Links
Comment on page

Configuration

This section explains each of the properties in the configuration blade.

General

Property
Description
Collection
This associates the Agent to a specific Collection (default to that of the current Data Stream).
Refer to Collections and Stream Hosts to understand more about Collections.

Details

Property
Description
Use Environment Variables
Tick to use the system Environment Variables on the host system for lookup paths.
User Server Variables
Tick to select saved variables for the Python Home, Python Path, and Python DLL Path, or enter the text manually (applies when Use Environment Variables is unticked).
Accelerator
Select whether the script will be executed on the CPU Only or GPU Accelerated (GPU option available when this capability is detected).
Python Home
The location of the standard Python libraries (applies when Use Environment Variables is not ticked). Eg. C:\Python37
Python Path
The location of the Python site packages folder (applies to CPU Only Accelerator only when Use Environment Variables is not ticked).
Eg. C:\Python37\Lib\site-packages
Python DLL Path
Path to python**.dll (e.g. python38.dll) for Windows hosts or libpython**.so (e.g. libpython3.8.so) for Ubuntu hosts (applies to CPU Only Accelerator only when Use Environment Variables is not ticked).
Python File
The Python script file to be executed.
Input Mapping
Input variables from the Python file are auto-populated when it is loaded. Each input variable is mapped to a column or attribute from the input payload.
Outputs
One or more Python variable names from the Python script file. These are appended to the output payload. When Accelerator is set to GPU Accelerated, the script should print the output in the following format: print({VariableName}:{VariableValue})
Columns Returned
The output values can either be appended to the incoming payload (Append to Existing), or a new payload is created that consists only of the output values (New).
Globals Grid
Optional global variables defined in the Python file that are available when next the Data Stream is published (applies to CPU Only Accelerator only). The value is saved to a Pickle file whenever the Data Stream is unpublished, which allows you to maintain the program state across sessions, e.g. to recall the previous high temperature without waiting on sensor input or a recalculation.
Before configuring the AI & ML Agent, please ensure that its input endpoint is connected to a parent Agent which will be sending data to it.

Endpoints

Name
Description
Input
This endpoint is used to receive data from the parent Agent.
Output
Events received from the parent Agent are made available to this endpoint in the format configured in Outputs and Columns Returned.
Error
Errors encountered during the execution of this Agent will be published on this endpoint.