Web scraper Python robot example

It is possible to implement robots using pure Python, without using Robot Framework. These robots are first-class citizens in Robocorp: they can be developed iteratively with our developer tools, and run and orchestrated in Robocorp Cloud, like any other robot. RPA Framework, our set of open-source libraries, provides APIs for both Robot Framework and Python.

This simple robot opens a web page, searches for a term, and takes a screenshot of the web page using the RPA.Browser.Selenium library.

Robot script (Python)

from RPA.Browser.Selenium import Selenium

browser = Selenium()
url = "https://robocorp.com/docs/"
term = "python"
screenshot_filename = "output/screenshot.png"


def open_the_website(url: str):
    browser.open_available_browser(url)


def search_for(term: str):
    input_field = "css:input"
    browser.input_text(input_field, term)
    browser.press_keys(input_field, "ENTER")


def store_screenshot(filename: str):
    browser.screenshot(filename=filename)

# Define a main() function that calls the other functions in order:
def main():
    try:
        open_the_website(url)
        search_for(term)
        store_screenshot(screenshot_filename)
    finally:
        browser.close_all_browsers()

# Call the main() function, checking that we are running as a stand-alone script:
if __name__ == "__main__":
    main()

• Import the Selenium library from the RPA.Browser.Selenium package.
• Initialize the browser and some variables.
• Define the functions that implement the operations the robot is supposed to do.
• Define the main function.
• Call the main function.

Important: Always use try and finally to ensure that resources such as open browsers are closed even if some of the functions fail. This is similar to using [Teardown] in Robot Framework scripts. Read the Errors and Exceptions document on python.org for more information.

robot.yaml

The robot.yaml file for the Python robot looks like this:

tasks:
  entrypoint:
    command:
      - python
      - tasks.py

condaConfigFile: conda.yaml
artifactsDir: output
ignoreFiles:
  - .gitignore
PATH:
  - .
PYTHONPATH:
  - .

Note the command section:

command:
  - python
  - tasks.py

In this case, we tell Python to run the tasks.py file. You can customize the command for your needs.

For comparison, the typical Robot Framework command looks like this:

command:
  - python
  - -m
  - robot
  - --report
  - NONE
  - -d
  - output
  - --logtitle
  - Task log
  - tasks.robot

In this case, we tell Python to run the robot module (python -m robot) with arguments.