pyLAVA Setup and Requirements

Setting Up a Virtual Environment

We recommend using a virtual environment to isolate dependencies for this project. While venv is the easiest option since it’s part of the standard Python library, you are free to use conda or any other environment management tool you prefer.

Create a virtual environment:

python -m venv ~/.venv/pyLAVA

Activate the environment:

source ~/.venv/pyLAVA/bin/activate

Verify that Python now points to the virtual environment:

which python
# Expected output:
# ~/.venv/pyLAVA/bin/python

Accessing pyLAVA

On the NAS Cluster

On the NAS cluster, pylava is provided as part of the LAVA modules. Information on loading the module is provided in the LAVA Installation Guide for NAS Users. Loading one of the LAVA modules will append the PYTHONPATH environment variable to include the path to the pylava package. This will provide python access to pylava without needing to install it in your own environment.

External Installation

To obtain pylava on a machine outside the NAS Cluster, first download the correct LAVA tarball by following the LAVA Installation Guide for External Users. Once extracted, the directory structure will look like this:

LAVA/
├── bin/
├── lib/
│   └── pylava/
│       ├── pylava/        # Actual Python package
│       └── ...
└── ...

Thus, the pylava package is available as an installed package in /path/to/LAVA/lib/pylava/pylava.

Method 1 : `pip` install

Start by navigating to the pyLAVA root folder, LAVA/lib/pylava. From this directory using pip should add it to your python, environment.

pip install .

Note

If you have permission errors when installing try changing the permission of the LAVA tarball directories to allow reading and writing.

Method 2 : Modify the `PYTHONPATH`

To make pylava accessible in your Python environment, add its parent directory to your PYTHONPATH:

export PYTHONPATH=$PYTHONPATH:/path/to/LAVA/lib/pylava

To verify the installation, run:

python -m pip show pylava

This will display details like version, location, and dependencies, confirming that Python can find the package.

Method 3 : Append the path in your Python scripts

If you prefer not to modify PYTHONPATH globally, add the following lines at the top of your Python script:

import sys
sys.path.append("/path/to/LAVA/lib/pylava")

Requirements for pyLAVA

The pylava package is not shipped with the necessary dependencies. You will need to install the required packages manually. The required packages depend on the type of case you plan to run: simply running aerodynamic analyses through pylava requires fewer packages than running shape optimization cases. We provide details below.

Aerodynamic Analysis Requirements

In addition to pylava itself, the following python packages are required for running the aerodynamic analysis cases (i.e. non-optimization cases).

Package Name	Package Site	Description	Tested Version
`numpy`	link	Fundamental package for array computing in Python	2.4.2
`pyYAML`	link	YAML parser and emitter for Python	6.0.3
`mpi4py`	link	Python bindings for MPI (only required if running MPI)	4.0.0

Optimization Requirements

In addition to pylava itself, the following python packages are required for running shape optimization cases through the pylava package. Some of these can be installed via pip, while others need to be installed from their respective GitHub repositories. Please visit the provided links for installation instructions.

Package Name	Package Site	Description	Tested Version
`numpy`	link	Fundamental package for array computing in Python	2.4.2
`pyYAML`	link	YAML parser and emitter for Python	6.0.3
`pandas`	link	Powerful data structures for data analysis, time series, and statistics	3.0.0
`tabulate`	link	Pretty-print tabular data	0.9.0
`pyspline`	link	Python tools for B-spline curves, surfaces, volumes	1.5.3
`pygeo`	link	Object-oriented geometry framework for MDO and CAD.	1.17.0
`pyoptsparse`	link	Framework for efficient large-scale nonlinear optimization	2.14.5
`mpi4py`	link	Python bindings for MPI	4.0.0

If you plan to vizualize the optimization history using pyoptsparse's optview, you will need to install the following packages

Package Name	Package Site	Description	Tested Version
`pyoptsparse.optview`	link	Interactive visualization of optimization history	2.14.5

OpenMDAO-Based Optimization Requirements

For connecting to OpenMDAO, the following packages are required,

Package Name	Package Site	Description	Tested Version
`numpy`	link	Fundamental package for array computing in Python	2.4.2
`pyYAML`	link	YAML parser and emitter for Python	6.0.3
`openMDAO`	link	Multidisciplinary optimization framework	3.42.0
`MPhys`	link	Multiphysics framework built on OpenMDAO	2.0.0

Additionally, for MPI support, you will require the following packages

Package Name	Package Site	Description	Tested Version
`mpi4py`	link	Python bindings for MPI	4.0.0
`petsc4py`	link	PETSc for Python	3.24.2