# Copyright 1999-2026 Gentoo Authors # Distributed under the terms of the GNU General Public License v2 EAPI=8 DISTUTILS_USE_PEP517=setuptools PYTHON_COMPAT=( python3_{11..14} ) inherit distutils-r1 pypi DESCRIPTION="Stack AV's Triton kernel repository for GPU-accelerated ML" HOMEPAGE=" https://github.com/stackav-oss/conch https://pypi.org/project/conch-triton-kernels/ " LICENSE="Apache-2.0" SLOT="0" KEYWORDS="~amd64" # Hard runtime dep is just numpy. Triton + torch are pulled in via # upstream's [cpu]/[cuda]/[rocm]/[xpu] extras and accessed lazily inside # kernel functions — calling a kernel without triton installed raises # ImportError, which is the same behavior we'd get from the wheel. # triton itself is not in any tree we depend on, so we can't declare it. RDEPEND=" >=dev-python/numpy-1.26.4[${PYTHON_USEDEP}] " BDEPEND=" dev-python/setuptools-scm[${PYTHON_USEDEP}] " export SETUPTOOLS_SCM_PRETEND_VERSION=${PV}