Selenium Python tutorial
1. 🤖 How it works
Applitools SDKs work with existing test frameworks to take screenshots of pages, elements, regions or iframes and upload them along with DOM snapshots to our Eyes server. Our AI then compares them with previous test executions' screenshots (aka Baselines) and reports if there is a bug or not. It's that simple!

1.1 Baseline vs. Checkpoint images
When you first run the test, our A.I. server stores those first set of screenshots as Baseline images. When you run the same test again (and everytime there after), the A.I. server compares the new set of screenshots, aka Checkpoint images, with the corresponding Baseline images and highlights differences in a pink color.

1.2 Marking the test as "Pass" or "Fail"
When the AI compares the baseline and the checkpoint image, if it finds a legitimate difference, it will mark the test as Unresolved. This is because the AI doesn't know if the difference is because of a new feature or a real bug and will wait for you to manually mark it as a Pass/Fail for the 1st time.
If you mark the Unresolved checkpoint image as "Failed", it'll only mark the current test result as Failed.

Note:
To automatically mark the checkpoint as a "Fail" in the future test runs, you need to do the following:
- Annotate at least one of differences as a "bug region"
- Select the "Fail tests" checkbox in the popup window
- Press "Thumbs Up" (not "Thumbs Down") button in the checkpoint image's toolbar (Note: this is counter-intuitive. But what happens is that, we now create a new Baseline along with this bug and "Failed" metadata. So if the same image with the exact bug appears, it'll fail again)
- Press "Save" in the main toolbar
If you mark the Unresolved checkpoint image as a "Pass", then it means that the difference is due to a new feature so we set the new checkpoint image as the new baseline and mark the current test as Pass. Going forward we'll compare any future tests with this new baseline.

Note:
Applitools AI has been trained with 100s of millions of images. It doesn't do pixel-to-pixel comparison as this can lead to a lot of false positives. It instead simulates human eyes that ignore differences that humans can't detect and highlight differences that humans can detect.
ACCURACY: Our A.I.'s current accuracy rate is 99.9999%! Which means for most applications the odds that you'll see false-positives are 1 in a million!
A powerful test results dashboard
We provide a state-of-the-art dashboard that makes it very easy for you to analyze differences, report bugs and much more. For more information on the Applitools dashboard check out these articles.

2. 🖼 Analyzing differences
The following Gifs show various tools Applitools provides to easily analyze various differences
Highlight differences between the baseline and checkpoint

Zoom into differences

Toggle between baseline and checkpoint

Show both the baseline and checkpoint side-by-side

3. 🐞 Reporting bugs (straight into Jira or Github)
You can select a section of the image and directly file a bug in Jira or Github. No need to manually take screenshots, write steps and explain things! To read more about bug regions check out this article.

