ChronoLaw

ChronoLaw is a web application that creates case timelines based on documents uploaded to it. The application analyzes documents using a locally hosted LLM to extract dates and relevant information, then displays them in a chronological timeline. Users can also chat with the timeline to ask questions about the documents and their dates.

Features

• Drag and drop document upload (PDF and DOCX)
• Automatic extraction of dates and related information from documents
• Interactive timeline display with filtering and sorting options
• Chat interface to ask questions about the timeline and documents
• Fully local processing with no external API calls

Prerequisites

• Node.js (v14 or higher)
• npm (v6 or higher)

Installation

1. Clone the repository:

   git clone <repository-url>
   cd chronolaw
   

2. Run the setup script to install all dependencies:

   npm run setup
   

   This script will:

   • Install server dependencies
   • Install client dependencies
   • Create necessary directories
   • Check if the LLM model is available

3. Download a LLaMA model:

   You need to download a LLaMA model in GGUF format to use with the application. We recommend using Mistral-7B-Instruct-v0.3.Q8_0.gguf for a good balance of performance and accuracy.

   a. Download the model from Hugging Face or another source.

   b. Place the downloaded model file in the models directory:

   mv /path/to/downloaded/Mistral-7B-Instruct-v0.3.Q8_0.gguf ./models/
   

   c. The setup script creates a .env file with default settings. You can customize the model path and other settings by editing this file.

Environment Variables

ChronoLaw uses environment variables for configuration. These are stored in a .env file in the project root. The setup script will create this file automatically from .env.example if it doesn't exist.

Variable	Description	Default
LLM_MODEL_PATH	Path to the LLM model file (relative to project root)	models/llama-2-7b-chat.gguf
LLM_CONTEXT_SIZE	Context size for the LLM	8192
LLM_BATCH_SIZE	Batch size for the LLM	512
LLM_GPU_LAYERS	Number of layers to offload to GPU (0 for CPU only)	0
PORT	Port for the server to listen on	5000

You can edit these variables to customize the application to your needs. For example, if you want to use a different model, you can change the LLM_MODEL_PATH variable.

Running the Application

1. Start the development server:

   npm run dev
   

   This will start both the backend server and the React frontend.

2. Open your browser and navigate to:

   http://localhost:3000
   

Usage

1. Upload Documents:

   • Drag and drop PDF or DOCX files into the upload area
   • Click "Process Documents" to analyze the files

2. View Timeline:

   • After processing, the application will display a timeline of events extracted from the documents
   • Use the search, sort, and filter options to navigate the timeline

3. Chat with Timeline:

   • Switch to the Chat tab to ask questions about the timeline
   • The AI will respond based on the information extracted from your documents

Technical Details

• Frontend: React with Chakra UI
• Backend: Node.js with Express
• Document Processing: pdf-parse and docx-parser
• LLM Integration: node-llama-cpp for local LLM inference
• File Upload: multer for handling multipart/form-data

Limitations

• The quality of date extraction depends on the clarity of dates in the documents
• Processing large documents may take several minutes depending on your hardware
• The LLM requires sufficient RAM to run efficiently (at least 8GB recommended)

License

MIT License