ChatWithPDFs

Overview

ChatWithPDFs is an interactive Streamlit web application that allows users to engage in a conversation with multiple PDF documents. The app extracts text from the uploaded PDFs, processes the text using a language model, and provides responses to user queries. It can also fetch additional information from the web when the answer is not found in the provided documents.

Features

• PDF Text Extraction: Upload multiple PDFs and extract text for analysis.
• Text Chunking: Split extracted text into manageable chunks for processing.
• Conversational Interface: Engage in a Q&A session with the extracted content using a conversational model.
• Web Search Integration: Fetch additional information from the web using Tavily when the answer is not found in the documents.
• Interactive UI: User-friendly interface built with Streamlit.

Technologies Used

• Python: Core programming language.
• Streamlit: Framework for creating the interactive web application.
• PyPDF2: Library for extracting text from PDF files.
• Langchain: Used for text chunking and prompt management.
• GPT4AllEmbeddings & FAISS: For creating vector stores from text chunks.
• HuggingFaceHub: For integrating large language models.
• Tavily: For querying the internet to fetch additional information.
• HTML/CSS: Custom styling for the chat interface.

Install local models

GPT4All is a locally run model and downloading this model is the major pre-requisite for execution of this project. Download from here

Installation

1. Clone the repository:

   git clone https://github.com/your-username/ChatWithPDFs.git
   cd ChatWithPDFs

2. Create and activate a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows use `venv\Scripts\activate`

3. Install the required packages:

   pip install -r chatwpdfproject.yaml

Usage

1. Run the Streamlit app:

   streamlit run app.py

2. Upload your PDF documents: Use the sidebar to upload multiple PDFs.
3. Ask questions: Enter your question in the provided text input box and get responses based on the content of your PDFs or additional web links.

Contributing

Contributions are welcome! Please submit a pull request or open an issue to discuss your ideas or report bugs.