🏛️ Legal Research Engine - Your AI Legal Assistant

A simple, powerful AI tool that helps lawyers, judges, and law students research legal cases quickly and efficiently.

💡 Think of it like having a super-smart legal assistant that can instantly read through thousands of legal documents and answer your questions!

🎯 What Does This Tool Do?

This Legal Research Engine is like having a personal AI lawyer that can:

✅ Answer legal questions by searching through your law documents
✅ Find similar cases to the one you're working on
✅ Create case summaries automatically
✅ Show connections between different legal cases
✅ Extract important citations from legal documents
✅ Compare different legal documents side by side

🌟 Perfect For:

Lawyers researching case law and precedents
Judges reviewing similar cases and legal principles
Law Students studying and understanding legal concepts
Legal Researchers analyzing large volumes of legal documents
Court Staff preparing case briefs and summaries

🚀 Quick Start (Super Easy!)

Step 1: Get the Prerequisites

You'll need these programs installed on your computer:

Python 3.8+ - Download here
LM Studio - Download here (This runs the AI on your computer)
Git - Download here (To download the code)

Step 2: Download the Legal Research Engine

Open your computer's terminal/command prompt and type:

# Download the legal research engine
git clone https://github.com/PiKa919/Research-Engine-For-Courts.git

# Go into the folder
cd "Research-Engine-For-Courts"

Step 3: Set Up Python Environment

# Create a special Python environment for this project
conda create -n legal python=3.11 -y

# Switch to this environment
conda activate legal

# Install all the required tools
pip install -r requirements.txt

Step 4: Set Up LM Studio (Your AI Brain)

Open LM Studio (the app you downloaded)
Download Models:
- Go to the "Search" tab
- Download: text-embedding-embeddinggemma-300m-qat (for understanding documents)
- Download any chat model like: Llama-2-7B-Chat or Mistral-7B (for answering questions)
Start the Server:
- Go to "Developer" tab
- Load both models
- Click "Start Server" - it should run on http://127.0.0.1:1234

Step 5: Add Your Legal Documents

# Put your PDF legal documents in the data folder
# Then tell the system to read them
python src/ingest.py

Step 6: Start the Legal Research Engine

# Make sure you're using the legal environment
conda activate legal

# Start the web application
streamlit run app.py

Step 7: Open in Your Browser

Open your web browser (Chrome, Firefox, etc.)
Go to: http://localhost:8501
Start researching! 🎉

📖 How to Use (Simple Guide)

🔍 Tab 1: Chat - Ask Legal Questions

What it does: Ask questions about your legal documents and get instant answers with references.

How to use:

Type your question like: "What are the main provisions of the Commercial Courts Act?"
Press Enter
Get a detailed answer with citations from your documents
Ask follow-up questions for more details

Example Questions:

"What is the limitation period for commercial disputes?"
"How are commercial court judges appointed?"
"What are the key differences between civil and commercial procedures?"

🎨 Tab 2: Themes - Change the Look

What it does: Change how the app looks to suit your preference.

Available Themes:

Light Theme - Bright and clean (default)
Dark Theme - Easy on the eyes
Legal Blue - Professional blue colors
Legal Classic - Traditional legal styling
Auto - Matches your system theme

📊 Tab 3: Citation Graph - See Case Connections

What it does: Shows you how legal cases are connected to each other in a visual network.

How to use:

The system automatically analyzes your documents
Click and drag to explore the network
See which cases cite which other cases
Understand the relationships between legal precedents

📈 Tab 4: Citation Analysis - Document Statistics

What it does: Shows you statistics about your legal document collection.

You'll see:

Most cited cases
Citation frequency charts
Document usage patterns
Legal trend analysis

📋 Tab 5: Case Brief Generator - Auto-Create Summaries

What it does: Automatically creates professional case briefs from legal documents.

How to use:

Select a document from your collection
Click "Generate Brief"
Get a structured summary with:
- Case facts
- Legal issues
- Court decision
- Key legal principles
- Citations

⚖️ Tab 6: Precedent Analysis - Find Similar Cases

What it does: Finds cases similar to your current legal issue.

