xpathwebdriver

A python wrapper for interacting with Selenium through XPath and CSS selectors. You can now use XPaths like //div/text():

from xpathwebdriver.browser import Browser

browser = Browser()

# xpath returns text
for idx, t in enumerate(browser.select_xpath('//div/text()')):
    print(idx, t)

# css selector returns elements
for idx, elem in enumerate(browser.select_css('.result__title')):
    print(idx, elem.text)

Which will return you a string as expected. Something webdriver API makes more complicated.

Means you can write all your tests based on XPath.

Also adds:

• Interactive shell for testing XPath manually and easily against a live browser
  • You can share the interactive shell with a script, to keep track of errors/debugging
• Multiple browser management
• Browser life management (whether to keep the browser open or kill it on exit)
  • Management is done through python contexts (with statement)
• Useful settings for local and remote (headless) testing
  • Also supports environment variables as settings
  • Plus allowing custom settings that you can also push through environment variables
• Screenshots comparison and diff management
• Virtual display management (so you can run "headless" in a remote instance)
  • you can use VNC to access the remote Browser
• Adding xpath, css, selector methods to returned WebElement objects, to keep the Xpath functionality

Ubuntu quick installation

You can opt to use Chromium to simplify installation:

sudo apt-get install -y python3-pip imagemagick findimagedupes tightvncserver xserver-xephyr xvfb unzip chromium-browser
# chromium-chromedriver (from 22.04 and on you no longer need to install this package, seems it's obsolete and comes with chromium itself)
sudo pip3 install xpathwebdriver Pillow ipython

You can quickly test it running:

xpathshell

That will open an interactive shell with a browser object. Use TAB to autocomplete available API. Use browser.driver to directly access the webdriver object.

General installation

pip install xpathwebdriver

1. Install xpathwebdriver using pip.
2. Install google chrome.
3. Download chromedriver for your Chrome version and install it in your path. https://chromedriver.chromium.org/downloads
4. That generally should work on a modern Linux System (not tested but should also work on other oses). Try example in section below
5. For image comparison install pip install Pillow, findimagedupes and imagemagick packages for your OS. (they are not automatically installed to keep basic requirements low) On ubuntu: sudo apt install imagemagick findimagedupes
6. For interactive shell install pip install ipython

Check "Installing Selenium" section for other browsers and details.

Example

from xpathwebdriver.browser import Browser

browser = Browser()
browser.get_url('https://duckduckgo.com/')
browser.fill(".//*[@id='search_form_input_homepage']", 'xpathwebdriver\n')
# Using xpath that returns text
for idx, t in enumerate(browser.select_xpath('//div/text()')):
    print(idx, t)
# Using css selector which returns elements
for idx, elem in enumerate(browser.select_css('.result__title')):
    print(idx, elem.text)

Documentation and tutorials

• Check examples directory
• The BrowserAPI.md file has a quick list of Browser's API
• Use xpathsell -e to print available environment variables for settings
• Use xpathsell --settings-help to print settings detailed documentation
  • or optionally check xpathwebdriver/default_settings.py

IPython interactive shell

First install ipython pip install ipython (not installed to keep basic requirements low) Run the xpathshell in your terminal, and you should see something like:

$ xpathshell
Python 3.7.5rc1 (default, Oct  8 2019, 16:47:45)
Type 'copyright', 'credits' or 'license' for more information
IPython 7.9.0 -- An enhanced Interactive Python. Type '?' for help.

XpathBrowser in 'b' or 'browser' variables
 Current url: data:,
In [1]: b.get('github.com/joaduo/xpathwebdriver/')
INFO 05:53:35:  Current url: https://github.com/joaduo/xpathwebdriver/

For a faster development and debugging cycles you can run an interactive shell which will let access the browser.

Or pass the url in the command arguments. Eg: xpathshell github.com/joaduo/xpathwebdriver/

Inside IPython you can enter browser.select_xpath? to get documentation and can access API docs.

More XpathBrowser details at:

• https://github.com/joaduo/xpathwebdriver/blob/master/BrowserAPI.md
• https://github.com/joaduo/xpathwebdriver/blob/master/xpathwebdriver/xpath_browser.py
• https://github.com/joaduo/xpathwebdriver/blob/master/xpathwebdriver_tests/test_XpathBrowser.py

Using unittest library

import unittest
from xpathwebdriver.webdriver_manager import get_browser

class SearchEnginesDemo(unittest.TestCase):
    def test_duckduckgo(self):
        with get_browser() as browser:
            browser.get_url('https://duckduckgo.com/')
            browser.fill('.//*[@id="search_form_input_homepage"]', 'xpathwebdriver\n')

Check a more options in the examples directory.

Installing Selenium

Under Ubuntu you should easily install Selenium/Webdrivar with sudo apt install chromium-browser chromium-chromedriver Firefox seems to come with webdriver out of the box on ubuntu.

Use the code below to test selenium and webdriver installation:

from selenium import webdriver
import time
driver = webdriver.Chrome() #or use another backend
driver.maximize_window()
driver.get('https://www.google.com')
print('You have 10 secs to check the browser window...')
time.sleep(10)

On other platforms find the easiest way to install selenium in your environment.

Here some references:

• https://www.seleniumhq.org/download/#thirdPartyDrivers
• http://chromedriver.chromium.org/
• https://github.com/mozilla/geckodriver/releases
• PhantomJs (abandoned)

Decompressed executables should be in your PATH. If you update python's webdriver package make sure you update browsers and drivers.

Useful links for working with XPath

• Xpath cheatsheets
  • http://ricostacruz.com/cheatsheets/xpath.html
  • https://web.archive.org/web/20200218033716/http://xpath.alephzarro.com/content/cheatsheet.html
• Firefox addons
  • https://addons.mozilla.org/es/firefox/addon/firebug/
  • https://addons.mozilla.org/es/firefox/addon/firepath/

Killing processes hanging around

Depending on your configuration from virtualdisplay and browser, processes like:

Xvnc
Xvfb
Xephyr
chromedriver
...

may keep hanging around. You may want to kill them (on linux) with:

# check the wanted process is alive
ps faux | grep Xvnc
# and you can kill it. If you are running as root, make sure you are not killing someone else's process too 
pkill Xvnc

On other OS google on how to do it (on windows you can use the Task Manager)