bharat-pii 0.1.1

India-first PII detection & redaction library

bharat-pii 🇮🇳

India-first PII detection & redaction library for GenAI pipelines.

bharat-pii helps detect and redact Indian-specific sensitive data (PAN, Aadhaar, UPI, IFSC, credit cards, etc.) before sending data to LLMs, logs, or analytics systems.

It is built with a regex-first, deterministic core and designed to be extensible with ML / Presidio adapters in the future.

---

✨ Features

• 🇮🇳 India-first PII entities
• ⚡ Fast & deterministic (regex + validation)
• 🔒 Credit card detection with strict Luhn validation
• 🔌 Extensible detector architecture
• 🧱 LLM-safe (local by default)
• 🧪 Fully testable
• 📦 uv / PyPI friendly packaging

---

📌 Supported Entities

Entity	Example
PAN	ABCDE1234F
Aadhaar	2345 6789 0123
Indian Phone	[phone]
Credit Card	4111 1111 1111 1111
UPI ID	rahul@upi
IFSC	HDFC0001234

---

🛣️ Roadmap

Planned additions:

• Email
• Passport
• Driving License
• Entity severity & priority
• Strict vs loose detection modes
• Presidio (ML) adapter
• CLI tool
• PDF / OCR support

---

📦 Installation

Using pip:

pip install bharat-pii

Using uv:

uv pip install bharat-pii

---

🚀 Usage

Detect PII

from bharat_pii.api import detect

text = "Pay 500 to rahul@upi using card 4111 1111 1111 1111"
results = detect(text)
print(results)

---

Redact PII

from bharat_pii.api import redact

safe_text = redact("My IFSC is HDFC0001234 and Aadhaar is 2345 6789 0123")
print(safe_text)

Output:

My IFSC is ******1234 and Aadhaar is ********0123

---

🧠 Design Philosophy

• Regex-first for precision and predictability
• Validation where applicable (e.g. Luhn for cards)
• Pluggable detector architecture
• No cloud dependency
• Minimal and stable API

Ideal for:

• LLM pipelines
• Fintech & healthcare systems
• Logs & analytics
• Internal tools
• Compliance & PII redaction

---

🤝 Contributing

Contributions are welcome.
Please see CONTRIBUTING.md for guidelines.

---

📄 License

MIT License © Rahul Sharma