Installation and quickstart

Requirements

Morana requires Python 3.12 or newer. Version 0.2.0 is available as a package distribution from PyPI, as a source release from GitHub, and from Zenodo under DOI 10.5281/zenodo.23118502. The concept DOI 10.5281/zenodo.22681315 identifies Morana across releases. See citing Morana for choosing between the version DOI and the concept DOI.

Supported platforms

Morana is developed and verified on Linux. Native Windows and macOS execution are not claimed as supported platforms. Windows users can use Windows Subsystem for Linux (WSL), which provides a Linux distribution and Bash environment on Windows. Run the Linux installation commands below inside that environment. WSL itself has not been separately verified for Morana.

User installation

Create an isolated virtual environment and install the exact package release:

python -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip
python -m pip install "morana==0.2.0"

The GitHub release remains the canonical source-release location and publishes the source archive checksum. To verify that archive before installation, compare its SHA-256 digest with the accompanying morana-0.2.0.tar.gz.sha256 file.

The installation brings in Morana’s runtime dependencies from pyproject.toml, including h5py, which supports the OpenMC runtime-MGXS material importer. OpenMC itself is not a Morana runtime dependency; the maintained transport comparison needs it for model construction, geometry checks, and CE/MGXS calculations. Python VTK is also unnecessary at runtime because Morana writes VTU and VTM files directly as VTK XML. Install ParaView or Python VTK only to inspect those files with an external tool.

Development installation

Contributors and maintainers use the project Conda environment. It installs Morana in editable mode together with its test, documentation, and file-reader verification dependencies.

Create the environment from the repository root using the checked-in environment.yml:

conda env create -f environment.yml
conda activate morana-dev

Update an existing environment after environment.yml changes:

conda env update -n morana-dev -f environment.yml --prune
conda activate morana-dev

Changes under the repository’s src/morana/ directory are immediately importable. See the contributor workflow for required development checks.

Quickstart

The following reflected six-cell ring has the uniform analytic flux \(\phi=Q/\Sigma_a=50\) in every active cell:

"""The reflected six-cell-ring quickstart case."""

from morana.solvers.finite_volume import solve_fixed_source
from morana import (
    BoundaryCondition,
    BoundaryConditionSet,
    CrossSections,
    HexPlanarMesh,
    Material,
    MaterialMesh,
    MaterialSlice,
    ProblemConfiguration,
    UniformSource,
)

mesh = HexPlanarMesh(num_rings=2, pitch=10.0)
medium = Material(
    "medium",
    xs=CrossSections(
        D=[1.2],
        sigma_a=[0.02],
        sigma_s=[[0.0]],
        fission=None,
    ),
)

configuration = ProblemConfiguration(
    mesh=mesh,
    materials={"medium": medium},
    material_mesh=MaterialMesh.stack(
        (
            MaterialSlice.from_openmc_rings(
                mesh,
                [["medium"] * 6, ["0"]],
                height=1.0,
            ),
        )
    ),
    boundary=BoundaryConditionSet(BoundaryCondition.reflective().globally()),
    source=UniformSource([1.0]),
)

result = solve_fixed_source(configuration)
print(result.flux_layer(0)[0])  # [50. 50. 50. 50. 50. 50.]

The complete executable is available as examples/quickstart.py in the source repository.

Even a one-layer mesh is a finite hex-z layer. The global reflective condition above covers its radial, bottom, top, and excluded-region faces. To model a two-dimensional radial problem with different radial conditions, assign reflective bottom and top conditions explicitly, then assign the radial condition. The six active cells are the outer ring; the center entry "0" is the built-in excluded-region key.

Capability boundaries

Before building a larger calculation, review the modeling and solver workflow and its capability boundaries.

Continue with the maintained examples, then consult geometry and indexing and the modeling and solver workflow. For nonuniform boundary assignments, use boundary conditions and face selection.