Before using QnA Maker, you need a series of questions and answers. You can either point it at a website and have it parse the questions and answers or upload a TSV. For this recipe, we will use a TSV. There is a sample in the Git repo for this book.
To create a QnA Maker project, go to the QnA Maker portal at https://www.qnamaker.ai/ and click on Create a Knowledge Base. It will take you through a five-step wizard to create a QnA bot. The first step deploys the resources in Azure. The second step has you choose a language and associate the bot with the new service you just created. You will then give your project a name and upload the files with questions and answers.
The most straightforward way of adding questions and answers is to use a TSV. There are a few fields you will need. These are question, answer, source, and meta. meta and source are fields you can use to query data. For example, in our nutrition FAQ, we may have several different ways of understanding...