Plugin

Execute sandboxed Python code with access to the pipeline data registry.

The Plugin module executes user-defined Python code. Scripts can read data from previous modules, transform it and export new values for later blocks.

Plugin editor

Configuration

Field	Type	Description
script_id	`string`	Script name shown in the editor and Codebase
code	`string`	The Python source code to execute

The Data API

Inside a Python script, you interact with pipeline data through the Data object:

Reading Data

# Get a value from the data registry
image = Data.get("Camera_0.frame")
score = Data.get("AnomalyDetection_0.pred_score")

Writing Data

# Export a value to the data registry
Data.export("result", my_result)
Data.export("processed_image", processed)

Exported variable names must not contain dots (.). The system automatically prefixes exports with <module_id>., so an export named result from a module with ID PythonScript_0 becomes PythonScript_0.result in the registry.

Input & Output Detection

The system automatically detects inputs and outputs by parsing the script code:

Inputs: extracted from Data.get("...") and Data.get_*("...") calls
Outputs: extracted from Data.export("...") and Data.export_*("...") calls

This allows the validation system to verify that all dependencies are satisfied before execution.

Console Output

Any print() output from the script is forwarded to the action plan console panel. Errors are displayed with an error indicator.

Compliance Override

Scripts can override the pipeline’s compliance value by exporting a special key:

# Override COMPLIANT=0, UNCERTAIN=1, DEFECTED=2
Data.set_compliance(COMPLIANT)

Inputs & Outputs

Direction	Key	Type	Description
Input	Any key via `Data.get(...)`	varies	Data from previous modules
Output	`<module_id>.<exported_name>`	varies	Data exported via `Data.export(...)`

Use Save to codebase to keep a reusable script. Existing Codebase entries can be loaded, overwritten or deleted from the editor.

Example Script

# Read the anomaly prediction score
score = Data.get("anomaly_0.pred_score")

# Apply custom business logic
if score < 0.3:
    verdict = "PASS"
    compliance = 0
elif score < 0.7:
    verdict = "REVIEW"
    compliance = 1
else:
    verdict = "FAIL"
    compliance = 2

# Export results
Data.export("verdict", verdict)
Data.set_compliance(compliance)

print(f"Verdict: {verdict} (score: {score})")