Sabhya23/legal_summarizer

Legal Document Summarizer (North Indian Focus)

A production-ready NLP system designed to summarize legal documents, specifically court judgments and policies related to environmental and deforestation cases in North India.

Features

• —Input Processing: Extract and clean text from PDF documents using PyMuPDF.
• —Section Segmentation: Automatically identifies Facts, Arguments, Judgment, and Reasoning.
• —Abstractive Summarization: Uses facebook/bart-large-cnn to generate concise, human-like summaries.
• —Extractive Summarization: Identifies key sentences using BERT-based ranking.
• —Domain Intelligence:
• —Detects North Indian states (Delhi, Haryana, Punjab, UP).
• —Highlights environmental impact keywords (Deforestation, Tree felling, etc.).
• —Extracts core legal principles.
• —Structured Output: Returns a detailed JSON object tailored for legal assistants.

Tech Stack

• —Backend: FastAPI
• —NLP: Transformers (Hugging Face), Sentence-Transformers, NLTK
• —Vector Math: Scikit-learn, NumPy
• —Containerization: Docker

Setup Instructions

Local Setup

1. 1.Clone the repository.
2. 2.Install Python 3.10+.
3. 3.Install dependencies:

   pip install -r requirements.txt

1. 1.Run the application:

   python main.py

The API will be available at http://localhost:8000.

Docker Setup

1. 1.Build the image:

   docker build -t legal-summarizer .

1. 1.Run the container:

   docker run -p 8000:8000 legal-summarizer

API Documentation

POST /summarize

Input: multipart/form-data

• —file: PDF Document

Output: JSON

{
  "case_title": "Delhi High Court Order - 2024",
  "court": "Delhi High Court",
  "year": "2024",
  "facts": "...",
  "legal_issues": "...",
  "arguments": "...",
  "judgment": "...",
  "reasoning": "...",
  "final_verdict": "...",
  "environmental_impact": "deforestation, tree felling",
  "detected_states": ["Delhi"],
  "legal_principles": ["The principle of absolute liability..."],
  "extractive_summary": "..."
}

Example Usage (cURL)

curl -X 'POST' \
  'http://localhost:8000/summarize' \
  -H 'accept: application/json' \
  -H 'Content-Type: multipart/form-data' \
  -F 'file=@your_legal_document.pdf;type=application/pdf'

Project Structure

• —main.py: FastAPI entry point.
• —utils/: Core logic modules.
• —pdf_parser.py: PDF extraction.
• —cleaner.py: Text preprocessing.
• —summarizer.py: BART summarization.
• —extractor.py: BERT extractive ranking.
• —intelligence.py: Domain entity detection.
• —data/: Temporary storage for processing.
• —models/: Fine-tuned model directory.