chromium.launch(
|
||
|
|
headless=False,
|
||
|
|
args=[
|
||
|
|
"--disable-dev-shm-usage",
|
||
|
|
"--disable-blink-features=AutomationControlled",
|
||
|
|
"--disable-component-extensions-with-background-pages"
|
||
|
|
]
|
||
|
|
)
|
||
|
|
ua = UserAgent(platforms='mobile')
|
||
|
|
random_mobile_ua = ua.random
|
||
|
|
logging.info("using user agent: {}".format(random_mobile_ua))
|
||
|
|
|
||
|
|
context = browser.new_context(user_agent=random_mobile_ua)
|
||
|
|
context.add_init_script("Object.defineProperty(navigator, 'webdriver', {get: () => undefined})")
|
||
|
|
page = context.new_page()
|
||
|
|
stealth_sync(page)
|
||
|
|
|
||
|
|
api_requests = {}
|
||
|
|
|
||
|
|
try:
|
||
|
|
page.goto(url, timeout=5000)
|
||
|
|
time.sleep(1)
|
||
|
|
page.reload()
|
||
|
|
with page.expect_response("**/wap/v2/product/detail**") as response:
|
||
|
|
api_requests = response.value.json()
|
||
|
|
except playwright._impl._errors.TimeoutError:
|
||
|
|
logging.info("Timeout occurred. Retrying.....")
|
||
|
|
page.reload()
|
||
|
|
with page.expect_response("**/wap/v2/product/detail**") as response:
|
||
|
|
api_requests = response.value.json()
|
||
|
|
finally:
|
||
|
|
browser.close()
|
||
|
|
|
||
|
|
return api_requests
|
||
|
|
except Exception as e:
|
||
|
|
logging.error(f"An error occurred: {str(e)}")
|
||
|
|
logging.info("Retrying...")
|
||
|
|
|
||
|
|
print(api_requests)
|
||
|
|
|
||
|
|
|
||
|
|
get_raw_product_data('https://hasaki.vn/san-pham/tinh-chat-chong-nang-sunplay-hieu-chinh-sac-da-50g-xanh-87613.html')
|