pyproject.toml

[build-system]
requires = ["setuptools>=80", "setuptools-scm>=10"]
build-backend = "setuptools.build_meta"

[project]
name = "structured-data-models"
dynamic = ["version"]
description = "Foundation Models for Structured Data"
readme = "README.md"
requires-python = ">=3.11"
license = "Apache-2.0"
license-files = [
    "LICENSE",
    "THIRD_PARTY_LICENSES.md",
    "sdm/models/tabiclv2/LICENSE",
    "sdm/models/tabfm/LICENSE",
    "sdm/models/timesfm3/LICENSE",
    "sdm/models/kumo/relational/NOTICE",
    "third_party/contributor-covenant/LICENSE",
    "third_party/pytorch/LICENSE",
    "third_party/tabpfn-extensions/LICENSE",
]
authors = [
    { name = "NVIDIA" },
]
maintainers = [
    { name = "NVIDIA" },
]
keywords = [
    "artificial intelligence",
    "foundation models",
    "machine learning",
    "nvidia",
    "schema",
    "structured data",
]
classifiers = [
    "Development Status :: 3 - Alpha",
    "Programming Language :: Python :: 3",
    "Programming Language :: Python :: 3.11",
    "Programming Language :: Python :: 3.12",
    "Programming Language :: Python :: 3.13",
    "Programming Language :: Python :: 3.14",
]
dependencies = [
    "torch>=2.7",
    "numpy",
    "pyarrow",
    "huggingface_hub",
    "safetensors",
    "typing_extensions",
]

[project.urls]
Download = "https://github.com/NVIDIA/structured-data-models/releases"
Homepage = "https://github.com/NVIDIA/structured-data-models"
Repository = "https://github.com/NVIDIA/structured-data-models"
Issues = "https://github.com/NVIDIA/structured-data-models/issues"

[dependency-groups]
test = [
    "pytest",
    "pytest-cov",
    "pandas",
    "cudf-cu13>=26.8; sys_platform == 'linux'",
    "pyg-lib; sys_platform == 'linux'",
    "pylibcugraph-cu13>=26.8; sys_platform == 'linux'",
    "sentence-transformers>=5.4",
]
doc = [
    "myst-parser",
    "pandas",
    "shibuya",
    "sphinx",
    "sphinx-copybutton",
]
dev = [
    { include-group = "test" },
    { include-group = "doc" },
]

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

[tool.setuptools_scm]
fallback_version = "0.1.0.dev0"

[tool.uv]
managed = true
package = true
find-links = ["https://data.pyg.org/whl/torch-2.14.0+cpu.html"]
exclude-newer = "7 days"

[tool.uv.exclude-newer-package]
cudf-cu13 = false
libcudf-cu13 = false
libcugraph-cu13 = false
libcuvs-cu13 = false
libkvikio-cu13 = false
libraft-cu13 = false
librmm-cu13 = false
pylibcudf-cu13 = false
pylibcugraph-cu13 = false
pylibraft-cu13 = false
rmm-cu13 = false
relbench = false
torch = false
triton = false

[tool.pytest.ini_options]
addopts = [
    "--capture=no",
    "--color=yes",
    "-vv",
    "--import-mode=importlib",
    "--strict-markers",
]
markers = [
    "cuda: tests that require a NVIDIA GPU",
]
testpaths = ["test"]
xfail_strict = true

[tool.coverage.run]
branch = false
relative_files = true

[tool.coverage.report]
exclude_also = [
    "pragma: no cover",
    "raise NotImplementedError",
    "except ImportError",
    "if __name__ == .__main__.:",
    "@(abc\\.)?abstractmethod",
]

[tool.coverage.html]
title = "Structured Data Models Coverage Report"
skip_covered = true
skip_empty = true

[tool.ty.src]
include = ["sdm", "test", "examples", "benchmark"]

[[tool.ty.overrides]]
include = ["examples/**", "benchmark/**"]

[tool.ty.overrides.rules]
unresolved-import = "ignore"
unresolved-attribute = "ignore"

