pyproject.toml

[project]
name = "pysail"
version = "0.2.0"
description = "Sail Python library"
authors = [
    { name = "LakeSail", email = "hello@lakesail.com" },
]
readme = "README.md"
license = { file = "LICENSE" }
# Spark 3.5 does not support Python 3.12 yet.
requires-python = ">=3.8,<3.12"
dependencies = []
keywords = ["data", "big data", "sql", "spark", "pyspark", "arrow", "datafusion"]
classifiers = [
    "Development Status :: 4 - Beta",
    "Environment :: Console",
    "Intended Audience :: Developers",
    "Intended Audience :: Science/Research",
    "License :: OSI Approved :: Apache Software License",
    "Operating System :: OS Independent",
    "Programming Language :: Python",
    "Programming Language :: Python :: 3",
    "Programming Language :: Python :: 3 :: Only",
    "Programming Language :: Python :: 3.8",
    "Programming Language :: Python :: 3.9",
    "Programming Language :: Python :: 3.10",
    "Programming Language :: Python :: 3.11",
    "Programming Language :: Rust",
    "Topic :: Scientific/Engineering",
]

[project.optional-dependencies]
spark = [
    "pyspark[connect]>=3.5,<4.0",
    "duckdb>=1.0,<2.0",
]

[project.urls]
Homepage = "https://lakesail.com"
Documentation = "https://docs.lakesail.com/sail/latest/"
Repository = "https://github.com/lakehq/sail"
Issues = "https://github.com/lakehq/sail/issues"

[project.scripts]
sail = "pysail.cli:main"

[build-system]
requires = ["maturin>=1.0,<2.0"]
build-backend = "maturin"

# All Hatch environments are configured to be "detached",
# otherwise Hatch triggers a fresh build to install the project, which can be slow.
# To install the project as an editable package, run `hatch run [<env>:]maturin develop` instead.

# We use `uv` in CI for faster package installation, but we still use `pip` for local development,
# since the IDE may only support `pip`. Also, if we used `uv` for local development, the `pip` command
# would then be associated with the global Python installation in the command line. In this case,
# the developer may incorrectly run `pip` instead of `uv pip`, resulting in unintended modification
# to the global Python installation.

[tool.hatch.envs.default]
detached = true
path = ".venvs/default"
python = "3.11"
installer = "pip"
dependencies = [
    "pyspark[connect]==3.5.1",
    # We do not specify the extra dependencies for Ibis since we install PySpark explicitly.
    "ibis-framework~=9.4",
    "pyarrow-hotfix>=0.4,<1", # (required by Ibis PySpark)
    "rich>=12.4.4,<14", # (required by Ibis PySpark)
    "pytest~=8.3",
    "duckdb~=1.1",
]

[tool.hatch.envs.default.overrides]
env.CI.installer = "uv"

[tool.hatch.envs.default.scripts]
install-pysail = "{env:HATCH_UV} pip install pysail --no-index -f target/wheels --force-reinstall"

[tool.hatch.envs.docs]
detached = true
path = ".venvs/docs"
python = "3.11"
installer = "pip"
dependencies = [
    "sphinx~=8.0",
]

[tool.hatch.envs.docs.scripts]
install-pysail = "{env:HATCH_UV} pip install pysail --no-index -f target/wheels --force-reinstall"
build = "sphinx-build python/pysail/docs python/pysail/docs/_build -b json"

[tool.hatch.envs.docs.overrides]
env.CI.installer = "uv"

[tool.hatch.envs.test]
detached = true
path = ".venvs/test"
python = "3.11"
installer = "pip"
dependencies = [
    # PySpark dependencies
    # PySpark itself is not added here since we install our patched version.
    "numpy~=1.15",
    "scipy~=1.14",
    "pandas==2.0.3",
    "pyarrow==12.0.1",
    "flask==3.0.2",
    "matplotlib~=3.9",
    "plotly~=4.8",
    "scikit-learn~=1.1",
    "mlflow~=2.14.2",
    "memory-profiler==0.60.0",
    "grpcio>=1.56,<1.57",
    "grpcio-status>=1.56,<1.57",
    "googleapis-common-protos==1.56.4",
    # pytest and plugins
    "pytest~=8.3",
    "pytest-xdist~=3.6.1",
    "pytest-timeout~=2.3.1",
    "pytest-reportlog~=0.4.0",
    # Ibis and its test dependencies
    "ibis-framework~=9.4",
    "pyarrow-hotfix>=0.4,<1",
    "rich>=12.4.4,<14",
    "filelock>=3.7.0,<4",
    "hypothesis>=6.58.0,<7",
    "packaging>=21.3,<25",
    "requests>=2,<3",
    "tomli>=2.0.1,<3",
    "pytest-httpserver>=1.0.5,<2",
    "pytest-mock>=3.6.1,<4",
    "pytest-repeat>=0.9.1,<0.10",
    "pytest-snapshot>=0.9.0,<1",
]

[tool.hatch.envs.test.env-vars]
PYTHONPATH = "{root}/scripts/spark-tests"
IBIS_TESTING_DATA_DIR = "{root}/opt/ibis-testing-data"

[tool.hatch.envs.test.scripts]
install-pyspark = "{env:HATCH_UV} pip install opt/spark/python/dist/pyspark-3.5.1.tar.gz"
install-pysail = "{env:HATCH_UV} pip install pysail --no-index -f target/wheels --force-reinstall"

[tool.hatch.envs.test.overrides]
env.CI.installer = "uv"

[tool.hatch.build.targets.sdist]
packages = ["python/pysail"]

[tool.hatch.build.targets.wheel]
packages = ["python/pysail"]

[tool.ruff.lint.per-file-ignores]
"crates/**/*.py" = ["INP001"]
"python/pysail/docs/conf.py" = ["INP001"]
"python/pysail/examples/**/*.py" = ["T201"]
"python/pysail/tests/**/*.py" = ["S101"]
"scripts/**/*.py" = ["SLF001"]

[tool.pytest.ini_options]
testpaths = ["python"]

[tool.maturin]
python-source = "python"
module-name = "pysail._native"
manifest-path = "crates/sail-python/Cargo.toml"
features = [
    "pyo3/extension-module",
    "pyo3/abi3-py38",
    "pyo3/generate-import-lib",
]