7. Installation

Note

NEMO is available on https://github.com/teuben/nemo

7.1. Installation from github

Installation is normally done via github. Here is a simple example, just 3 lines in your (bash) shell, using a configurable helper script:

wget https://teuben.github.io/nemo/install_nemo.sh
bash install_nemo.sh  nemo=$HOME/opt/nemo yapp=pgplot bench5=1 python=1
source  $HOME/opt/nemo/nemo_start.sh

where the arguments to the install_nemo.sh script are optional, but a few are given to show some often use non-defaults. See that script for more details.

A more manual install, bypassing this script, can be:

git clone https://github.com/teuben/nemo
cd nemo
./configure --with-yapp=pgplot
make build check bench5 python
source nemo_start.sh

On a Mac with their new SIP protection, the --disable-shared flag needs to be added

git clone https://github.com/teuben/nemo
cd nemo
./configure --with-yapp=pgplot --disable-shared
make build check bench5
source nemo_start.sh

Disabling SIP is not recommended, so we’ve been told.

7.2. Rebuilding

If you have an existing installation, but many things have change, this is probably the preferred method:

cd $NEMO
git pull
make rebuild

this will also preserve the possibly peculiar options for configure that you passed the first time it was installed. Or more importantly, if you had edited the $NEMOLIB/makedefs file.

7.3. Advanced Installation

It’s a fact of life that you will not be satisified with the compiler or libraries that your system provides. Add to this that if you don’t have admin privilages, and you might be in for a rude awakening.

No worries, NEMO has you covered (to some degree). We provide an environment (a poor man’s container) where most open source libraries can be installed with a supported $NEMO/opt prefix. This means you can configure packages using

--with-prefix=$NEMO/opt

of for cmake based packages

-DCMAKE_INSTALL_PREFIX=$NEMO/opt

as NEMO generally adds the $NEMO/opt tree search for include and library files, as well as adding its binaries to the search path.

For some packages this has been automated using the mknemo command, described in the next section.

7.4. mknemo

Although the mknemo script was intended to quickly compile a NEMO program (from any directory), and without the need to know where the source code lives. It is now also used to aid the installation of a number of supported libraries that can be used by NEMO. They are compiled within $NEMO/local, and will be installed in $NEMO/opt, as described in the previous section. The supporting scripts are generally located $NEMO/src/scripts/mknemo.d for you to examine.

Examples:

mknemo cfitsio fftw gsl hdf4 hdf5 hypre netcdf4 wcslib

The Programmers Guide (*) will give some advanced examples how to deal with other libraries, and writing your own programs or one of the plugins.

7.5. python

With so many useful python packages around, and so many different methods (anaconda, conda, venv etc.), we will not recommend a method, as this will likely depend on your own situation. The installation examples below should give you enough information how to adapt it for your python installation. It goes without saying (it 2021 here) we only support python3.

However, if you install python from within NEMO, there will be a $NEMO/anaconda3 directory, that gets automatically activated once NEMO is loaded. Here is how you can install that version:

cd $NEMO
make python

This will install a few python modules we often wind up using: amuse-framework, amuse-galactics, amuse-gadget2, amuse-bhtree, astromartini, gala, galpy, pynbody, python-unsio, python-unsiotools, and yt

For a number of these we have small test scripts to see if they are functional:

cd $NEMO/src/scripts/python
make tests

For the cases where you want some control and be in developer mode, we suggest the recommended practice of placing the code in $NEMO/local, as can be seen in the example below

cd $NEMO/local
git clone https://github.com/webbjj/clustertools
pip install -e clustertools

For a few packages, we have a few existing examples in the $NEMO/usr tree (e.g. amuse, martini, unsio and uns_projects)