OA Detection and Fusion Model App with RAG-based Reporting

This repository contains a Flask-based web application for osteoarthritis (OA) prediction and detection. The application integrates multiple machine learning models to process clinical data and X-ray images, and it uses a Retrieval-Augmented Generation (RAG) pipeline to automatically generate patient reports. A separate FAISS index is used to quickly retrieve context from a document (e.g., report_rules.txt) to guide report generation.

Features

• Clinical Data Processing
  • Accepts clinical parameters via a form.
  • Uses a pre-trained TensorFlow model and XGBoost classifiers for multi-class (including KL scoring) and binary OA predictions.
• Image Processing
  • Upload X-ray images for model prediction using TensorFlow-based image processing.
• Fusion of Predictions
  • Combines predictions from both clinical data and image analysis.
• RAG-based Patient Report Generation
  • Uses a separate FAISS index built from a report rules document (report_rules.txt) to retrieve relevant context.
  • Augments the patient report query with retrieved context and uses the OpenAI API (via a proxy) to generate a concise, evidence-based report.
• Chatbot Integration
  • Provides a chatbot interface (via the /chat route) that supports a RAG-based dialogue system for answering user queries.
• Markdown Support
  • A custom Jinja filter converts text wrapped in ** to bold HTML, enhancing report readability.
• Security & Configuration
  • Uses environment variables (via python-dotenv) to securely manage API tokens and configuration settings.

Prerequisites

• Python 3.8+
• pip
• Virtualenv (recommended)

Installation

1. Clone the repository:

   git clone https://github.com/yourusername/oafinalllm.git
   cd oafinalllm/oa_detection_app

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 requirements.txt

   A sample requirements.txt might include:

   Flask
   tensorflow
   numpy
   pandas
   joblib
   xgboost
   scikit-learn
   faiss-cpu
   openai
   python-dotenv
   

Configuration

1. Environment Variables:

   Create a .env file in the root of the project (or in the oa_detection_app directory) with the following keys:

   
   AIPROXY_TOKEN=your_actual_token_here
   API_BASE_URL=https://aiproxy.sanand.workers.dev/openai/v1 replace it with your base and token