Skip to main content

Python

Python can be used to run processes in the OpenDataDSL platform.

Processes can be triggered:

  • On a schedule by setting a cron property
  • Event-driven, using an automation
  • Manually in the web portal, REST API or ODSL code

Creating a process​

There are 3 specific properties to set on a process to run it as a Node.js process:

  • image - The name of docker image to run, you can use the ODSL standard one - odslacr.azurecr.io/odsl_python:latest
  • job - This is always set to prod_python
  • commandline - The command to run, see below

Commandline​

The commandline is what the compute pool runs in the docker container, what you specify depends on how you want to run your python code.

Using odslrun.py​

odslrun.py has the following features:

  • Creates context variables that can be used in your script:
    • PROCESS_CONFIG - contains the JSON of the process configuration
    • TASK - contains the JSON of the task configuration
    • SDK - the python sdk, already connected using the process runners credentials
    • PROCESS - the process sdk class which allows you to communicate with the process execution log
  • Checks for the name of a private script as the input parameter
  • Pulls the script from the platform and applies to the current context
  • Starts the process and runs your script in a MAIN phase
Example commandline for the process​
"/bin/sh -c 'cd /app;python3 odslrun.py odsl-extensions\\extensions\\python\\python_sample'"

This will pull the python_sample script from the platform and execute it.

python_sample​

async def main():

	# Get the #ECB object
	obj = SDK.get('object', 'public', '#ECB')

	# Log the description
	await PROCESS.logMessage("INFO", obj['description'])