hmnshudhmn24/legal-policy-summarizer-nlp

🏛️ Legal / Policy Text Summarizer NLP

An advanced Transformer-based NLP model that simplifies legal, governmental, and policy documents into three easy-to-understand summary formats:

🔹 3-line summary

🔹 1-paragraph summary

🔹 Bullet points (3–7 bullets)

This project includes the full ML pipeline: preprocessing, PDF extraction, dataset creation, training, evaluation, inference, FastAPI deployment, Gradio UI, tests, and a HuggingFace model card.

🚀 Features

✔ Summarizes long policies, laws, govt documents
✔ Output styles: 3line, paragraph, bullets
✔ Full training/evaluation pipeline
✔ Works with PDFs
✔ Built on google/flan-t5-base
✔ Apache 2.0 licensed
✔ HuggingFace-ready metadata

📁 Project Structure

legal-policy-summarizer-nlp/
├── data/
├── src/
├── tests/
├── app/
├── notebooks/
├── huggingface/
├── model/
├── README.md
├── LICENSE
└── requirements.txt

📦 Installation

python -m venv venv
source venv/bin/activate
pip install -r requirements.txt

📄 Dataset Preprocessing

python -m src.dataset_preprocessing --input data/raw/dataset.csv --output data/processed/dataset_clean.jsonl

🏋️ Training

python -m src.train

🧪 Evaluation

python -m src.evaluate

🤖 Inference

from src.inference import summarize
print(summarize("policy text...", mode="paragraph"))

🌐 API (FastAPI)

uvicorn app.api:app --reload --port 8000

🎨 Gradio UI

python app/ui.py