pyproject.toml

[project]
name = "ttpy"
version = "2.0.0rc3"
description = "Tensor Train toolbox: TT/QTT decompositions, cross approximation, AMEn solvers — pure Python, numpy/torch backends"
readme = "README.md"
requires-python = ">=3.10"
license = "MIT"
license-files = ["LICENSE"]
authors = [{ name = "Ivan Oseledets" }]
keywords = ["tensor train", "TT", "QTT", "low-rank", "cross approximation", "AMEn"]
classifiers = [
    "Development Status :: 4 - Beta",
    "Intended Audience :: Science/Research",
    "Operating System :: OS Independent",
    "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",
    "Topic :: Scientific/Engineering :: Mathematics",
]
dependencies = [
    "numpy>=1.24",
    "scipy>=1.10",
    "einops>=0.7",
]

[project.optional-dependencies]
torch = ["torch>=2.2"]
# Compiled kernels for the innermost AMEn loops. numba ships wheels, so this
# still needs no compiler on the user's machine; without it the package falls
# back to the numpy path and every result is identical (tests/test_fast.py).
fast = ["numba>=0.59"]
test = ["pytest>=7", "pytest-xdist"]

[project.urls]
Homepage = "https://github.com/oseledets/ttpy"

[build-system]
requires = ["hatchling"]
build-backend = "hatchling.build"

[tool.hatch.build.targets.wheel]
packages = ["tt"]

[tool.pytest.ini_options]
testpaths = ["tests"]
filterwarnings = ["error::DeprecationWarning:tt.*"]