如何在 Botbuilder v4 框架中使用 Luis Intent 而无需从 LUIS 下载 json 文件

Question

I am using the below code:

https://github.com/microsoft/BotBuilder-Samples/tree/master/experimental/adaptive-dialog/csharp_do.netcore/

I have loaded csharp_do.netcore.sln in visual studio file and able to see core-bot todo-bot,using-cards and multi-turn-prompt. I am able to build the project. It is mentioned in README that I have to download the json file from LUIS and put in specific location.However, even after downloading the file and putting in the location, intents are not getting recognized.

Please help me with adding a LUIS intent in this project.

And also where to do the necessary changes to identify the intents provided in chat bot. Will it be in RootDialog.cs? I have also put the LUIS app id and key and hostname in CoreBot.luis.json.

In the attached screenshot,sample output can be seen. Where it is only producing the output from CoreBot.luis.json file and its kind of a static output, even though after changing the json file content i am getting the same response in chatbot and i also want it to detect the output directly from LUIS portal

Please help.

Answer 1

Ok, your issue here is sheer confusion. You list 3 separate projects that you're saying aren't working.

Both the CoreBot and the ToDoBot require LUIS. You have to add the LUIS model on your LUIS.ai portal using either the CLI or manually through the portal upload:

Once you have your LUIS app created, you can use the appID and Password you get from the LUIS portal to add to the appsettings.json file:

In the question, you say you're using CoreBot, in the first comment, you say you're using Cards Bot, and in a later comment you say you're using ToDo Bot. Those are three different bots in the solution, so you're probably putting your keys in the wrong appsetttings.json file. Double check that you're using the right appsettings.json file.

Also to note: the bots in the experimental branch are meant to see different possible approaches that the Bot Framework might take going forward, and should not be used in production environments. They are not guaranteed to work, and are not supported.