CodeStyleDiffuser πŸš€

CodeStyleDiffuser is an AI-powered code rewriting system that transforms source code to match different developer coding stylesβ€”while preserving logic and behavior.

Supported styles include:

  • Clean (readable, modern)
  • Competitive (compact, fast-paced)
  • Enterprise (strict, verbose)
  • Legacy (wide indents, older conventions)

The project is designed for developer tools, refactoring systems, and AI coding assistants, and is ready for GitHub, Hugging Face, and Gradio Spaces.


πŸ” What Problem Does It Solve?

Different developers and teams follow different coding styles.
CodeStyleDiffuser allows you to:

  • Convert code between styles instantly
  • Standardize codebases
  • Adapt code to team or project conventions
  • Prepare code for reviews, contests, or enterprise systems

✨ Key Features

  • 🧠 Style-aware code rewriting
  • 🎨 Multiple predefined style profiles
  • 🧩 Logic-preserving transformations
  • ⚑ Lightweight, fast baseline
  • πŸ€— Hugging Face–ready pipeline
  • πŸŽ›οΈ Gradio web demo included
  • πŸ§ͺ Test-covered core components

πŸ“‚ Project Structure

codestyle-diffuser/
β”œβ”€β”€ config/
β”œβ”€β”€ data/
β”œβ”€β”€ src/
β”œβ”€β”€ training/
β”œβ”€β”€ pipelines/
β”œβ”€β”€ scripts/
β”œβ”€β”€ tests/
β”œβ”€β”€ notebooks/
β”œβ”€β”€ app.py
β”œβ”€β”€ README.md
β”œβ”€β”€ model_card.md
β”œβ”€β”€ requirements.txt
└── LICENSE

βš™οΈ Installation

pip install -r requirements.txt

πŸš€ Quick Usage

from src.inference import CodeStylePipeline

pipeline = CodeStylePipeline()

code = """
def add(a,b):
 return a+b
"""

styled_code = pipeline(code, style="enterprise")
print(styled_code)

πŸŽ›οΈ Gradio Demo

Run locally:

python app.py

🧠 How It Works

  1. Code Parsing
  2. Style Encoding
  3. Style Diffusion
  4. Rewrite Engine
Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support

Evaluation results