datafaker-ai 0.1.3

Generate synthetic datasets from natural language using Gemini 1.5 Flash

🧬 Gemini Dataset Generator

Generate synthetic datasets from natural language descriptions using Google's Gemini 1.5 Flash model. Ideal for data science, machine learning prototyping, and testing workflows with customizable, structured synthetic data.

---

✨ Features

• ⚡ Powered by Gemini 1.5 Flash (fast + cost-effective)
• 🧠 Natural language prompt → structured data
• 📦 Output in pandas, CSV, and JSON
• 🧪 Optional edge case injection for testing
• 🧾 Save datasets to local disk
• 🔐 Easy API key setup using .env

---

📦 Installation

Install from PyPI:

pip install gemini-dataset-generator


Or clone manually:


git clone https://github.com/ahsanraza1457/deepfaker_ai.git
cd deepfaker_ai
pip install -r requirements.txt

🔐 Setup
Create a .env file in the root directory of your project
GEMINI_API_KEY=your_google_generativeai_api_key


🚀 Usage
from generator import generate_dataset

df = generate_dataset(
    description="Customer name, email, age, and signup date",
    num_samples=50
)

print(df.head())



Save Output as CSV or JSON
generate_dataset(
    description="IoT device logs with timestamp, device_id, temperature",
    num_samples=100,
    save_as='csv'  # Options: 'csv', 'json', 'both'
)



🗂 Project Structure
├── generator/
│   ├── __init__.py
│   ├── generator.py           # Main interface
│   ├── formatter.py           # Formats model output
│   ├── prompts.py             # Builds prompt from description
│   ├── edge_case_handler.py   # Injects edge cases
│
├── .env                      
├── README.md
├── requirements.txt
└── setup.py / pyproject.toml