RAG Proyecto resumidor de artículos con AI de OpenAI

Structure

The structure of the RAG is relatively simple:

• Installation of libraries and import of installed libraries.

• Loading OpenAI API keys. For security reasons, I am not including the OpenAI API key. To run it, you will need to generate an API key and insert it.

• Load a database with text without embeddings; the CSV database includes: 10 papers about of the Ukranian - Russia War.

• Generate embeddings with the text. Embeddings are numerical representations that capture the similarities and relationships between concepts.

• Define specific functions to perform the search, define the response message, and respond.

• Perform the search.

Background

This script is the fourth iteration of the RAG. Each of the previous RAGs had unacceptable errors. T-1.

Future Adjustments

There are three main improvements I need to make:

• Generate a graphical interface, probably using Flask.
• Expand the database.
• Modify the model to respond using other AIs.

Usage Instructions

• Load the database included here and then run the attached script. IT IS IMPORTANT TO INCLUDE YOUR OPENAI API KEY; OTHERWISE, IT WILL NOT WORK.