Scuffed version of RAG (Retrieval-Augmented Generation) Helps the AI to get data from a database

How it works

1. Create our data base
2. Build our index of the data base
3. Search that index for relevant information/keywords with our query

Methods

• RAG_utan.buildIndex(DataBase: { string }): SearchDatabases

  • Builds a searchable index from an array of text strings.

• RAG_utan.Search(QueryText: string, Databases: {[number]: SearchDatabases}, MaxSnippets: number, FuzzThreshold: number): string

  • Searches the indexed database for documents matching the query.

  • Parameters

    • QueryText
      • The search string used to query the database. This is the input text you're trying to find relevant information for
    • Databases
      • A table containing all the necessary data structures for the search operation from buildIndex
    • MaxSnippets
      • Maximum amount of snippets the search can return
    • FuzzThreshold
      • A numeric value that determines how tolerant the search is to differences between the query and document content.
      • A lower value means stricter matching (fewer differences allowed), while a higher value allows for more fuzzy (approximate) matches.