Watson Conversation API响应与Watson Web控制台上的工作区响应不一致 [英] Watson Conversation API response inconsistent with workspace response on Watson web console

问题描述

我建立了一个对话对话框模型，当在www.ibmwatsonconversation.com工作区上进行测试时，该模型可以完美地工作.

I built a conversations dialog model that works perfectly when tested on the www.ibmwatsonconversation.com workspace.

但是，当我使用API​​调用Web应用程序上的同一工作区时，通过API给出的响应是不一样的.

However, when I use the API calling the same workspace on my web app, the response given through the API not the same.

以下是流程:

意图1

意图2->实体1

意图3

预期行为:

1.询问意图2的问题，得到意图2节点的答复.

1.Ask question with intent 2, get reply from intent 2 node.

2.输入实体1，得到实体1节点的答复.

2.Enter entity 1, get reply from entity 1 node.

实际行为(仅来自API)

Actual behaviour (only from API):

1. 用意图2询问问题，得到意图2节点的答复.

1. Ask question with intent 2, get reply from intent 2 node.

输入实体1，获得意图1的回复

Enter entity 1, get reply from intent 1

推荐答案

最可能的原因是您没有在每次调用时都传递上下文对象.会话是无状态的，因此如果没有上下文对象，会话将无法确定您的位置，并且默认为root.

The most likely cause for this is that you are not passing back the context object at every call. Conversation is stateless, so without the context object it can't determine where you are, and will default to root.

您的第一个调用将创建上下文对象，您可以继续将其传递回去.

Your first call will create the context object, and you can keep passing that back.

如果不是问题所在，则需要提供带有虚拟数据的问题演示或对话框流程的屏幕截图(仅限相关部分).

If this isn't is the issue, you need to supply a demo of the issue with dummy data, or a screenshot of your dialog flow (related part only).

这篇关于Watson Conversation API响应与Watson Web控制台上的工作区响应不一致的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！