Workflow Context

​

Understanding Workflow Context

​

The WorkflowRunContext Class

​

Properties

• backend_token: Authentication token for making API calls to the SampleHC backend
• current_user: Information about the user who initiated the workflow (dict with id, email, and name fields). This is None when the run was not started by a user — for example, runs started by a cron schedule or an event — so guard access with a check like ctx.current_user["name"] if ctx.current_user else "Automated".
• resume_data: Data used when resuming a suspended workflow
• scope: Additional scoped data for the step that you’re on. For example, loops will have i and item in scope.
• workflow_run_id: The ID of the current workflow run. This is useful when you need to look up other tasks from the same run, for example inside a next_task routing override.

​

Key Methods

​

get_step_result(step_id: str)

# Example: Get the result from a previous step
previous_result = ctx.get_step_result("document-upload")
documents = previous_result["documents"]

​

get_start_data()

# Example: Access the initial workflow data
start_data = ctx.get_start_data()
initial_params = start_data.get("params", {})

​

Using Context in Workflow Steps

​

In Automated Steps

def process_documents(ctx):
    # Access previous step results
    uploaded_docs = ctx.get_step_result("upload-step")

    # Access user information
    user_name = ctx.current_user["name"]
    user_email = ctx.current_user["email"]

    # Process the documents
    processed_docs = []
    for doc in uploaded_docs["documents"]:
        # Your processing logic here
        processed_docs.append(process_single_doc(doc))

    return {"processed_documents": processed_docs}

workflow.then(
    Step("process-docs", process_documents)
)

​

Best Practices

1. Step ID Consistency: Use clear, descriptive step IDs that make it easy to reference them in later steps.
2. Data Validation: Ensure the output is what you expect from previous steps before using it.
3. Context Security: The context includes authentication tokens, so be careful when logging or exposing context data.