For some use cases and workflows, more flexibility is required than what can be offered through Neko's user interface or library. To accommodate these usage patterns, Neko provides a C-based API that enables simple integration of Neko into custom workflows and interactive computing environments such as Jupyter notebooks. On top of the API, Neko provides interfaces to different high-level languages, including Python (pyneko) and Julia (Neko.jl).

Both the API (neko_api) and high-level interfaces provide functionality for defining cases, setting up and controlling simulation workflows, creating callbacks for Neko (e.g., initial and inflow conditions), and methods for direct (zero-copy) access to the underlying data inside Neko.

Installation

The API is always compiled and installed together with Neko; no special configuration options are necessary to use the C-based API. An example on how to use the C interface is shown in examples/api/c/cylinder.c.

Python (pyneko)

To install the python interface, configure and install Neko with the flags --enable-shared and --enable-python

Finally set PYTHONPATH and library paths (e.g. LD_LIBRARY_PATH) to the correct paths under installation prefix. On some systems you might need to preload (e.g. LD_PRELOAD) the MPI library to avoid missing symbols.

An example on how to use the Python interface is shown in examples/api/pyneko/cylinder.py.

Julia (Neko.jl)

To interface Neko with Julia, first configure and install Neko with shared libraries enabled --enable-shared. Then install the Neko.jl packge using:

(@v1.11) pkg> add https://gitlab.com/ExtremeFLOW/Neko.jl

Finally, set library paths (e.g., LD_LIBRARY_PATH) to the correct paths under the installation prefix. On some systems, you might need to preload (e.g., LD_PRELOAD) the MPI library to avoid missing symbols.

An example on how to use the Julia interface is shown in examples/api/Neko.jl/cylinder.jl.