4. ✅ Prerequisites
Create a free Applitools account and get the Applitools API KEY
Please set the
APPLITOOLS_API_KEY
environment variableMac:
export APPLITOOLS_API_KEY='YOUR_API_KEY'
Windows:
set APPLITOOLS_API_KEY='YOUR_API_KEY'
TIP
- You may skip this step if you want to hard code the API KEY inside the tutorial project's code.
- It's better to store APPLITOOLS_API_KEY in the system variables (in Windows) or in the
~/.bash_profile
(in Mac) so that it is accessible from all Terminal shells
Install Python 3. This is a good resource
Install git from https://git-scm.com
TIP
Installing
git
is optional. You need this mainly to clone the demo project from the Github repository. Instead of installinggit
, you can simply download the Zip file from the repo. Further, If you are Mac, you already havegit
.Install Google Chrome browser from https://www.google.com/chrome/
Install ChromeDriver on your machine and make sure it's is in the
PATH
.
Below are some resources from the internet that'll help you
Tip: On Mac, place the chromedriver
executable in /usr/local/bin
folder so Eclipse and IntelliJ can find it.
5. 🚀 Run Tests
There are two main ways in which you can run tests.
The Applitools Ultrafast Grid (option 1) Test across all browsers and viewports 30-70 times faster than traditional test clouds. Using this approach, your tests will execute on a local browser once with the DOM, CSS, and additional artifacts captured by the SDK and sent to the Ultrafast Grid for parallel rendering and Visual AI comparison. This is the preferred execution method, however please note that it is a premium feature.
The second approach is to Run Local (option 2). This is the classic test execution in which your tests execute locally and each call to eyes via the SDK will send a screenshot to Eyes for Visual AI comparison.
While running tests using the classic approach is a great start - we recommend running tests on the Ultrafast Grid, the fastest way to ensure quality and consistency across all browsers, viewports, and devices.
Option 1 - Run With The Ultrafast Grid
1. Introduction To the Ultrafast Grid
In real world testing, you'll need to test your app on multiple browsers, viewports and mobile devices. So you'll need to run and re-run ALL the tests in different browsers and multiple viewports to ensure things are fine - which could take hours and hours to complete. You may also encounter browsers not opening properly, or hanging and what not.
What if you could run in just one browser, just once, and still do cross-browser tests across various browsers, viewports and multiple mobile emulators? That's where Ultrafast Grid comes in.
TL;DR
When you execute this tutorial, it'll run the test locally in Chrome browser but will show results in the following 5 combinations:
- Chrome browser (800 X 600 viewport)
- Chrome browser (700 X 500 viewport)
- Firefox browser (1200 X 800 viewport)
- Firefox browser (1600 X 1200 viewport)
- An iPhone4 emulator
... all in ~30 seconds!⏰🚀
2. What Is The Ultrafast Grid?
The Ultrafast Grid provides an elegant and an ultra fast way to render pages in different browsers, viewports and mobile simulators and then take screenshots and then send those screenshots to Applitools AI server. With Ultrafast Grid you just need to run your test in a single browser but perform end-to-end and cross-browser tests in various browsers such as Chrome,Firefox, IE, and also numerous combinations of viewports and in mobile simulators -- all in under a minute. Without Ultrafast Grid, you'll be wasting a lot of time and money running those tests in multiple browsers which could take hours and are error-prone.
And since Ultrafast Grid is integrated with our Eyes AI, you'll actually find a lot more functional and visual bugs than you'd normally find.
Further, you can use our Root-Cause-Analysis (RCA) feature to figure out exactly which DOM element or CSS caused difference! So you can go backwards, from the screenshot to the code!
Best part: All you need to do is to simply make eyes.check
calls in your existing Selenium tests after adding our SDK and instantiating Ultrafast Grid. And that's all there is to it!
How it works
Once you add and instantiate the Applitools SDK, simply add eyes.check
calls wherever you want to check if the state of the page looks OK. Then run the Selenium test as usual. And while running the tests, anytime it encounters eyes.check
(or similar methods), Applitools SDK uploads the DOM resource of that page to our Ultrafast Grid behind the scenes and in batches. The Ultrafast Grid then renders every page in different browsers and mobile emulators (in parallel) and takes screenshots as per the configuration you set. Finally, it sends those screenshots to our AI Server.
When you first run the test, our AI server simply stores those 1st set of screenshots as Baseline images. When you run the same test again, the AI server will compare the new set of screenshots, aka Checkpoint images, with the corresponding Baseline images and highlights bugs.