[[tool.ty.overrides]]
include = ["sdm/_kernels/triton/**"]

[tool.ty.overrides.rules]
unresolved-import = "ignore"
unused-ignore-comment = "ignore"

[tool.ruff]  # https://docs.astral.sh/ruff/rules
src = ["pyproject.toml", "sdm", "test", "examples", "benchmark"]
line-length = 79
indent-width = 4
target-version = "py311"

[tool.ruff.lint]
# https://docs.astral.sh/ruff/rules/
select = [
    "A",    # flake8-builtins
    "B",    # bugbear
    "BLE",  # blind-except
    "C4",   # comprehensions
    "D",    # pydocstyle
    "E",    # pycodestyle errors
    "F",    # pyflakes
    "G",    # flake8-logging-format
    "I",    # import sorting
    "LOG",  # flake8-logging
    "NPY",  # numpy-specific rules
    "PIE",  # flake8-pie
    "PLC0415",  # pylint: imports should be at module scope
    "PLE",  # pylint errors
    "PT",   # pytest style
    "RET",  # return statement style
    "RUF",  # ruff-specific rules
    "SIM",  # simplification
    "T10",  # flake8-debugger
    "T20",  # flake8-print
    "TID252",  # flake8-tidy-imports: require absolute imports
    "TID253",  # flake8-tidy-imports: keep optional imports lazy
    "UP",   # pyupgrade
    "W",    # pycodestyle warnings
    "YTT",  # flake8-2020
]
ignore = [
    "B905",  # Ignore "zip() without explicit `strict`"
    "D100",  # Ignore "Missing docstring in public module"
    "D105",  # Ignore "missing docstring in magic method"
    "D107",  # Ignore "missing docstring in __init__"
    "D413",  # Ignore "Missing blank line after last section ("{name}")"
    "RUF022",  # Ignore sorting __all__; order controls public docs.
]

[tool.ruff.lint.isort]
known-first-party = ["benchmark", "sdm"]

[tool.ruff.lint.pydocstyle]
convention = "google"

[tool.ruff.lint.flake8-builtins]
ignorelist = ["id"]

[tool.ruff.lint.flake8-tidy-imports]
ban-relative-imports = "all"
banned-module-level-imports = [
    "cudf",
    "cupy",
    "graphviz",
    "pandas",
    "pyg_lib",
    "pylibcudf",
    "pytest",
    "sentence_transformers",
    "transformers",
    "tokenizers",
]

[tool.ruff.lint.per-file-ignores]
"**/__init__.py" = ["I"]
"benchmark/**/__init__.py" = ["D104"]
"sdm/_kernels/**/*.py" = ["D"]
"examples/**/*.py" = [
    "D101",  # Ignore "Missing docstring in public class"
    "D102",  # Ignore "Missing docstring in public method"
    "D103",  # Ignore "Missing docstring in public function"
    "T201",  # Allow print.
    "TID253",  # Example dependencies should be imported at module scope.
]
"benchmark/**/*.py" = [
    "D101",  # Ignore "Missing docstring in public class"
    "D102",  # Ignore "Missing docstring in public method"
    "D103",  # Ignore "Missing docstring in public function"
    "T201",  # Allow print.
    "TID253",  # Benchmark dependencies should be imported at module scope.
]
"test/**/*.py" = [
    "D101",  # Ignore "Missing docstring in public class"
    "D102",  # Ignore "Missing docstring in public method"
    "D103",  # Ignore "Missing docstring in public function"
    "D104",  # Ignore "Missing docstring in public package"
    "D106",  # Ignore "Missing docstring in public nested class"
    "TID253",  # Test dependencies should be imported at module scope.
]

[tool.ty.analysis]
allowed-unresolved-imports = [
    "cudf.**",
    "cupy.**",
    "pyg_lib.**",
    "pylibcudf.**",
    "graphviz.**",
    "pylibcugraph.**",
    "sentence_transformers.**",
]