ContextualQAPDF

ContextualQAPDF is an intelligent application designed to help you interact with and extract meaningful insights from your PDF documents. Leveraging the power of AWS Bedrock and various AI models, ContextualQAPDF allows users to upload PDFs, ask questions, and receive detailed, contextually accurate answers.

Features

Upload PDF Documents: Easily upload PDF files to the application.
Intelligent Question Answering: Ask questions related to the content of the PDFs and receive detailed answers.
Vector Embedding: Utilize Titan Embeddings to generate embeddings for document content.
Vector Store Management: Create and manage a FAISS-based vector store for efficient retrieval.
Multiple AI Models: Switch between different AI models for question answering.

Installation

Clone the repository:

git clone https://github.com/yourusername/contextualqapdf.git
cd contextualqapdf

Set up a virtual environment:

python3 -m venv venv
source venv/bin/activate

Install the required packages:
```
pip install -r requirements.txt
```

Set up environment variables:

Create a .env file in the project root and add your AWS credentials:

AWS_ACCESS_KEY_ID=your_access_key_id
AWS_SECRET_ACCESS_KEY=your_secret_access_key
AWS_SESSION_TOKEN=your_session_token  # Only if using temporary credentials

Usage

Run the application:
```
streamlit run app.py
```
Upload PDF: Use the sidebar to upload your PDF documents.
Ask Questions: Enter your questions in the text input field and get answers based on the content of your PDFs.

Configuration

The application uses the following AI models:

Claude: AI21 model for general question answering.
Llama 2: Meta model for advanced context handling and detailed answers.

You can switch between models and manage vector stores using the sidebar options.

Code Overview

Data Ingestion: Load and process PDF documents using PyPDFDirectoryLoader and RecursiveCharacterTextSplitter.
Vector Embedding and Store: Generate embeddings with BedrockEmbeddings and manage them with FAISS.
Question Answering: Use RetrievalQA to handle user queries and provide detailed answers.

Troubleshooting

Ensure your AWS credentials are correctly set in the .env file.
Make sure the FAISS index file exists before attempting to query the vector store.
Check the console for detailed error messages and tracebacks.

Contributing

Contributions are welcome! Please open an issue or submit a pull request on GitHub.

Happy Document Searching with ContextualQAPDF!

10-disha/ContextualQAPDF