ss-pyworkflow 0.1.2

This is for workflow tracing on python environment.

My Dear Colleagues

This repository is for QA tools(workflows), you can use this on your python projects.

SS-PyWorkflow

This project is designed to simplify and standardize workflow execution using decorators to automate task execution and result tracking. It is primarily used for llm-graph, but it can be extended to other projects as well.

---

🚀 Installation

pip install ss-pyworkflow

Environment Variables

To configure the Workflow Monitoring Service, please add the following environment variables to your .env file or your environment settings:

Variable Name	Description	Example Value
WORKFLOW_MONITORING_URL	The endpoint URL for sending data.	https://example.com/api/v1/log
WORKFLOW_MONITORING_API_KEY	The API key for authentication.	your-secret-api-key

Example:

WORKFLOW_MONITORING_URL=https://example.com/api/v1/log
WORKFLOW_MONITORING_API_KEY=your-secret-api-key

📖 Usage

SS-PyWorkflow provides two main modes for managing workflows:

1️⃣ Single Function Mode

When you only need to track a single function, you can use the @workflow_entry decorator:

from ss_pyworkflow import workflow_entry

@workflow_entry(name="llm-graph")
async def _run_multigraph(**kwargs):
    # Your logic here...
    ...

After the function execution is completed, the result of this node will automatically be sent to the specified workflow backend and trigger the workflow end signal.

2️⃣ Chained Function Mode

If your workflow consists of multiple functions that execute sequentially, and each function's output is the next function's input, use the @workflow_lifecycle decorator:

from ss_pyworkflow import workflow_lifecycle

@workflow_lifecycle()
async def aexecute(self, state: AgentStateT) -> AgentStateT:
    # Each function's output will be passed to the next function
    ...

In this mode, after each function execution, data is automatically sent to the workflow and visualized as step-by-step nodes in the backend.

⚙️ Kwargs Explanation

Both @workflow_entry and @workflow_lifecycle require workflow_trace_data to be passed in kwargs for proper tracking and data transmission:

"workflow_trace_data": {
    "enable": true,
    "traceId": "592225192fb1ac17022e80c85fb8a749",
    "prevSpanId": "3f7decc34dc8d03a",
    "userId": "202505125600020250512145500563QSLVDGS96F",
    "projectId": "mtr-kiosk-pquzibd",
    "componentName": "llm-graph",
    "post_url": "https://dev.setsailapi.com/workflow/v1/event",
    "log_enabled": true
}

Parameter Description:

Parameter Name	Type	Description
enable	bool	Whether to enable the workflow. If False, no operations will be performed.
traceId	str	A unique identifier for the entire workflow execution (32-character hexadecimal format).
prevSpanId	str	The parent span ID (16-character hexadecimal format). Can be null for root spans.
userId	str	The user ID that triggers this workflow.
projectId	str	The project ID associated with this workflow.
componentName	str	Specifies the name of the workflow (e.g., llm-graph).
post_url	str	The backend API URL to send workflow data.
log_enabled	bool	Whether to enable logging.

📌 Examples

Single Function Mode

from ss_pyworkflow import workflow_entry

@workflow_entry(name="llm-graph")
async def run_task(**kwargs):
    print("Executing task...")

Chained Function Mode

from ss_pyworkflow import workflow_lifecycle

@workflow_lifecycle()
async def task_one(state):
    print("Task One")
    return state

@workflow_lifecycle()
async def task_two(state):
    print("Task Two")
    return state

🌐 Extension Support

Currently, it is mainly focused on llm-graph, but it can be extended to other workflow scenarios by modifying the decorators or adding new ones.

License

This project is licensed under the Apache License 2.0.
Copyright © 2025 Set Sail Venture Limited