pyproject.toml

[project]
name = "benchkit"
version = "0.1.0"
description = "Benchmark your local LLMs with real eval suites"
readme = "README.md"
requires-python = ">=3.11"
license = "Apache-2.0"
dependencies = [
    "evalplus==0.3.1",
    "httpx>=0.28.1",
    "rich>=15.0.0",
    "python-dotenv>=1.2.3",
    "textual>=8.2.8",
]

[project.optional-dependencies]
# treejs-arena renders generated pages in headless Chromium. Install with
# `uv sync --extra browser`, then `uv run playwright install chromium`.
browser = ["playwright>=1.62.0"]

[project.scripts]
benchkit = "benchkit.cli:main"

[dependency-groups]
dev = ["pytest>=9.1.1", "pre-commit>=4.6.2", "ruff>=0.16.5"]

[build-system]
requires = ["setuptools>=84.0.0"]
build-backend = "setuptools.build_meta"

[tool.setuptools.packages.find]
where = ["src"]
include = ["benchkit*"]

[tool.setuptools.package-data]
benchkit = [
    "answer_key_guard.ts",
    "datasets/*.jsonl",
    "datasets/*.csv",
    "datasets/*.md",
    "datasets/*.txt",
    "git_surgery/**/*.sh",
    "git_surgery/**/*.md",
    "mc_viewer/dist/*",
    "pi_package/*.json",
    "templates/*.html",
    "tui/*.tcss",
]

[tool.ruff]
line-length = 88
target-version = "py311"
src = ["src", "tests"]

[tool.ruff.lint]
select = [
    "E",    # pycodestyle errors
    "W",    # pycodestyle warnings
    "F",    # pyflakes
    "I",    # import sorting
    "UP",   # pyupgrade
    "B",    # bugbear
    "C4",   # comprehensions
    "SIM",  # simplifications
    "RUF",  # ruff-specific
]
ignore = [
    "E501",    # line length is the formatter's job
    "RUF001",  # the TUI leans on box-drawing and other unicode glyphs
    "RUF003",
]

[tool.ruff.lint.per-file-ignores]
# Textual declares BINDINGS/COMMANDS as plain class-level lists by design.
"src/benchkit/tui/**" = ["RUF012"]

[tool.pytest.ini_options]
testpaths = ["tests"]
addopts = "-q"