This is Henri Augusto's submission for the Thoughtful Automation's fresh-news challenge.
This sections descibres aspects that different when executing the robot in Robocloud's Cloud Workers vs Self-Hosted Workers/Locally.
At first I've had success running the robot i a couple of Self-Hosted workers but couldn't manage to run the robot in any Cloud Worker.
I would get errors such as:
Message: unknown error: Chrome failed to start: exited abnormally.
2023-08-31 19:34:08: (unknown error: DevToolsActivePort file doesn't exist)
2023-08-31 19:34:08: (The process started from chrome location /home/worker/.cache/selenium/chrome/linux64/116.0.5845.96/chrome is no longer running, so ChromeDriver is assuming that Chrome has crashed.)
I've pinpointed the error to RPA.Browser.Selenium's open_browser() method.
After a lot of trial and error, experimenting with different parameters for open_browser() and also different chromedriver versions in conda.yaml, I've managed to get the robot running in the Cloud Workers by exchanging open_browser() for open_chrome_browser() in BrowserInitializer.
Aftwerwards i've start receiving errors like this:
'utf-8' codec can't decode byte 0x92 in position 292: invalid start byte
Which were solved by specifying which encoding to use when reading and writing to the CSV in the DataManager class. The offending characters were and U+2018 (‘) and U+2019 (’)
This robot would run sucessfully on
- Cloud Workers (Preview)
- Cloud Workers (Ubuntu 18.04 Legacy)
but not on
- Cloud Workers
- Cloud Workers (early access)
Cecause in those 2 the following error appears while interacting with the search button:
element not interactable
Further investigation is required.
I've already tried adding python-chromedriver-binary=114.0.5735.16.0 to conda.yaml in hope it's a versioning issue but with no success.
Open the command palette and run Robocorp: Debug Robot. It should give you the following options:
1. <No Work item as input>
2. run-1 - Output
Choose the second option so it runs with the preconfigured test itens. (devdata\work-items-out\run-1\work-items.json)
All results in the CSV file are written and loaded by the DataManager class.
Our main task checks for every results if it was previously written to the file before writing, i.e. ignores any result processed in previous executions
A Log class is provided which acts as an interface. Changing, removing or adding logging strategies can be done without changing the client code.
- Locators that are only used in one place are defined in the function they're used.
- Locators used more than one place are defined as class variables.