pyproject.toml

[build-system]
requires = ["setuptools>=61.0", "wheel"]
build-backend = "setuptools.build_meta"

[project]
name = "dflash-mlx-universal"
version = "0.2.0"
description = "Universal DFlash block diffusion speculative decoding for MLX — any model family on Apple Silicon"
readme = "README.md"
license = {text = "MIT"}
authors = [
    {name = "Raaz Kumar"},
    {name = "Community Contributors"},
]
classifiers = [
    "Development Status :: 4 - Beta",
    "Intended Audience :: Developers",
    "Intended Audience :: Science/Research",
    "License :: OSI Approved :: MIT License",
    "Programming Language :: Python :: 3",
    "Programming Language :: Python :: 3.9",
    "Programming Language :: Python :: 3.10",
    "Programming Language :: Python :: 3.11",
    "Programming Language :: Python :: 3.12",
    "Topic :: Scientific/Engineering :: Artificial Intelligence",
    "Environment :: MacOS X",
    "Operating System :: MacOS :: MacOS X",
]
keywords = [
    "mlx", "llm", "speculative-decoding", "diffusion", "dflash", 
    "inference", "apple-silicon", "qwen3", "llama", "mistral", "gemma",
    "block-diffusion", "text-generation", "inference-acceleration"
]
requires-python = ">=3.9"
dependencies = [
    "mlx>=0.25.0",
    "mlx-lm>=0.24.0",
    "transformers>=4.57.0",
    "torch>=2.9.0",
    "safetensors>=0.4.0",
    "huggingface-hub>=0.25.0",
    "datasets>=2.14.0",
    "numpy>=1.24.0",
]

[project.optional-dependencies]
server = [
    "fastapi>=0.100.0",
    "uvicorn>=0.23.0",
]
dev = [
    "pytest>=7.0.0",
    "pytest-cov>=4.0.0",
    "black>=23.0.0",
    "ruff>=0.1.0",
]

[project.urls]
Homepage = "https://huggingface.co/tritesh/dflash-mlx-universal"
Repository = "https://huggingface.co/tritesh/dflash-mlx-universal"
Documentation = "https://huggingface.co/tritesh/dflash-mlx-universal/blob/main/README.md"
Issues = "https://huggingface.co/tritesh/dflash-mlx-universal/discussions"

[project.scripts]
dflash-mlx-serve = "dflash_mlx.serve:main"
dflash-mlx-convert = "dflash_mlx.convert:main"

[tool.setuptools.packages.find]
where = ["."]
include = ["dflash_mlx*", "tests*", "examples*"]

[tool.black]
line-length = 100
target-version = ['py311']

[tool.ruff]
line-length = 100
select = ["E", "F", "W", "I"]
ignore = ["E501"]

# ──────────────────────────────────────────────────────────────────────────────
# UV (Ultra-fast Python package manager) configuration
# ──────────────────────────────────────────────────────────────────────────────

[tool.uv]
# Use the project Python version
python-preference = "only-managed"

# Default environment
default-groups = ["dev"]

# Allow pre-releases for MLX (which often has beta releases)
prerelease = "allow"

# Resolution strategy
resolution = "highest"

# Cache directory (default is fine, but can be overridden)
# cache-dir = "~/.cache/uv"

[tool.uv.pip]
# Use PEP 517 builds for packages that need compilation
no-build-isolation = false

[tool.uv.sources]
# Pin mlx-lm to a known-good version range if needed
# mlx-lm = { url = "https://files.pythonhosted.org/packages/..." }

[dependency-groups]
dev = [
    "pytest>=7.0.0",
    "pytest-cov>=4.0.0",
    "black>=23.0.0",
    "ruff>=0.1.0",
    "mypy>=1.5.0",
]
server = [
    "fastapi>=0.100.0",
    "uvicorn>=0.23.0",
]