This Python module is used to talk to Automation Anywhere 10.5.x and start Tasks on selected machines. What it does is basically call the Automation Anywhere REST API 2 times:

1. To login
2. To execute the task

Automation Anywhere does have a basic check if the task was "asked to be deployed" successfully, albeit very simple, it's still better then nothing.

There are some minimum dependencies:

pip install pyodbc>=4.0 sqlalchemy>=1.3 requests>=2.21 

After installing them, you can proceed with the installation:

pip install automation_anywhere

As is, for the version 10.5.x, the API usage is basically:

from automation_anywhere import base
import logging

logging.basicConfig(level=logging.DEBUG)

try:
    aa = base.Executor(host='localhost', protocol='http', port=8080, username='bot_user', password='bot_password')
    aa.deploy_task(task='My Tasks\\Task1.atmx', client='BOTRUNNER01')
except Exception as error:
    print(error)
    raise

The errors mainly come from the requests() that we create witht the JSON, and while there are some errors from the Automation Anywhere side, they get back mainly as "Internal Error", unless it's username/password issues.

The Automation Anywhere API (at least on version 10) does not return the task status, it only returns the deploy status. Knowing that I did a little reverse engineering on the database, and came up with some simple queries that can check the status of the task.

There are a few caveats and warnings that you should know, before using this:

1. It'll block the execution, making it a sync execution, instead of async
2. You need to configure the ODBC directly for the Control Room database
3. There'll be a lot of queries to get the correct task ID, and status
4. This simply does not works on V11
5. This works, but it wasn't tested a lot, so there may be some bugs

If you want to, here's an example on how to use it

from automation_anywhere import base
import logging

logging.basicConfig(level=logging.DEBUG)

try:
    aa = base.Executor(host='localhost', protocol='http', port=8080, username='bot_user', password='bot_password')
    aa.database_options = {
        'DSN': 'CR_ODBC_LINK',
        'username': 'username',
        'password': 'password'
    }
    aa.set_check_status(True)
    if aa.deploy_task(task='My Tasks\\Task1.atmx', client='BOTRUNNER01') is False:
        logging.info('Task deploy or execution failed!')
    logging.info('Task return information:')
    logging.info('Status  : {status}'.format(status=aa.task_status['status']))
    logging.info('Complete: {complete}'.format(complete=aa.task_status['complete']))
    logging.info('Error   : {error}'.format(error=aa.task_status['error']))
except Exception as error:
    print(error)
    raise

There's one big thing that needs to be taken account of, always: the Control Room Credential Vault must be opened. If not, the Control Room API is unable to deploy the task. The Automation Anywhere API however will return this simply as an "Internal Error", so when this happens you need to go to the Control Room's Audit Log and check for the error.

From our experience, 90% of the time the error is "Credential Vault not open". For fixing this the user simply has to login as admin on the "Control Room", so if you use an admin user to check all those things, congratulations, by checking it you already fixed it.

There is however an API call for checking if the Credential Vault is open or not, however this has proven extremely unreliable and there's no official documentation for it, so I haven't added it here. Maybe on v11 of Automation Anywhere this is executing better.

While I was developing this simple Python module, Automation Anywhere released the v11 of it's software. It has a lot of new and shiny things, however I wasn't able to test it so far. That basically means that this should work, but I'll never guarantee it, until I've tested it.

While I don't have a timeline (remember I'm pro bono here) there's a few things I want to do (in no specific order):

• Test on v11
• Increase the ease of use of this module