Skip to content

Latest commit

 

History

History
88 lines (61 loc) · 2.5 KB

README.md

File metadata and controls

88 lines (61 loc) · 2.5 KB

TalkToFiles: Conversational RAG Chatbot

TalkToFiles is a web application built using Gradio and Haystack, designed to enable users to upload documents and interact with their content through a conversational AI interface. This chatbot leverages Retrieval-Augmented Generation (RAG) to provide accurate and context-aware answers by retrieving information from document content.

App Demo Link

Features

  • Document Upload: Upload multiple files in PDF, Markdown, or Text format.
  • Document Store Creation: Automatically preprocess and store document content for querying.
  • Conversational Interface: Ask questions and get responses based on uploaded document content.
  • Memory Integration: Maintains context throughout the conversation for more natural interactions.

Tech Stack

  • Gradio: Provides the web interface for document uploads and chat functionality.
  • Haystack: Powers the RAG pipeline for document preprocessing and retrieval.
  • Cohere: Used for query rephrasing and response generation.
  • SentenceTransformers: For creating document embeddings.

Installation

  1. Clone the repository:

    git clone <repository-url>

  2. Navigate to the project directory:

    cd TalkToFiles

  3. Install dependencies:

    pip install -r requirements.txt

  4. Set up environment variables:

    • Create a .env file in the root directory.
    • Add your API key for Cohere: 
      COHERE_API_KEY=your_api_key_here

Usage

  1. Start the application:

    python main.py

  2. Upload Documents:

    • Use the interface to upload PDF, Markdown, or Text files.
    • Click the "Create Document Store" button to preprocess and store the documents.

  3. Chat with Documents:

    • Type your query in the chat interface.
    • Receive contextually accurate responses based on the document content.

File Structure

  • main.py: Contains the Gradio interface and application logic.
  • module.py: Defines the pipelines and components for preprocessing, retrieval, and query handling.

Example Workflow

  1. Upload a PDF document containing information about Artificial Intelligence.
  2. Initialize the document store.
  3. Ask the chatbot: "What is Artificial Intelligence?"
  4. Receive an accurate, document-based response.

License

This project is licensed under the MIT License. See the LICENSE file for details.