PaperQA is a Python-based question-answering application for research papers and documents. It uses OpenAI-powered embeddings and FAISS for vector storage to provide accurate answers based on the content of uploaded papers in .pdf, .docx, and .txt formats.

• File Support: Reads and processes .pdf, .docx, and .txt files.
• Embeddings & Vector Search: Converts document text to embeddings with OpenAI for similarity-based search.
• Interactive Q&A: Provides an interactive prompt for querying the content.
• Extensible Design: Modular code structure for ease of expansion and customization.

• Python 3.7+
• OpenAI API Key
• FAISS
• LangChain and other dependencies in requirements.txt

1. Clone the Repository

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

2. Install Dependencies

   pip install -r requirements.txt

3. Environment Setup

   Create a .env file in the root directory and add your OpenAI API key:

   OPENAI_API_KEY=your_openai_api_key_here
   

1. Place your documents in the data folder.

2. Run the application:

   python main.py

3. Enter questions related to the content of the papers in the prompt. Type end to exit.

• document_reader.py: Functions to read .pdf, .docx, and .txt files.
• text_processing.py: Splits large texts into manageable chunks.
• embeddings.py: Generates and stores embeddings for document similarity search.
• main.py: The main application logic, including the interactive prompt.

After running main.py, you can input questions like:

Ask a question about the paper: What are the main conclusions?

The model will return a response based on the document content.

Contributions are welcome! Please fork the repository and submit a pull request.

This project is licensed under the MIT License.