Legal Document Analyzer

Overview

The Legal Document Analyzer is a FastAPI application designed to analyze legal documents using a fine-tuned BERT model. The application provides endpoints for uploading PDF files and analyzing their content. It also includes a web interface for interacting with the application.

Main Features

• Upload and analyze legal documents (PDF)
• Extract text from PDF files
• Classify legal clauses using a fine-tuned BERT model
• Track experiments and log metrics using MLflow
• Web interface for uploading documents and viewing results
• API endpoints for document analysis and question answering

Setup

Dependencies

The project requires the following dependencies:

• Python 3.8+
• FastAPI
• Uvicorn
• Transformers
• Torch
• PyPDF2
• MLflow
• Docker
• DVC
• Datasets
• Python-Multipart

Installation

1. Clone the repository:

git clone https://github.com/harshitIIITD/legal-doc-analyzer.git
cd legal-doc-analyzer

2. Create a virtual environment and activate it:

python -m venv venv
source venv/bin/activate  # On Windows, use `venv\Scripts\activate`

3. Install the dependencies:

pip install -r requirements.txt

4. Set up MLflow:

mlflow server --backend-store-uri sqlite:///mlflow.db --default-artifact-root ./mlruns

Running the Application

1. Start the FastAPI application:

uvicorn app.main:app --reload

2. Open your browser and navigate to http://127.0.0.1:8000 to access the web interface.

Using the API

Analyze Document

Endpoint: POST /analyze

Upload a PDF document to analyze its content.

Request

curl -X POST "http://127.0.0.1:8000/analyze" -F "file=@path/to/document.pdf"

Response

{
  "filename": "document.pdf",
  "classification": [0, 1, 0, 1],
  "text_preview": "Extracted text preview...",
  "confidence": 0.95,
  "model_version": "1234567890abcdef",
  "timestamp": "2023-01-01T12:00:00"
}

Ask Question

Endpoint: POST /ask

Ask a question about the analyzed document.

Request

curl -X POST "http://127.0.0.1:8000/ask" -H "Content-Type: application/json" -d '{"question": "What is the main clause?"}'

Response

{
  "answer": "The main clause is..."
}

Architecture

Data Flow

Key Components

• app/main.py: Main FastAPI application
• app/routes/questions.py: Question answering endpoint
• app/models/legal_bert: Fine-tuned BERT model
• app/templates/index.html: Web interface template
• tests: Test cases for the application
• data/dvc.yaml: Data version control configuration