The Story:

One of the approaches to solve captchas, like Google ReCaptcha, is to try to imitate the human mouse actions: movements, hovering and clicks.

Some users reported that making mouse moves as B-spline curves worked for them.

The Question:

How to move the mouse to a particular element following the B-spline trajectory via Selenium?


Note that the regular browser.actions().mouseMove(elm).perform(); would "jump" to the element straight and far too quickly. My understanding is that it is a matter of slowing down the movement speed, "jumping" from point to point smoothly following the mathematical model for the B-spline trajectory.

We are using Protractor/JavaScript, but the question is really language-agnostic.Note that I'm not trying to solve the captcha, or contribute to the "captcha-solving making new evil bots violating terms of use here and there" space. I'm just curious and eager to obtain more skills in the test automation space.


You can use scipy.interpolate to interpolate B-spline curves like you can see in this question.

Here I'll use one of the B-spline examples to get values to x and y:

import numpy as np
import scipy.interpolate as si

# Curve base:
points = [[0, 0], [0, 2], [2, 3], [4, 0], [6, 3], [8, 2], [8, 0]];
points = np.array(points)

x = points[:,0]
y = points[:,1]


t = range(len(points))
ipl_t = np.linspace(0.0, len(points) - 1, 100)

x_tup = si.splrep(t, x, k=3)
y_tup = si.splrep(t, y, k=3)

x_list = list(x_tup)
xl = x.tolist()
x_list[1] = xl + [0.0, 0.0, 0.0, 0.0]

y_list = list(y_tup)
yl = y.tolist()
y_list[1] = yl + [0.0, 0.0, 0.0, 0.0]

x_i = si.splev(ipl_t, x_list) # x interpolate values
y_i = si.splev(ipl_t, y_list) # y interpolate values

With values of x and y, you can move the mouse cursor with ActionChains:

from selenium import webdriver
from selenium.webdriver.common.action_chains import ActionChains

url = "https://codepen.io/falldowngoboone/pen/PwzPYv"
driver = webdriver.Chrome(executable_path="/home/selenium/chromedriver2.25")
driver.get(url)

action =  ActionChains(driver);

startElement = driver.find_element_by_id('drawer')

# First, go to your start point or Element:
action.move_to_element(startElement);
action.perform();

for mouse_x, mouse_y in zip(x_i, y_i):
    action.move_by_offset(mouse_x,mouse_y);
    action.perform();
    print(mouse_x, mouse_y)

If you were running this from desktop wand wanted to use an actual mouse movement, with AutoIt you can make mouse movements delayed.


@ODIUM @Guilherme , or anyone still looking for a fix. What ODIUM described in Guilherme's answer as a jump to the first curve position then back to the beginning, then second curve position and back to the beginning again is caused by a small bug in the provided code. It will be fixed by "resetting" the action chain after each perform, like this:

action =  ActionChains(driver);
startElement = driver.find_element_by_id('drawer')

# First, go to your start point or Element:
action.move_to_element(startElement);
action.perform();

for mouse_x, mouse_y in zip(x_i, y_i):
    # Here you should reset the ActionChain and the 'jump' wont happen:
    action =  ActionChains(driver)
    action.move_by_offset(mouse_x,mouse_y);
    action.perform();
    print(mouse_x, mouse_y)