Add TabGAN - synthetic tabular data generation with GANs, Diffusion, and LLMs

[Diyago]

Add TabGAN — Synthetic Tabular Data Generation

TabGAN is a Python library for generating high-quality synthetic tabular data using multiple generative approaches through a unified API:

• CTGAN (Conditional Tabular GAN) for mixed data types
• ForestDiffusion (tree-based diffusion) for structured data
• GReaT (Large Language Models) for semantic dependencies

Key Features

• Unified API across GANs, Diffusion Models, and LLMs
• Adversarial filtering ensures distribution consistency
• Privacy metrics (DCR, NNDR, membership inference)
• Constraint enforcement (range, uniqueness, formula, regex)
• HTML quality reports with distribution comparisons
• sklearn TabGANTransformer for Pipeline integration
• 100K+ PyPI downloads, 115 tests, Apache 2.0

Paper: Tabular GANs for uneven distribution