Building from source¶

Note

End users should not need this page. The published PyPI wheels are expected to cover the vast majority of use cases — see Basic installation. Build from source if you are contributing to NEML2 itself, debugging a build flavor the wheels don’t ship, or experimenting with a custom LibTorch.

Quick start¶

git clone -b main https://github.com/applied-material-modeling/neml2.git
cd neml2
pip install -e ".[dev]" -v

This drives scikit-build-core to build the bundled C++ runtime and lays down an editable Python install plus everything the dev workflow needs (pytest, pre-commit, sphinx, …).

CMake presets¶

For pure C++ development (the wheel build is invoked separately by pip), CMakePresets.json defines two presets, each serving a different purpose:

dev — the day-to-day build preset

Debug build of libneml2_aoti (the compiled-model runtime). The pybind extension modules are not included (NEML2_WHEEL=OFF); use pip install -e ".[dev]" to rebuild those. This is the preset for iterating on C++ sources:

cmake --preset dev -S .
cmake --build --preset dev

cc — a configure-only preset for tooling

Configures with CMAKE_EXPORT_COMPILE_COMMANDS=ON and NEML2_WHEEL=ON so the resulting compile_commands.json covers both the C++ runtime sources and the pybind extension .cpp files. A compile_commands.json symlink is created in the repo root for clangd / clang-tidy / other static-analysis tools to pick up. This preset is configure-only; there is no corresponding build step.

cmake --preset cc -S .

One-off CMake variable overrides¶

The presets cover the vast majority of cases. For a single-shot override (for example, switching the build type or pointing at a custom LibTorch), append -D<NAME>=<VALUE> to the configure line in the usual way:

cmake --preset dev -DCMAKE_BUILD_TYPE=RelWithDebInfo -S .

Custom LibTorch¶

If you want to build NEML2 against a libtorch other than the one shipped by the active Python environment’s torch package, set torch_ROOT before configuring:

cmake --preset dev -Dtorch_ROOT=/path/to/libtorch -S .

Equivalent environment variables and the active Python’s torch site-packages are also consulted; see the bundled Findtorch.cmake module for the full discovery procedure.

Running the test suite¶

See Testing in the contributing guide.