How to use:

Describe your legal situation
The AI finds similar cases from your documents
See how courts decided similar issues
Use these as precedents for your case

📄 Tab 7: Document Comparison - Compare Legal Docs

What it does: Compares two or more legal documents side by side.

How to use:

Select documents to compare
See similarities and differences highlighted
Understand how legal positions differ
Export comparison reports

📤 Tab 8: Export - Save Your Research

What it does: Save your research results in different formats.

Available Formats:

📄 PDF - Professional reports for printing
📝 Word Document - Editable legal documents
💾 Text File - Simple text format
📊 JSON - Data format for further analysis
🌐 HTML - Web page format

� Tab 9: Analytics - Performance Dashboard

What it does: Shows you how well the system is working and usage statistics.

You'll see:

Query response times
System performance metrics
Usage patterns
Success rates

🔧 Tab 10: Workflow Builder - Custom Research Processes

What it does: Create custom research workflows for specific types of legal work.

How to use:

Choose your research type (case law, statutory analysis, etc.)
Configure the steps
Save as a template for future use
Share workflows with colleagues

📚 Tab 11: Knowledge Graph - Legal Concept Network

What it does: Shows relationships between legal concepts, cases, and statutes.

Features:

Interactive concept mapping
Legal principle connections
Statute-case relationships
Searchable knowledge network

⚙️ Tab 12: System Settings - Configure the Tool

What it does: Adjust settings to customize how the system works.

Options:

Model selection
Performance settings
Export preferences
Theme customization
Cache management

🎯 Real-World Use Cases

For Lawyers:

Case Research: "Find all cases related to breach of contract in commercial disputes"
Precedent Analysis: "Show me similar cases to my current client's situation"
Brief Preparation: "Generate a case brief for [specific case]"

For Judges:

Case Review: "What are the key legal principles in this type of case?"
Precedent Research: "How have similar cases been decided?"
Legal Analysis: "What are the statutory provisions relevant to this matter?"

For Law Students:

Study Aid: "Explain the main points of the Contract Act"
Case Analysis: "Break down this judgment for me"
Research Help: "Find cases that illustrate this legal principle"

🔧 Troubleshooting (If Something Goes Wrong)

❌ Can't Connect to LM Studio

Problem: The AI can't talk to LM Studio
Solution:

Make sure LM Studio is running
Check that the server is started in LM Studio
Verify the URL is http://127.0.0.1:1234

❌ App Won't Start

Problem: Error messages when starting the app
Solution:

Make sure you're in the right folder
Activate the conda environment: conda activate legal
Reinstall packages: pip install -r requirements.txt

❌ No Documents Found

Problem: The system can't find your legal documents
Solution:

Put PDF files in the data/ folder
Run: python src/ingest.py
Wait for processing to complete

❌ Slow Responses

Problem: AI takes too long to answer
Solution:

Use smaller AI models in LM Studio
Close other programs to free up memory
Process fewer documents at once

❌ Out of Memory

Problem: Computer runs out of memory
Solution:

Increase RAM allocation in LM Studio
Use smaller models
Process documents in smaller batches
Restart the application

💡 Tips for Best Results

📚 Document Preparation:

Use clear, text-based PDFs (not scanned images)
Name files clearly (e.g., "Supreme_Court_2023_Contract_Case.pdf")
Organize by topic or case type for better results

❓ Asking Good Questions:

Be specific: Instead of "Tell me about contracts", ask "What are the essential elements of a valid contract under Indian law?"
Use legal terms when you know them
Ask follow-up questions to get more detail

� Using Results:

Always verify AI answers with original sources
Cross-reference with multiple cases
Use citations provided to find original documents

🤝 Getting Help

Need Support?

Check this README first
Look at the troubleshooting section above
Open an issue on GitHub if you find bugs
Review LM Studio documentation for AI model issues

Want to Improve the Tool?

