tabicl/pyproject.toml at main · soda-inria/tabicl · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
[project]
name = "tabicl"
dynamic = ["version"]
description = "TabICL: A state-of-the-art tabular foundation model"
readme = "README.md"
authors = [
    { name = "Jingang Qu" },
    { name = "David Holzmüller" } ,
    { name = "Marine Le Morvan" } ,
    { name = "Gaël Varoquaux" } ,
]
license = { file = "LICENSE" }
requires-python = ">=3.10"
keywords = ["tabular data", "foundation model", "TabICL", "in-context learning"]
classifiers = [
    "Intended Audience :: Science/Research",
    "Intended Audience :: Developers",
    "Topic :: Scientific/Engineering",
    "Development Status :: 4 - Beta",
    "Programming Language :: Python",
    'Programming Language :: Python :: 3',
    "Programming Language :: Python :: 3.10",
    "Programming Language :: Python :: 3.11",
    "Programming Language :: Python :: 3.12",
    "Programming Language :: Python :: 3.13",
    "Programming Language :: Python :: 3.14",
]
dependencies = [
    "torch>=2.2",  # to enjoy Flash Attention v2
    "scikit-learn>=1.3.0",
    "numpy",
    "scipy",
    "einops>=0.7",  # this is only used for rope, so it could be replaced
    "psutil",  # used for memory calculation at inference
    "tqdm>=4.64.0",
    "huggingface-hub",
]

[project.optional-dependencies]
doc = [
  "sphinx",
  "sphinx-copybutton",
  "sphinx-gallery",
  "matplotlib>=3.10.3",
  "pydata-sphinx-theme",
  "sphinxext-opengraph",
  "skrub",
  "pandas",
  "shap>=0.42",
  "shapiq>=1.0",
  "numba",  # Add it here to facilitate the life of the installer
]
forecast = [
    "pandas>=2.1.2",
    "gluonts>=0.16.0",  # calendar features
    "statsmodels>=0.14.5",  # seasonal detection
    "matplotlib>=3.10.3",  # plotting
]
shap = [
    "shap>=0.42",
    "shapiq>=1.0",
    "matplotlib>=3.10.3",
    "numba",  # Add it here to facilitate the life of the installer
]
pretrain = [
    "transformers",  # only used in optim for schedules
    "xgboost",  # used in the TabICLv1 prior
    "wandb",  # logging in pre-training
]
finetune = [
    "transformers",  # get_scheduler reuses the pretraining cosine-warmup schedules
    "wandb",  # optional, only engaged when wandb_kwargs is passed
]
test = [
    "pandas",
]
all = [
    "tabicl[doc,forecast,shap,pretrain,finetune,test]",
]

[build-system]
requires = ["hatchling>=1.26.1"]  # https://github.com/pypa/hatch/issues/1818
build-backend = "hatchling.build"

[tool.hatch.version]
path = "src/tabicl/__about__.py"

[tool.hatch.build.targets.sdist]
only-include = ["src/tabicl", "figures", "LICENSE", "README.md"]

[tool.hatch.build.targets.wheel]
packages = ["src/tabicl"]

[tool.hatch.envs.default]
installer = "uv"
features = []

[tool.hatch.envs.hatch-test]
default-args = ["tests"]
installer = "uv"
features = ["test"]

[tool.hatch.envs.types]
extra-dependencies = [
    "mypy>=1.0.0",
]
[tool.hatch.envs.types.scripts]
check = "mypy --install-types --non-interactive {args:tabicl tests}"

[tool.coverage.run]
source = ["src"]
branch = true
parallel = true
omit = [
    "src/tabicl/__about__.py",
]

[tool.coverage.paths]
tabicl = ["src"]

[tool.coverage.report]
exclude_lines = [
    "no cov",
    "if __name__ == .__main__.:",
    "if TYPE_CHECKING:",
]

[tool.uv]
# gluonts pins an old numpy version which is not available for Python 3.14. The
# following override is a temporary workaround to avoid having uv build numpy
# from source on Python 3.14 and later.
override-dependencies = [
    "numpy>=2.4; python_version >= '3.14'",
    "numpy; python_version < '3.14'",
]