KO
|
EN
gitlite — search
Search
#python
#java
#python3
#arduino
#golang
#machine-learning
#rust
#html
#flask
#javascript
#seismology
#nodejs
clausenaut
★ 16
Open GitHub ↗
LLM-Powered Document Query System
Download README (.md)
Explore Similar Repositories
verdex-mcp
:
MCP Server for Playwright Test Writing
AAGATE
:
No description available.
mind_haven_backend
:
No description available.
langchain4f-paperfly
:
尚硅谷的langchain4j学习笔记，并补充了尚硅谷关于RAG高级部分没有讲的代码调用示例
holla-at-your-boy
:
GitHub Action that lets you holla at your boy (comment on PRs with code comments)
// repository documentation
Was this content helpful?
★ 0
(0 ratings)
Select Rating:
★
★
★
★
★
Submit Feedback
Recent Feedback
×
Download README
Do you want to download the
README.md
file for
clausenaut
?
Download (.md)
--- title: Clausenaut emoji: 🔥 colorFrom: yellow colorTo: red sdk: gradio sdk_version: 5.42.0 app_file: app.py pinned: false short_description: 'insurance Document analyzer ' license: mit --- # 🚢 ClauseNaut **An LLM-Powered Document Query System for Intelligent Claim Adjudication** ClauseNaut is a **Retrieval-Augmented Generation (RAG)** system designed to analyze policy documents and make informed decisions on user claims. It features a **user-friendly web interface** built with [Gradio](https://www.gradio.app/), allowing users to upload documents (PDFs, DOCX), ask complex questions in natural language, and receive a **structured JSON-based decision** with justifications. <img width="2880" height="1800" alt="image" src="https://github.com/user-attachments/assets/18318d09-1365-435b-ab0d-139397c4aab6" /> --- ## ✨ Key Features - 📄 **Multi-Format Document Support** – Upload and process `.pdf` and `.docx` files. - 🧠 **Intelligent Query Structuring** – Uses **Google’s Gemini model** to parse natural language queries into structured key-value pairs. - 🔍 **Semantic Search** – Employs **LangChain + Chroma vector store** for retrieving the most relevant clauses. - 🤖 **AI-Powered Adjudication** – Gemini LLM compares clauses against the user’s claim to render a decision. - 📝 **Structured Output** – Returns a JSON object containing: - Decision (Approved/Rejected) - Approved amount (if applicable) - Detailed justification citing source clauses - 💻 **Simple Web UI** – A seamless **3-step Gradio interface** for claim adjudication. --- ## ⚙️ How It Works (The RAG Pipeline) ### Phase 1: Knowledge Base Creation 1. **Document Upload** – Upload one or more policy documents. 2. **Document Loading & Chunking** – LangChain loaders split documents into manageable chunks. 3. **Embedding & Indexing** – Chunks converted into embeddings using HuggingFace → stored in **Chroma DB**. ✅ Knowledge base is ready for queries. ### Phase 2: Claim Processing 1. **User Query** – e.g., *“46-year-old male, knee surgery in Pune, 3-month-old insurance policy”*. 2. **Query Structuring** – First LLM call extracts fields into JSON (Age, Procedure, Location, etc.). 3. **Context Retrieval** – Semantic search fetches relevant policy clauses. 4. **Final Decision Generation** – Second LLM call adjudicates claim → produces structured JSON with verdict & justification. --- ## 🚀 Getting Started ### Prerequisites - Python **3.8+** - Google Gemini API Key ([Get it here](https://aistudio.google.com/)) --- ### 1️⃣ Clone the Repository ```bash git clone https://github.com/your-username/ClauseNaut.git cd ClauseNaut ``` 2. Set Up Virtual Environment Windows ```bash python -m venv venv venv\Scripts\activate ``` macOS/Linux ```bash python3 -m venv venv source venv/bin/activate ``` 4. Install Dependencies ```bash gradio google-generativeai langchain-community pypdf docx2txt sentence-transformers chromadb warnings ``` Then install: ```bash pip install -r requirements.txt ``` 6. Configure Environment Variables ```bash Create a .env file in the project root: GOOGLE_API_KEY="YOUR_GEMINI_API_KEY_HERE" Or set it as a system environment variable. ``` 8. Run the Application ```bash python app.py ``` Open browser → http://127.0.0.1:7860 📖 Usage Guide Upload Documents – Add PDF/DOCX policy files. Wait for Knowledge Base – Confirm readiness status. Define Extraction Fields – Default fields are provided (can be customized). Ask a Claim Query – Enter claim details in natural language. Submit – Click Submit Query. Review Output – View final structured JSON verdict in the Final Decision panel. 🛠️ Tech Stack Frontend/UI: Gradio LLM: Google Gemini Frameworks: LangChain, HuggingFace Embeddings Vector DB: Chroma Document Parsing: PyPDF, docx2txt 📌 Roadmap (Future Enhancements) ✅ Add support for Excel & TXT files ✅ Multi-user session handling 🚧 Integration with hospital/insurance APIs 🚧 Advanced claim analytics dashboard 🤝 Contributing Contributions are welcome! Fork this repo Create a feature branch (git checkout -b feature-name) Commit your changes Submit a PR 🎉 📜 License This project is licensed under the MIT License. Made with ❤️ by Eswar 👉 This version is clean, professional, and visually appealing for GitHub. Do you also want me to **add a small architecture diagram (RAG pipeline workflow)** inside the README using Markdown + Mermaid, so it looks more visually descriptive?