Suggest new features by opening a GitHub issue
Report bugs with detailed descriptions
Contribute code if you're a developer

📄 License

This project is free to use under the MIT License. You can use it, modify it, and share it for any purpose, including commercial use.

🎉 Final Words

Congratulations! You now have a powerful AI legal research assistant at your fingertips. This tool can help you:

Save hours of manual research time
Find relevant cases you might have missed
Create professional documents quickly
Understand complex legal relationships
Make your legal work more efficient

Remember: This AI assistant is a tool to help you, not replace your legal expertise. Always verify important information and use your professional judgment.

Happy Legal Researching! ⚖️✨

Built with ❤️ for the legal community | Last updated: September 29, 2025

🆕 Human-in-the-Loop: Built-in approval workflows for critical decisions

📊 Advanced Tools

Citation Graph Visualization: Interactive network of legal citations
🆕 Automated Case Brief Generation: AI-generated structured legal summaries
🆕 Advanced Precedent Analysis: Multi-agent case law research
Document Comparison: Side-by-side analysis of multiple documents
Timeline Builder: Chronological case progression tracking
🆕 Workflow Monitoring: Real-time tracking of analysis progress

🎯 Legal Research Capabilities

Indian Law Focus: Specialized for Indian legal system and courts
🆕 Automatic Citation Extraction: AI-powered legal reference identification
Case Law Research: Intelligent precedent finding and analysis
Document Similarity: Semantic search across legal documents
🆕 Confidence Scoring: Quality assessment of analysis results

Installation

Prerequisites

Python 3.8+: Programming language runtime
LM Studio: Local AI model server (download)
Git: Version control system

System Requirements

RAM: Minimum 8GB, Recommended 16GB+
Storage: 5GB+ free space for models and documents
OS: Windows 10/11, macOS, or Linux

Setup Steps

Clone the repository

git clone https://github.com/PiKa919/Research-Engine-For-Courts.git
cd Research-Engine-For-Courts

Create conda environment

conda create -n legal python=3.11 -y
conda activate legal

Install dependencies
```
pip install -r requirements.txt
```
Install LM Studio
- Download from lmstudio.ai
- Install and launch LM Studio
- Download models:
  - Embedding: text-embedding-embeddinggemma-300m-qat
  - Chat: Any compatible model (Llama 2, Mistral, etc.)
Configure LM Studio
- Go to "Developer" tab
- Load your downloaded models
- Start local server (should run on http://127.0.0.1:1234)

Usage

Prepare Documents

# Add PDF documents to the data/ folder
# Then run ingestion
python src/ingest.py

Run the Application

# Start the Streamlit app
streamlit run app.py --server.headless true --server.port 8501

# Access at: http://localhost:8501

Main Features

💬 Chat - Interactive Legal Research

Ask questions about your legal documents and get AI-powered answers with citations.

📊 Citation Graph - Visual Citation Network

Interactive visualization of legal citations and case relationships.

� Citation Analysis - Document Analytics

Frequency analysis and citation patterns in your document collection.

📋 Case Brief Generator - Automated Briefs

Generate structured case briefs with key legal elements.

⚖️ Precedent Analysis - Case Similarity

Find similar legal cases and analyze legal principles.

📄 Document Comparison - Side-by-Side Analysis

Compare multiple legal documents with legal implications.

📊 Evaluation - System Performance

DeepEval metrics and system accuracy assessment.

Embedding Model: text-embedding-embeddinggemma-300m-qat or similar
Chat Model: Any compatible model (Llama 2, Mistral, etc.)

