ksvmuralidhar
commited on
Commit
•
046cc67
1
Parent(s):
152aad4
Update scraper.py
Browse files- scraper.py +4 -1
scraper.py
CHANGED
@@ -11,9 +11,12 @@ def scrape_text(url, n_words=15):
|
|
11 |
try:
|
12 |
driver = None
|
13 |
logging.warning("Initiated Scraping")
|
|
|
|
|
|
|
14 |
opts = FirefoxOptions()
|
15 |
opts.add_argument("--headless")
|
16 |
-
driver = webdriver.Firefox(options=opts)
|
17 |
driver.set_page_load_timeout(60)
|
18 |
driver.get(url)
|
19 |
elem = driver.find_element(By.TAG_NAME, "body").text
|
|
|
11 |
try:
|
12 |
driver = None
|
13 |
logging.warning("Initiated Scraping")
|
14 |
+
firefox_profile = webdriver.FirefoxProfile()
|
15 |
+
user_agent = "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/123.0.0.0 Safari/537.36"
|
16 |
+
firefox_profile.set_preference("general.useragent.override", user_agent)
|
17 |
opts = FirefoxOptions()
|
18 |
opts.add_argument("--headless")
|
19 |
+
driver = webdriver.Firefox(options=opts, firefox_profile=firefox_profile)
|
20 |
driver.set_page_load_timeout(60)
|
21 |
driver.get(url)
|
22 |
elem = driver.find_element(By.TAG_NAME, "body").text
|