Wednesday, May 24, 2017

Updates to Chatbot API and Tooling

Build Chatbots with IBM Bluemix
If you have been coding up your own chatbot using the IBM Watson Conversation service on IBM Bluemix, then you might be aware of the new features in Conversation API. Earlier this year I wrote a small tool to manage chatbot workspaces (read here the blog about the management tool). I updated the tool to the newest API version and added basic support for one of the new API functions.

You can now invoke my "Watson Conversation Tool" with the "-logs" parameter. It retrieves recent interactions, i.e., processed messages with all the details. The logs are great to see what's going on (of course), to understand why a specific response was generated by Watson Conversation and hence to improve a chatbot.

[henrik@mymachine] python wctool.py -logs -id "123123b8-123b-1234-a656-6dxx5a2xxf15"


{
  "pagination": {},
  "logs": [
    {
      "log_id": "fba37b8c-a4a8-4ec1-a2d0-7c361d24f441",
      "response_timestamp": "2017-05-24T08:02:20.882Z",
      "request": {
        "input": {
          "text": "@db2bm Was liegt heute an, wie wird das Wetter?"
        },


...

      "request_timestamp": "2017-05-24T08:02:18.494Z"
    }
  ]
}



The API allows to filter, sort and paginate the output. If you want that added to my tool, let me know by leaving feedback or opening an issue.

No comments:

LinkWithin

Related Posts with Thumbnails