Legal-Bot:

Legal-Bot is a Streamlit-based application designed to democratize access to the Indian Penal Code (IPC). It empowers users to quickly find simple, accurate explanations of IPC sections through an intuitive conversational interface.

---

🧩 Problem Statement: "The Legal Knowledge Gap"

• Complexity of Law: IPC sections are written in dense legal language, making them hard to understand for non-lawyers.
• Accessibility Issues: Citizens often lack affordable, reliable access to legal experts for basic queries.
• Fragmented Resources: People must search across multiple sources (books, websites, forums), wasting time and risking misinformation.

Legal-Bot bridges this gap by providing a simple, conversational way to understand IPC sections.

---

💡 Our Approach & Importance

Legal-Bot uses a Retrieval-Augmented Generation (RAG) pipeline to combine fast document retrieval with LLM-powered explanations:

1. Retrieval (Precision): FAISS + Sentence-Transformers quickly find the most relevant IPC sections.
2. Generation (Clarity): Llama 3.1 (via Ollama) rewrites complex legal text into plain language.
3. Local Privacy: Runs entirely on your machine — no cloud dependency, ensuring confidentiality.
4. User-Centric Design: Streamlit UI makes interaction intuitive, lowering the barrier to legal knowledge.

Why it matters:
Legal-Bot empowers citizens, students, and professionals to understand their rights and responsibilities without needing a lawyer for every query. It’s a step toward accessible justice.

---

🏗️ System Architecture

graph TD
    A[User Query] --> B[Sentence-Transformers Embeddings]
    B --> C[FAISS Vector Search]
    C -->|Relevant IPC Sections| D{RAG Pipeline}
    
    subgraph Intelligence
        D --> E[Llama 3.1 via Ollama]
        E --> F[Plain-Language Legal Explanation]
    end
    
    F --> G[Streamlit Chat UI]
    G --> H[User]
    
    subgraph Data Layer
        I[legal_data.jsonl] --> C
        C --> J[ipc_faiss.index + ipc_metadata.json]
    end

Loading

---

✨ Features

• Intelligent IPC Explanations: Converts complex IPC sections into simple language.
• RAG Pipeline: Combines retrieval + generation for contextual accuracy.
• Local LLM Integration: Runs Llama 3.1 locally via Ollama.
• FAISS Vector Search: Lightning-fast retrieval.
• Streamlit UI: Clean, user-friendly interface.
• Dynamic Data Handling: Auto-generates FAISS index + metadata.
• Comprehensive Logging: Configurable via logging_config.yaml.

---

🚀 Getting Started

1. Clone the Repository

git clone https://github.com/arjunrd07/Legal-Bot.git
cd Legal-Bot

2. Set Up Ollama & Llama 3.1

ollama run llama3.1

(Downloads the model and starts an interactive session. Exit with /bye once complete.)

---

⚡ Running the Application

Option A: Automated Setup (Recommended)

• Download install.bat (Windows) from the latest GitHub Release.
• Double-click to run.
• Script handles Ollama setup, model download, venv creation, dependency installation, and app launch.

Option B: Manual Setup

pip install uv
python3 -m venv venv
source venv/bin/activate   # Linux/macOS
venv\Scripts\activate      # Windows
uv pip install -r requirements.txt
streamlit run app.py

App runs at http://localhost:8501.

---

🖥️ Usage

• Open the app in your browser.
• Enter queries like:
  • "What is IPC section 302?"
  • "Explain crimes against children under IPC."
  • "What are the punishments for theft?"
• Legal-Bot retrieves relevant IPC sections and generates simplified explanations.

---

📦 Dependencies

• Python 3.8+
• Streamlit
• Sentence-Transformers
• FAISS-CPU
• PyTorch
• Ollama Python Library
• Utilities: langchain, pydantic, etc.

---

🎯 Impact

Legal-Bot makes legal knowledge accessible by simplifying IPC sections into conversational explanations.
It reduces the barrier to understanding complex laws, empowering students, professionals, and citizens alike.

---

🔮 Future Roadmap

• Expanded Coverage: Beyond IPC to other Indian laws.
• Advanced Querying: Multi-section reasoning with cross-references.
• Enhanced Privacy: Fully offline inference with Ollama.
• Dockerized Deployment: Seamless containerized setup for production.