3. Getting Started With The Ultrafast Grid
Pre-requisites:
- Python 3 is installed on your machine. Install Python 3.
- Package manager pip is installed on your machine. Install pip
- Chrome browser is installed on your machine. Install Chrome browser
- Chrome Webdriver is on your machine and is in the environment variable PATH. Here are some resources from the internet that'll help you.
- Git is installed on your machine. Install git
- If you want to run example from IDE, install any IDE for Python (e.g. PyCharm )
Run the example
Start with the ready-to-run code Github repo
- Git clone this repo
git clone https://github.com/applitools/tutorial-selenium-python-ultrafastgrid.git
, or download this as a Zip file and unzip it cd tutorial-selenium-python-ultrafastgrid
- Open in any editor file
ultrafastgrid_demo.py
and set your ApiKey in string.set_api_key('...')
- Install requirements
pip install -r requirements.txt
- Run
ultrafastgrid_demo.py
by callingpython ultrafastgrid_demo.py
- If you want run from IDE - start PyCharm, open just cloned project, set project interpreter by File > Settings > Project: > Project Interpreter choose interpreter by dropdown box; tap Run and choose
ultrafastgrid_demo
.
Sample code
from webdriver_manager.chrome import ChromeDriverManager
from selenium.webdriver import Chrome
import os
from applitools.selenium import (
logger,
VisualGridRunner,
Eyes,
Target,
BatchInfo,
BrowserType,
DeviceName,
)
def set_up(eyes):
# You can get your api key from the Applitools dashboard
eyes.configure.set_api_key("APPLITOOLS_API_KEY")
# create a new batch info instance and set it to the configuration
eyes.configure.set_batch(BatchInfo("Ultrafast Batch"))
# Add browsers with different viewports
# Add mobile emulation devices in Portrait mode
(
eyes.configure.add_browser(800, 600, BrowserType.CHROME)
.add_browser(700, 500, BrowserType.FIREFOX)
.add_browser(1600, 1200, BrowserType.IE_11)
.add_browser(1024, 768, BrowserType.EDGE_CHROMIUM)
.add_browser(800, 600, BrowserType.SAFARI)
.add_device_emulation(DeviceName.iPhone_X)
.add_device_emulation(DeviceName.Pixel_2)
)
def ultra_fast_test(web_driver, eyes):
try:
# Navigate to the url we want to test
web_driver.get("https://demo.applitools.com")
# Call Open on eyes to initialize a test session
eyes.open(
web_driver, "Demo App", "Ultrafast grid demo", {"width": 800, "height": 600}
)
# check the login page with fluent api, see more info here
# https://applitools.com/docs/topics/sdk/the-eyes-sdk-check-fluent-api.html
eyes.check("", Target.window().fully().with_name("Login page"))
web_driver.find_element_by_id("log-in").click()
# Check the app page
eyes.check("", Target.window().fully().with_name("App page"))
# Call Close on eyes to let the server know it should display the results
eyes.close_async()
except Exception as e:
eyes.abort_async()
print(e)
def tear_down(web_driver, runner):
# Close the browser
web_driver.quit()
# we pass false to this method to suppress the exception that is thrown if we
# find visual differences
all_test_results = runner.get_all_test_results(False)
print(all_test_results)
# Create a new chrome web driver
web_driver = Chrome(ChromeDriverManager().install())
# Create a runner with concurrency of 1
runner = VisualGridRunner(1)
# Create Eyes object with the runner, meaning it'll be a Visual Grid eyes.
eyes = Eyes(runner)
set_up(eyes)
try:
# ⭐️ Note to see visual bugs, run the test using the above URL for the 1st run.
# but then change the above URL to https://demo.applitools.com/index_v2.html
# (for the 2nd run)
ultra_fast_test(web_driver, eyes)
finally:
tear_down(web_driver, runner)
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
Option 2 - Run Locally
Running the Example Project
- Clone or download the repository and navigate to that folder
git clone https://github.com/applitools/tutorial-selenium-python.git
cd tutorial-selenium-python
2
Note: you can alternatively download the project as a Zip file and extract it
- Install the dependencies
pip3 install -r requirements.txt
- Run the example test
APPLITOOLS_API_KEY="[Your API Key]" pytest tutorial.py
After you run one set of tests, you now have the baseline. Run the same test but with this URL: https://demo.applitools.com/index_v2.html
. This version of the demo app has some visual bugs so you can see how it all works. And the second set of screenshots are called "Checkpoint images".
Sample test
import os
import pytest
from selenium import webdriver
from applitools.selenium import Eyes, Target, BatchInfo, ClassicRunner
from webdriver_manager.chrome import ChromeDriverManager
@pytest.fixture(scope="module")
def batch_info():
"""
Use one BatchInfo for all tests inside module
"""
return BatchInfo("Some general Test cases name")
@pytest.fixture(name="driver", scope="function")
def driver_setup():
"""
New browser instance per test and quite.
"""
driver = webdriver.Chrome(ChromeDriverManager().install())
yield driver
# Close the browser.
driver.quit()
@pytest.fixture(name="runner", scope="session")
def runner_setup():
"""
One test runner for all tests. Print test results in the end of execution.
"""
runner = ClassicRunner()
yield runner
all_test_results = runner.get_all_test_results()
print(all_test_results)
@pytest.fixture(name="eyes", scope="function")
def eyes_setup(runner, batch_info):
"""
Basic Eyes setup. It'll abort test if wasn't closed properly.
"""
eyes = Eyes(runner)
# Initialize the eyes SDK and set your private API key.
eyes.api_key = os.environ["APPLITOOLS_API_KEY"]
eyes.configure.batch = batch_info
yield eyes
# If the test was aborted before eyes.close was called, ends the test as aborted.
eyes.abort_if_not_closed()
def test_tutorial(eyes, driver):
# Start the test and set the browser's viewport size to 800x600.
eyes.open(driver, "Test app", "First test", {"width": 800, "height": 600})
# Navigate the browser to the "hello world!" web-site.
driver.get("https://demo.applitools.com")
# Visual checkpoint #1.
eyes.check("Login Window test", Target.window())
# End the test.
eyes.close(False)
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
6. ⚙️ Troubleshooting Common Issues
Forgetting to set your API key (or getting 401 exception).
Applitools is not opening the app (appears to crash).
- In order to make sure the screenshots are consistent across different tests runs. Applitools uses Viewport as an input. A viewport is simply the size of the screenshot you are taking (width and height). Applitools tries to resize the browser to that size before taking the screenshot. If you are running the app on a small laptop and/or have set the Viewport much larger than the laptop’s screen, then you’ll see this error.
- Solution: Reduce the height or the width. For more see here.
Not properly loading the API key from the environment variable into your IDE (like Eclipse).
- After setting the APPLITOOLS_API_KEY in the environment variable to hold your Applitools API key, open the IDE from the command line terminal (and not from the IDE directly). On Mac, it’d look like this: Open a Terminal and then type:
open ~/Applications/Eclipse.app
. This will load Eclipse with all the environment variables.
- After setting the APPLITOOLS_API_KEY in the environment variable to hold your Applitools API key, open the IDE from the command line terminal (and not from the IDE directly). On Mac, it’d look like this: Open a Terminal and then type:
Trying to run the test but nothing happens
- Check if the APPLITOOLS_API_KEY is null
Incompatibility between Chrome browser and Chrome driver.
- Make sure for your version of Google Chrome, you are using the corresponding Chrome driver
Debug logs
- See this article to enable debug logs to help file support ticket
Resources
Terms & Conditions Privacy Policy GDPR© 2020 Applitools. All rights reserved.