How to implement streaming response when a Dify agent invokes a workflow?

[Aoaoao7]

Self Checks

• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit report (我已阅读并同意 Language Policy).
• [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:）
• Please do not modify this template :) and fill in all the required fields.

Provide a description of requested docs changes

I want my workflow conversations to return data in a stream when invoked.

[dosubot[bot]]

@Aoaoao7 你好! 我是Dosu，一个可以帮助解决 bug、回答问题并帮助你成为贡献者的机器人。我会在此等待人类维护者，有什么我可以帮助你的吗？

To implement a streaming response when a Dify agent invokes a workflow, you can use the invoke method of the workflow interface with the response_mode set to "streaming". This will return a generator that yields data as it becomes available. Here is an example implementation:

import json
from typing import Mapping
from werkzeug import Request, Response
from dify_plugin import Endpoint

class Duck(Endpoint):
    def _invoke(self, r: Request, values: Mapping, settings: Mapping) -> Response:
        """
        Invoke the endpoint with the given request.
        """
        app_id = values["app_id"]
        def generator():
            response = self.session.app.workflow.invoke(
                app_id=app_id, inputs={}, response_mode="streaming", files=[]
            )
            for data in response:
                yield f"{json.dumps(data)} <br>"
        return Response(generator(), status=200, content_type="text/html")

In this example, the generator function is used to yield each piece of data from the streaming response, which is then returned as an HTTP response with a content type of text/html ^[1].

To continue talking to Dosu, mention @dosu.

---

^{Help Dosu learn! Give it feedback: Great Response | Irrelevant Answer | Incorrect Sources | Too Verbose | Hallucination | Other}