Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

PolyDocs is a Streamlit application that enables users to interactively chat with the contents of their PDF files using advanced AI models. This app leverages Google Generative AI for embedding and querying, allowing detailed and context-aware question-answering from PDF documents.

• Upload Multiple PDFs: Easily upload multiple PDF files for processing.
• Intelligent Text Extraction: Extracts text from uploaded PDFs efficiently.
• Context-Aware Question Answering: Ask questions about the content of your PDFs and receive detailed answers.
• Interactive User Interface: Simple and intuitive interface for seamless user experience.
• Persistent Vector Store: Stores embeddings of text chunks locally using FAISS for fast similarity search.

A demo of the app can be found here: PolyDocs Demo

To run this application locally, follow these steps:

1. Clone the repository:

git clone https://github.com/shubhendu-ghosh-DS/polydocs.git
cd polydocs

2. Set up a virtual environment (optional but recommended):

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

3. Install the required packages:

pip install -r requirements.txt

4. Set up your Google API key: Create a .env file in the root directory of your project and add your Google API key:

GOOGLE_API_KEY=your_google_api_key_here

To run the application, execute:

streamlit run app.py

1. Upload PDFs: Use the sidebar menu to upload your PDF files.
2. Process PDFs: Click the "Submit" button to process the uploaded PDFs.
3. Ask Questions: Enter your questions in the text input field and receive answers based on the content of the PDFs.

• Streamlit: For building the web application.
• Google Generative AI: For embedding and querying text.
• FAISS: For creating and querying the vector store.
• PyPDF2: For extracting text from PDF files.
• LangChain: For managing the conversational chain and text processing.
• otenv: For loading environment variables.

app.py

• Imports and Configuration: Sets up necessary imports and configures the Google Generative AI API key.
• Helper Functions:
  • get_pdf_text(pdf_docs): Extracts text from the uploaded PDF files.
  • get_text_chunks(text): Splits the extracted text into manageable chunks.
  • get_vector_store(text_chunks): Creates and saves a FAISS vector store from the text chunks.
  • get_conversational_chain(): Sets up the conversational chain using a custom prompt template.
  • user_input(user_question): Processes the user's question by querying the FAISS index and generating a response.
  • Main Function: Manages the user interface, handles file uploads, and processes user input.

requirements.txt Lists all the dependencies required to run the application:

streamlit
google-generativeai
python-dotenv
langchain
PyPDF2
chromadb
faiss-cpu
langchain_google_genai
langchain-community

Contributions are welcome! Please fork the repository and create a pull request with your changes.

This project is licensed under the apache-2.0 License.