Skip to main content

Creating Domain Assistant + HTTPClient + CustomDatasource

Initial Assistant Setup​

  • Created a Virtual Assistant called "Movie Assistant"
  • Configured the context: answer questions about movies
  • Instructed it to always use emojis in responses

Integration with the IMDB API​

Steps Taken:​

  1. Created an HTTP Client to query the IMDB API
  2. Used Key Vault to protect the access token
  3. Configured a Custom Data Source to populate a table with popular movies

Features Implemented:​

  • Query for the list of most popular movies
  • Created a dataset called "all_movies"
  • Groundwork to restrict the bot's responses based on the IMDB database
info

The Dataset's default pagination is 20, but make sure to also pass pagination to your HTTPClient request via CustomHttpDataSource. Otherwise, the CustomHttpDataSource will keep bringing in all records from the HTTPClient request.

Final Goal​

Create a movie assistant that:

  • Answers specific questions about movies
  • Uses a reliable database (IMDB)
  • Avoids incorrect or made-up information

Next Steps​

  • Configure the user's interaction with the assistant
  • Implement searching and getting details for specific movies