Splunk Apps » Website Input

Apparently, you need to install these executables on the path to get this to work: https://github.com/mozilla/geckodriver/releases.

Might want to try updating to Selenium 3.0: http://www.seleniumhq.org/download/

Here is something that worked.

1. Download selenium 3.0.1: https://pypi.python.org/pypi/selenium
2. Deploy client in same directory as script: https://github.com/mozilla/geckodriver/releases
3. Update page to include local directory

from selenium import webdriver
import os
os.environ["PATH"] += os.pathsep + "." 
browser = webdriver.Firefox()
browser.get('http://seleniumhq.org/')

Will need to use sys.platform to determine platform. platform.architecture() will indicate whether the host is 32-bit or 64-bit.

Here is a test script. This works.

import os
from selenium import webdriver
new_path = os.path.join(os.getcwd(), "browser_drivers", "darwin")
print new_path
os.environ["PATH"] += ":" + new_path
browser = webdriver.Firefox()
browser.get('http://seleniumhq.org/')

This is failing when the script is run within Splunk. This is because it is running from the root directory and the drivers can't write the log file in this directory.

Some options for this:

1. Run selenium extraction code from another process with a different working directory
2. Get selenium to make the logs in a different location

Changing the path should be possible per https://github.com/SeleniumHQ/selenium/pull/2700/commits/95f5ef23bb454b9a02274fe4d41fe03e014e704c

Need to make sure that the binary is set to executable when it is packaged.