rag-lite-app / project /tutorial /tutorial.md
Barry Quinn
Initial commit: RAG-lite PDF Query App
2207f46
|
Raw History Blame Contribute Delete
1.44 kB

A newer version of the Gradio SDK is available: 6.31.0

Upgrade

RAG-lite PDF Query App Tutorial

This tutorial will guide you through using the RAG-lite PDF Query App to extract insights from your PDF documents.

Getting Started

  1. Upload a Document

    • Click the "Upload PDF" button
    • Select any PDF document from your computer
    • The app will process the document and create searchable embeddings
  2. Query Your Document

    • Type your question in the query box
    • The app will return the most relevant passages from your document
    • You'll receive 3 most relevant text chunks as results

Sample Documents

In the sample_docs folder, you'll find two example documents:

  • financial_report.pdf: A sample financial report
  • trading_document.pdf: A sample trading strategy document

Example Queries

Try these example queries with the sample documents:

For financial_report.pdf:

  • "What were the key financial metrics for Q4?"
  • "Describe the revenue growth trends"

For trading_document.pdf:

  • "What are the main trading strategies discussed?"
  • "Explain the risk management approach"

Tips for Better Results

  1. Be specific in your queries
  2. Use complete sentences for better semantic matching
  3. Try reformulating your question if the initial results aren't satisfactory

Limitations

  • The app processes text in chunks of 500 characters
  • Only text content is processed (images and tables are ignored)
  • The app returns the 3 most relevant chunks per query