Skip to content

Latest commit

 

History

History
111 lines (75 loc) · 3.61 KB

INSTALL.md

File metadata and controls

111 lines (75 loc) · 3.61 KB

Installing MOOSE

Installing released version from PyPI using pip

MOOSE is available on PyPI. To install the latest release, run pip install pymoose.

Installing from a binary wheel using pip

Binary wheels for MOOSE are available on its github page under Releases. You can download a wheel suitable for your platform and install it directly with pip. The last three components of the wheel filename indicate what platform it was built for: *-{python-version}-{operating-system}_{architecture}.whl.

For example,

pymoose-4.1.0.dev0-cp312-cp312-manylinux_2_28_x86_64.whl

is built for CPython version 3.12 on Linux for 64 bit Intel CPU (x86_64). MOOSE also depends on GSL, and the specific version should match. This should be available in the release description. The above was with GSL 2.7. So you will need an environment with these.

The easiest way to setup a custom environment is with Anaconda (or similar tools like mamba, miniforge, micromamba, etc.) See https://docs.conda.io/projects/conda/en/latest/user-guide/install/linux.html to find how to install conda or its variants. The channel conda-forge provides binary gsl package for all major platforms.

To create an environment named moose with these using conda run (Note: in the commands below replace conda with mamba or micromamba if you are using one of those):

conda  create -n moose python=3.12 gsl numpy vpython matplotlib -c conda-forge

in the terminal, and then activate this environment with

conda activate moose

Now install the wheel file with

pip install pymoose-4.1.0.dev0-cp312-cp312-manylinux_2_28_x86_64.whl

assuming you have the file in the current directory.

Installing from source-code in github repository

To build MOOSE from source, you need meson, ninja, pkg-config, and python-setuptools. We recommend to use Python 3.9 or higher. To build and install with pip, you also need meson-python.

For platform specific instructions, see:

Briefly,

  1. Build environment: make sure that the following packages are installed on your system.

    • gsl-1.16 or higher.
    • python-numpy
    • pybind11 (if the setup fails to find pybind11, try running pip install pybind11[global])
    • python-libsbml
    • pyneuroml
    • clang compiler 18 or newer
    • meson
    • ninja
    • meson-python
    • python-setuptools
    • pkg-config

  2. Now use pip to download and install pymoose from the github repository.

$ pip install git+https://github.com/BhallaLab/moose-core --user

This will install the master branch of MOOSE. If you want a specific fork/branch, modify the address:

$ pip install git+https://github.com/subhacom/moose-core@fix495merge --user

will install the fix495merge branch from subhacom's fork of moose-core.

Post installation

You can check that moose is installed and initializes correctly by running:

$ python -c "import moose; ch = moose.HHChannel('ch'); moose.le()"

This should show

Elements under /
    /Msgs
    /clock
    /classes
    /postmaster
    /ch

Now you can import moose in a Python script or interpreter with the statement:

>>> import moose

Uninstall

To uninstall moose, run

$ pip uninstall pymoose

If you are building moose from source, make sure to get out of the source directory, or you may encounter a message like this:

Found existing installation: pymoose {version}
Can't uninstall 'pymoose'. No files were found to uninstall.