Fusion RAG: A Streamlit-Powered PDF Q&A System with Fusion-in-Decoder

Description

Welcome to Fusion RAG, a powerful web-based application designed to answer questions based on the content of uploaded PDF documents. This project leverages the Fusion-in-Decoder (FiD) approach for Retrieval-Augmented Generation (RAG), combining semantic similarity, technical term relevance, and recency to deliver accurate and contextually relevant responses. Built with modern technologies like Streamlit, LangChain, Chroma, and Ollama, Fusion RAG is ideal for researchers, students, or professionals who need to extract insights from documents efficiently.

The application features a user-friendly interface where you can upload PDFs, ask questions in a conversational format, and receive detailed, real-time responses with citations to the source documents for transparency and verifiability.

Features

Feature	Description
PDF Upload & Processing	Upload multiple PDF files, which are automatically split into chunks and indexed for querying.
Advanced Retrieval	Uses a custom FiDRetriever that combines semantic similarity (via Chroma), technical term relevance (via TF-IDF), and recency to fetch the most relevant document chunks.
Conversational Interface	Ask questions in a chat-like interface and receive detailed answers based on document content.
Real-Time Responses	Answers are streamed in real-time for a seamless user experience.
Error Handling	Robust error handling and cleanup mechanisms ensure reliable operation.
Source Citations	Responses include references to the source documents, enhancing trust and verifiability.

How It Works

The Fusion RAG system employs a sophisticated pipeline to process documents and generate answers. Here’s a detailed breakdown of the process:

Document Processing

• PDF Loading: Uploaded PDFs are loaded using PyPDFLoader from LangChain.
• Text Splitting: Documents are split into manageable chunks (1000 characters, 200-character overlap) using RecursiveCharacterTextSplitter.
• Metadata Tagging: Each chunk is tagged with metadata, such as the source file name, for traceability.

Retrieval

• Vector Search: Documents are embedded into a high-dimensional vector space using OllamaEmbeddings (model: nomic-embed-text:latest) and stored in a Chroma vector store. This enables semantic similarity searches.
• FiDRetriever: A custom retriever that scores documents based on:
  • Semantic Similarity: Using vector embeddings to find contextually relevant chunks.
  • Technical Term Relevance: Using TF-IDF to prioritize chunks with important keywords.
  • Recency: Prioritizing recently seen documents for relevance.
• The top 2 most relevant documents are retrieved for each query.

Generation

• FiDChain: Combines the retriever with an OllamaLLM (model: qwen2.5:latest, temperature: 0.3) to generate a detailed response.
• The response is structured to include an analysis summary, recommended approach, reference standards, and citations to the source documents.

Diagram of the RAG Pipeline

flowchart LR
    A(["User Query"])
    A --> B["FiDRetriever"]
    B --> C["Vector Search (Chroma)"]
    C --> D["Embeddings (OllamaEmbeddings)"]
    B --> E["TF-IDF Scoring"]
    B --> F["Recency Scoring"]
    C --> G["Retrieved Documents"]
    E --> G
    F --> G
    G --> H["Formatter"]
    H --> I["Prompt Template"]
    I --> J["LLM"]
    J --> K["Output Parser"]
    K --> L["Response"]

Loading

This diagram can be rendered in GitHub to visualize the pipeline from query to response.

Environment Setup

To run Fusion RAG, you’ll need to set up the following:

• Python 3.8 or later: Ensure Python is installed on your system. Download from python.org.
• Ollama: A tool for running large language models locally. Install it based on your operating system:
  • Windows: Download the installer from Ollama Download and run it.
  • macOS: Download the installer from Ollama Download, unzip it, and drag the Ollama.app to your Applications folder.
  • Linux: Run the installation script as per the Ollama GitHub repository.
• Python Libraries: Install the required dependencies listed in requirements.txt.

Installation

Follow these steps to set up the project:

1. Clone the Repository:

   git clone https://github.com/armanjscript/Fusion-RAG.git

2. Navigate to the Project Directory:

   cd Fusion-RAG

3. Install Dependencies:

   pip install -r requirements.txt

   The requirements.txt file includes dependencies like streamlit, langchain, chromadb, langchain-ollama, scikit-learn, and others.

4. Start Ollama: Ensure the Ollama service is running on your system. Follow the instructions from the Ollama GitHub repository to start the service.

5. Run the Streamlit App:

   streamlit run fusion_rag.py

   This will launch the app in your default web browser.

Usage

1. Upload PDFs:

   • In the sidebar, use the file uploader to select one or more PDF files.
   • The files are saved locally in the uploaded_pdfs directory and indexed in a Chroma database (chroma_db).

2. Ask Questions:

   • Enter your query in the chat input field in the main interface.
   • The chatbot will retrieve relevant document chunks, generate a response, and stream it in real-time.
   • Responses include citations to the source documents for reference.

3. Clear Documents:

   • Use the "Clear All Documents" button in the sidebar to reset the application by deleting uploaded files and the vector store.

Configuration

The system uses fixed parameters for optimal performance:

• Temperature: Set to 0.3 for controlled randomness in the language model’s responses.
• Retrieval: Retrieves the top 2 documents based on a combination of semantic similarity, TF-IDF scoring, and recency.

To modify these parameters, you can edit the fusion_rag.py file to adjust the OllamaLLM temperature or the FiDRetriever scoring logic.

Contributing

We welcome contributions to enhance Fusion RAG! To contribute:

• Fork the repository.
• Make your changes in a new branch.
• Submit a pull request with a clear description of your changes.

For detailed guidelines, refer to the CONTRIBUTING.md file.

License

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

Contact

For questions, feedback, or collaboration opportunities, please reach out:

• Email: [[email protected]]
• GitHub Issues: Open an issue on this repository for bug reports or feature requests.

Acknowledgments

This project builds on the following open-source technologies:

• Streamlit for the web interface
• LangChain for document processing and RAG pipeline
• Chroma for vector storage
• Ollama for local language models and embeddings
• Scikit-learn for TF-IDF vectorization

Citations

• LangChain Documentation
• Streamlit Documentation
• Chroma Documentation
• Ollama Documentation

Thank you for exploring Fusion RAG! We hope it simplifies your document analysis tasks.

#AI #RAG #FiD #PDFQ&A #Streamlit #LangChain #Chroma #Ollama #MachineLearning #NaturalLanguageProcessing