📦 Installation

1. Clone the Repository

git clone https://github.com/PiKa919/Research-Engine-For-Courts.git
cd "Research Engine For Courts"

2. Create Conda Environment

# Create new environment
conda create -n legal python=3.11 -y

# Activate environment
conda activate legal
## Recent Updates

### v1.0.1 - September 2025
- **✅ Fixed** `langchain_experimental` import error - added missing package dependency
- **✅ Fixed** RAG chain TypeError - corrected chain structure for proper string input handling
- **✅ Improved** error handling and logging throughout the application
- **✅ Enhanced** documentation and setup instructions

## Project Structure

Research-Engine-For-Courts/ ├── app.py # Main Streamlit application ├── src/ │ ├── init.py │ ├── api.py # FastAPI REST endpoints │ ├── config.py # Configuration management │ ├── retrieval.py # RAG implementation & vector search │ ├── ingest.py # Document ingestion pipeline │ ├── enhanced_document_processor.py # Advanced document processing │ ├── precedent_analyzer.py # Legal precedent analysis │ ├── document_comparator.py # Document comparison tools │ ├── case_brief_generator.py # Automated case brief generation │ ├── knowledge_graph.py # Citation graph visualization │ ├── evaluation.py # System evaluation & metrics │ ├── caching.py # Performance caching system │ ├── monitoring.py # Performance monitoring │ └── evaluation.py # Model evaluation tools ├── data/ # Legal document storage ├── chroma/ # Vector database storage ├── requirements.txt # Python dependencies ├── pyrightconfig.json # Python type checking └── README.md # This file


