depot/third_party/nixpkgs/pkgs/development/libraries/galario/default.nix

79 lines
2.5 KiB
Nix
Raw Normal View History

{ lib, stdenv
, fetchzip
, fetchFromGitHub
, cmake
, fftw
, fftwFloat
, enablePython ? false
, pythonPackages ? null
, llvmPackages
}:
let
# CMake recipes are needed to build galario
# Build process would usually download them
great-cmake-cookoff = fetchzip {
url = "https://github.com/UCL/GreatCMakeCookOff/archive/v2.1.9.tar.gz";
sha256 = "1yd53b5gx38g6f44jmjk4lc4igs3p25z6616hfb7aq79ly01q0w2";
};
in
stdenv.mkDerivation rec {
pname = "galario";
version = "1.2.2";
src = fetchFromGitHub {
owner = "mtazzari";
repo = pname;
rev = "v${version}";
sha256 = "0dw88ga50x3jwyfgcarn4azlhiarggvdg262hilm7rbrvlpyvha0";
};
nativeBuildInputs = [ cmake ];
buildInputs = [ fftw fftwFloat ]
++ lib.optional enablePython pythonPackages.python
++ lib.optional stdenv.isDarwin llvmPackages.openmp
;
propagatedBuildInputs = lib.optional enablePython [
pythonPackages.numpy
pythonPackages.cython
pythonPackages.pytest
];
checkInputs = lib.optional enablePython [ pythonPackages.scipy pythonPackages.pytest-cov ];
preConfigure = ''
mkdir -p build/external/src
cp -r ${great-cmake-cookoff} build/external/src/GreatCMakeCookOff
chmod -R 777 build/external/src/GreatCMakeCookOff
'';
preCheck = ''
${if stdenv.isDarwin then "export DYLD_LIBRARY_PATH=$(pwd)/src/" else "export LD_LIBRARY_PATH=$(pwd)/src/"}
${if enablePython then "sed -i -e 's|^#!.*|#!${stdenv.shell}|' python/py.test.sh" else ""}
'';
doCheck = true;
postInstall = lib.optionalString (stdenv.isDarwin && enablePython) ''
install_name_tool -change libgalario.dylib $out/lib/libgalario.dylib $out/lib/python*/site-packages/galario/double/libcommon.so
install_name_tool -change libgalario_single.dylib $out/lib/libgalario_single.dylib $out/lib/python*/site-packages/galario/single/libcommon.so
'';
meta = with lib; {
description = "GPU Accelerated Library for Analysing Radio Interferometer Observations";
longDescription = ''
Galario is a library that exploits the computing power of modern
graphic cards (GPUs) to accelerate the comparison of model
predictions to radio interferometer observations. Namely, it
speeds up the computation of the synthetic visibilities given a
model image (or an axisymmetric brightness profile) and their
comparison to the observations.
'';
homepage = "https://mtazzari.github.io/galario/";
license = licenses.lgpl3;
maintainers = [ maintainers.smaret ];
platforms = platforms.all;
};
}