nirmal 1.0.0

AI-powered code generation streamed live to your terminal.

nirmal ⚡

AI-powered code generation, streamed live to your terminal.

import nirmal

nirmal.make("create a snake game in pygame")

Code starts appearing instantly — character by character — as the AI writes it in real time.

---

Installation

pip install nirmal

Requires Python 3.8+.

---

Quick Start

import nirmal

# Generate a complete script from a single sentence
nirmal.make("build a REST API with FastAPI that has CRUD routes for a todo list")

That's it. The generated code streams directly into your terminal.

---

More Examples

import nirmal

# Games
nirmal.make("create a tetris clone using pygame")

# Web
nirmal.make("build a flask web app with user login and a dashboard")

# Scripts
nirmal.make("write a python script that watches a folder and renames files by date")

# Data
nirmal.make("create a pandas script that reads a CSV, cleans missing values, and plots a bar chart")

# Automation
nirmal.make("write a selenium script that logs into a website and scrapes product prices")

---

How It Works

Your Python code
     │
     ▼
nirmal.make("your prompt")
     │
     ▼  POST /nirmal/generate.php
PHP Backend (Hostinger)
     │
     ▼  Groq API  (llama-3.3-70b-versatile)
Streaming AI response
     │
     ▼  Chunk by chunk
Live terminal output ✓

1. You call nirmal.make() with a plain-English description.
2. The package sends a POST request to the PHP backend.
3. The backend forwards the prompt to Groq with streaming enabled.
4. Each token is forwarded back to Python as it arrives.
5. Python prints each chunk live — no waiting for the full response.

---

Configuration

If you host your own backend, edit nirmal/config.py:

BACKEND_URL    = "https://your-domain.com/nirmal/generate.php"
REQUEST_TIMEOUT = 60  # seconds

---

Package Structure

nirmal/
├── nirmal/
│   ├── __init__.py   ← public API (make)
│   ├── client.py     ← streaming HTTP client
│   ├── config.py     ← endpoint + timeout settings
│   └── utils.py      ← markdown cleanup helper
├── backend/
│   ├── generate.php  ← PHP streaming proxy to Groq
│   └── .htaccess     ← security rules for Hostinger
├── pyproject.toml    ← pip package config
├── README.md
└── LICENSE

---

Publishing to PyPI

1. Install build tools

pip install build twine

2. Build the distribution

cd nirmal-project
python -m build

This creates dist/nirmal-1.0.0.tar.gz and dist/nirmal-1.0.0-py3-none-any.whl.

3. Upload to PyPI

twine upload dist/*

You'll need a PyPI account and an API token.

4. Install from PyPI

pip install nirmal

---

Deploying the PHP Backend on Hostinger

1. Connect via FTP or File Manager

Log in to Hostinger → hPanel → File Manager (or use FileZilla).

2. Upload the backend files

Upload backend/generate.php and backend/.htaccess to:

public_html/nirmal/

3. Set your Groq API key

Open generate.php and replace the placeholder:

define('GROQ_API_KEY', 'your-real-groq-api-key-here');

Get your key at console.groq.com.

4. Test the endpoint

curl -X POST https://hub-5.angelthesis.com/nirmal/generate.php \
     -d "prompt=print hello world in python"

You should see code stream back in your terminal.

---

Security Recommendations

Concern	Recommendation
API key exposure	Never commit generate.php with a real key. Use environment variables if possible.
Open access	Add a shared secret token validated in generate.php (see code comments).
Rate limiting	Hostinger plans support .htaccess rate limiting; see .htaccess comments.
HTTPS	Always use HTTPS — Hostinger provides free SSL via Let's Encrypt.
Input length	generate.php caps prompt length at 2000 characters server-side.

---

License

MIT — see LICENSE.