pyproject.toml

[build-system]
requires = ["setuptools>=75.3.2", "wheel"]
build-backend = "setuptools.build_meta"

[project]
name = "mtdata"
dynamic = ["version"]
description = "mtdata is a tool to download datasets for machine translation"
readme = "README.md"
#license = {text = "Apache-2.0"}
#license-files = [ "LICENSE" ]  # twine complaints about this;
license-files = [] #TODO: revisit this after twine is fixed
requires-python = ">=3.10"
authors = [
    { name = "Thamme Gowda", email = "tgowdan@gmail.com" }
]
classifiers = [
    "Development Status :: 5 - Production/Stable",
    "Intended Audience :: Developers",
    "Topic :: Utilities",
    "Topic :: Text Processing",
    "Topic :: Text Processing :: General",
    "Topic :: Text Processing :: Filters",
    "Topic :: Text Processing :: Linguistic",
    "Programming Language :: Python :: 3 :: Only"
]
keywords = [
    "machine translation",
    "datasets",
    "NLP",
    "natural language processing",
    "computational linguistics"
]
dependencies = [
    "requests~=2.32",
    "rich~=14.0",
    "portalocker~=3.2",
    "pybtex~=0.25",
    "ruamel.yaml~=0.18"
]

[project.urls]
homepage = "https://github.com/thammegowda/mtdata"
documentation = "https://github.com/thammegowda/mtdata"
repository = "https://github.com/thammegowda/mtdata"
#changelog = ""

[tool.setuptools.packages.find]
include = ["mtdata*"]  # ["*"] by default
exclude = ["tests*", "tmp*", "build*", "dist*", "crawler*", "*.egg-info*"]


[project.optional-dependencies]
hf = ["datasets>=4.0.0"]
xlsx = ["openpyxl>=3.1.0"]
test = [ "pytest", "pytest-cov[all]", "black", "isort", "mypy"]


[project.scripts]
mtdata = "mtdata.__main__:main"
mtdata-iso = "mtdata.iso.__main__:main"
mtdata-bcp47 = "mtdata.iso.bcp47:main"
mtdata-map = "mtdata.map:main"


[tool.setuptools.dynamic]
version = {attr = "mtdata.__version__"}