## Configuration

### Model Configuration
```python
# LM Studio models (local)
EMBEDDING_MODEL = "text-embedding-embeddinggemma-300m-qat"
LLM_MODEL = "local-model"

# Performance settings
MAX_TOKENS = 2048
TEMPERATURE = 0.1
TOP_P = 0.9

Environment Variables

# Optional: Create .env file for additional configuration
LM_STUDIO_BASE_URL=http://127.0.0.1:1234
DATA_PATH=./data
CHROMA_PATH=./chroma

API Reference

REST API Endpoints

The system provides REST API endpoints for integration:

GET /health - Health check
POST /query/ - Synchronous legal research query
POST /query/async - Asynchronous legal research query
GET /documents/ - List available documents
POST /documents/ingest - Ingest new documents

Python API

from src.retrieval import create_rag_chain

# Initialize the system
async_rag_chain, sync_rag_chain = create_rag_chain()

# Perform legal research
result = sync_rag_chain("What are the provisions for commercial courts?")
print(result["answer"])

Troubleshooting

Common Issues

LM Studio Connection

# Check if LM Studio server is running
curl http://127.0.0.1:1234/v1/models

Import Errors

# Reinstall dependencies
pip install -r requirements.txt --force-reinstall

Vector Store Issues

# Reset vector database
rm -rf chroma/
python src/ingest.py

Memory Issues

Increase RAM allocation in LM Studio
Use smaller models
Process documents in smaller batches

Performance

Benchmarks

Document Processing: ~50 pages/minute
Query Response: < 3 seconds average
Vector Search: < 100ms for similarity search
Memory Usage: ~2GB base + 0.5GB per 1000 documents

Optimization Tips

Use GPU acceleration in LM Studio for better performance
Batch document ingestion for large collections
Configure appropriate chunk sizes for your documents
Monitor system resources during heavy usage

Contributing

Development Setup

Fork the repository
Create a feature branch: git checkout -b feature/your-feature
Make your changes
Test thoroughly
Submit a pull request

Code Standards

Follow PEP 8 for Python code
Add docstrings to all functions
Include type hints where possible
Write comprehensive tests

Adding New Features

Create feature in appropriate module in src/
Update configuration if needed
Add to Streamlit interface in app.py
Update documentation

License

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

Acknowledgments

LangChain - Framework for LLM applications
Streamlit - Web application framework
LM Studio - Local AI model server
Chroma - Vector database
PyVis - Network visualization

Built with ❤️ for legal professionals and researchers

Last updated: September 22, 2025

# Make sure you're in the project directory
cd "c:\FILES_PIKA\Research Engine For Courts"

# Activate environment
conda activate legal

# Run the application
streamlit run app.py --server.headless true --server.port 8501

Access the Application

Open your browser
Navigate to: http://localhost:8501
The application will load with multiple tabs

📖 Usage

Main Interface Tabs

1. 💬 Chat - Interactive Legal Research

Ask questions about your legal documents
Get AI-powered answers with citations
Explore document relationships

2. 📊 Citation Graph - Visual Citation Network

Interactive visualization of legal citations
Explore connections between cases
Network analysis of legal precedents

3. 📈 Citation Analysis - Document Analytics

Frequency analysis of citations
Document usage statistics
Citation patterns and trends

4. 📋 Case Brief Generator - Automated Briefs

Generate structured case briefs
Extract key legal elements
Professional legal document formatting

5. ⚖️ Precedent Analysis - Case Similarity

Find similar legal cases
Analyze legal principles and holdings
Precedent research and validation

6. 📄 Document Comparison - Side-by-Side Analysis

Compare multiple legal documents
Identify similarities and differences
Legal implications analysis

7. 📊 Evaluation - System Performance

DeepEval metrics and benchmarks
System accuracy assessment
Performance optimization insights

Example Usage

Basic Q&A

Question: "What are the key principles of Section 138 of the Negotiable Instruments Act?"

Response: The system will search your documents and provide:
- Relevant case law
- Section references
- Legal analysis with citations

Document Upload

# Add new documents
# 1. Place PDF in data/ folder
# 2. Run: python src/ingest.py
# 3. New documents are automatically indexed

📁 Project Structure

Research Engine For Courts/
├── app.py                          # Main Streamlit application
├── src/
│   ├── __init__.py
│   ├── app.py                     # Alternative app interface
│   ├── config.py                  # Configuration settings
│   ├── retrieval.py               # RAG implementation
│   ├── ingest.py                  # Data ingestion pipeline
│   ├── enhanced_document_processor.py  # Document processing
│   ├── precedent_analyzer.py      # Legal precedent analysis
│   ├── document_comparator.py     # Document comparison
│   ├── case_brief_generator.py    # Case brief generation
│   ├── knowledge_graph.py         # Citation graph visualization
│   └── evaluation.py              # System evaluation
├── data/                          # Legal document storage
│   ├── Commercial Courts Act, 2015.pdf
│   ├── Commercial Courts Rules, 2019.pdf
│   └── ...
├── chroma/                        # Vector database storage
