blend2d-python 1.0.0

Blend2D Python Bindings (nanobind version)

Blend2D Python Bindings (nanobind version)

Python bindings for the Blend2D rendering engine using nanobind.

Overview

This project has been converted from the original Cython-based implementation to use nanobind, which offers:

• Improved performance
• Better compatibility with modern C++
• Simplified binding code
• Cleaner implementation

Requirements

• C++17 compatible compiler
• CMake 3.15 or newer
• Python 3.8 or newer
• NumPy

Installation

From PyPI

Pre-built wheels are available for:

• Windows (x86_64)
• macOS (x86_64, arm64/Apple Silicon)
• Linux (x86_64, aarch64/ARM64)
• WebAssembly (via Emscripten)

pip install blend2d-python

From Source

pip install .

Or for development:

pip install -e .

Usage

import numpy as np
from blend2d import Context, Image, Gradient, Path, Format

# Create an image
img = Image(480, 480, Format.PRGB32)

# Attach a rendering context to it
ctx = Context(img)

# Clear the image with white color
ctx.fill_all((1.0, 1.0, 1.0, 1.0))

# Create a linear gradient
gradient = Gradient.create_linear(0, 0, 480, 480)
gradient.add_stop(0.0, (1.0, 0.0, 0.0, 1.0))
gradient.add_stop(0.5, (0.0, 1.0, 0.0, 1.0))
gradient.add_stop(1.0, (0.0, 0.0, 1.0, 1.0))

# Create a path (a star, for example)
path = Path()
path.move_to(240, 80)
for i in range(5):
    angle = (i * 2.0 * np.pi / 5.0) + np.pi / 2.0
    path.line_to(240 + 160 * np.cos(angle), 240 + 160 * np.sin(angle))

# Fill the path with the gradient
ctx.set_fill_style(gradient)
ctx.fill_path(path)

# Save the image to a file
img.write_to_file("star.png")

Building Wheels

This project uses cibuildwheel to build wheels for multiple platforms:

• Windows (x86_64)
• macOS (x86_64, arm64/Apple Silicon)
• Linux (x86_64, aarch64/ARM64)
• WebAssembly (via Emscripten)

To build wheels locally:

pip install cibuildwheel
python -m cibuildwheel --platform auto

License

MIT License

Acknowledgements

This project was originally created by John Wiggins using Cython, and has been converted to use nanobind.