├── requirements.txt               # Python dependencies
├── README.md                      # This file
└── pyrightconfig.json             # Python type checking

⚙️ Configuration

Model Configuration (`src/config.py`)

# LM Studio models (local)
EMBEDDING_MODEL = "text-embedding-embeddinggemma-300m-qat"
LLM_MODEL = "local-model"  # LM Studio will use loaded model

# Paths
DATA_PATH = "data/"
CHROMA_PATH = "chroma/"

Environment Variables

Create a .env file for sensitive configurations:

# .env file
GOOGLE_API_KEY=your_key_here  # Only if using Google models (not recommended)

🔧 Troubleshooting

Common Issues

1. Import Errors

# Reinstall dependencies
pip install -r requirements.txt --force-reinstall

2. LM Studio Connection Issues

Check if LM Studio is running
Verify server URL: http://127.0.0.1:1234
Ensure models are loaded in LM Studio
Check firewall settings

3. Vector Store Errors

# Reset vector store
rm -rf chroma/
python src/ingest.py

4. Memory Issues

Increase RAM allocation in LM Studio
Use smaller models
Process documents in smaller batches

5. Port Conflicts

# Use different port
streamlit run app.py --server.port 8502

Performance Optimization

For Better Speed:

Use GPU in LM Studio (if available)
Optimize model parameters
Use smaller embedding models
Increase batch sizes in ingestion

For Better Accuracy:

Use larger models
Fine-tune prompts
Add more legal documents
Improve document preprocessing

🤝 Contributing

Development Setup

Fork the repository

Create a feature branch:

git checkout -b feature/your-feature-name

Make your changes
Test thoroughly
Submit a pull request

Code Standards

Follow PEP 8 for Python code
Add docstrings to all functions
Include type hints where possible
Write comprehensive tests

Adding New Features

Create feature in appropriate module
Update configuration if needed
Add to Streamlit interface
Update documentation

📄 License

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

🙏 Acknowledgments

LangChain - Framework for LLM applications
Streamlit - Web application framework
LM Studio - Local AI model server
Chroma - Vector database
PyVis - Network visualization

📞 Support

For support and questions:

Open an issue on GitHub
Check the troubleshooting section
Review LM Studio documentation

Built with ❤️ for legal professionals and researchers

Last updated: September 16, 2025

Name		Name	Last commit message	Last commit date
.roo		.roo
data		data
src		src
.gitignore		.gitignore
CONTRIBUTING.md		CONTRIBUTING.md
ENHANCEMENT_ROADMAP.md		ENHANCEMENT_ROADMAP.md
FEATURE_SUMMARY.md		FEATURE_SUMMARY.md
LANGGRAPH_GUIDE.md		LANGGRAPH_GUIDE.md
NEXT_FEATURES.md		NEXT_FEATURES.md
README.md		README.md
README_NEW.md		README_NEW.md
app.py		app.py
citation_graph.html		citation_graph.html
data_collector.py		data_collector.py
demo_features.py		demo_features.py
example_langgraph_usage.py		example_langgraph_usage.py
project.md		project.md
rag_pipeline.ipynb		rag_pipeline.ipynb
requirements.txt		requirements.txt
test_knowledge_graph.html		test_knowledge_graph.html
timeline.md		timeline.md

Latest commit

History

Folders and files

Repository files navigation

🏛️ Legal Research Engine - Your AI Legal Assistant

🎯 What Does This Tool Do?

🌟 Perfect For:

🚀 Quick Start (Super Easy!)

Step 1: Get the Prerequisites

Step 2: Download the Legal Research Engine

Step 3: Set Up Python Environment

Step 4: Set Up LM Studio (Your AI Brain)

Step 5: Add Your Legal Documents

Step 6: Start the Legal Research Engine

Step 7: Open in Your Browser

📖 How to Use (Simple Guide)

🔍 Tab 1: Chat - Ask Legal Questions

🎨 Tab 2: Themes - Change the Look

📊 Tab 3: Citation Graph - See Case Connections

📈 Tab 4: Citation Analysis - Document Statistics

📋 Tab 5: Case Brief Generator - Auto-Create Summaries

⚖️ Tab 6: Precedent Analysis - Find Similar Cases

📄 Tab 7: Document Comparison - Compare Legal Docs

📤 Tab 8: Export - Save Your Research

� Tab 9: Analytics - Performance Dashboard

🔧 Tab 10: Workflow Builder - Custom Research Processes

📚 Tab 11: Knowledge Graph - Legal Concept Network

⚙️ Tab 12: System Settings - Configure the Tool

🎯 Real-World Use Cases

For Lawyers:

For Judges:

For Law Students:

🔧 Troubleshooting (If Something Goes Wrong)

❌ Can't Connect to LM Studio

❌ App Won't Start

❌ No Documents Found

❌ Slow Responses

❌ Out of Memory

💡 Tips for Best Results

📚 Document Preparation:

❓ Asking Good Questions:

� Using Results:

🤝 Getting Help

Need Support?

Want to Improve the Tool?

📄 License

🎉 Final Words

📊 Advanced Tools

🎯 Legal Research Capabilities

Installation

Prerequisites

System Requirements

Setup Steps

Usage

Prepare Documents

Run the Application

Main Features

💬 Chat - Interactive Legal Research

📊 Citation Graph - Visual Citation Network

� Citation Analysis - Document Analytics

📋 Case Brief Generator - Automated Briefs

⚖️ Precedent Analysis - Case Similarity

📄 Document Comparison - Side-by-Side Analysis

📊 Evaluation - System Performance

📦 Installation

1. Clone the Repository

2. Create Conda Environment

Environment Variables

API Reference

REST API Endpoints

Python API

Troubleshooting

Common Issues

LM Studio Connection

Import Errors

Vector Store Issues

Memory Issues

Performance

Benchmarks

Optimization Tips

Model Configuration (`src/config.py`)