3.4.2. MacOS X Installation

Warning

This installation method is no longer maintained, and is unlikely to work as written. Please refer to Installation for more up-to-date instructions.

There are a number of dependencies for RMG-Py. This page will guide you through installing them. You will need the Command Line Tools for XCode. If you are not using Anaconda to install RMG-Py, we highly recommend the Homebrew package manager. The following instructions assume that you have installed Homebrew and its requirements. We recommend using a Virtual Environment for your Python packages, but this is optional (without it you may need to add sudo before some commands to solve permission errors).

You will also need gfortran, Python, Numpy and Scipy. We typically install them using homebrew-python (which used to be at https://github.com/Homebrew/homebrew-python) but other methods may work as well.

• For example:

  brew tap homebrew/python
  brew install numpy
  brew install scipy
  brew install matplotlib --with-cairo --with-ghostscript --with-ticl-tk --with-pyqt --with-pygtk --withgtk3
  

• Install git if you don’t already have it (you may also like some graphical interfaces like GitX or GitUp or dozens more):

  brew update
  brew install git
  

• Optional (but recommended for Nitrogen-chemistry nomenclature): install OpenBabel:

  brew install open-babel --with-python --HEAD
  

• Install RDKit:

  brew tap rdkit/rdkit
  brew install rdkit --with-inchi
  brew link --overwrite rdkit
  

  You’ll need to set an environment variable to use it, eg. put this in your ~/.bash_profile file:

  export RDBASE=/usr/local/share/RDKit
  

• The following dependencies are also required for core RMG functions and must be installed from source before building RMG:

  pyrdl: RingDecomposerLib, used for ring perception. Download from https://github.com/rareylab/RingDecomposerLib. Requires CMAKE to compile.

  lpsolve: Mixed integer linear programming solver. Download from https://sourceforge.net/projects/lpsolve/. Python extension also required.

• Make a directory to put everything in:

  mkdir ~/Code
  

• Get the RMG-Py source code and the RMG-database from GitHub:

  cd ~/Code
  git clone https://github.com/ReactionMechanismGenerator/RMG-database.git
  git clone https://github.com/ReactionMechanismGenerator/RMG-Py.git
  

• Install the Python dependencies listed in the RMG-Py/requirements.txt file using pip (do easy_install pip if you don’t already have it):

  pip install -r RMG-Py/requirements.txt
  

• Get and build PyDQED:

  cd ~/Code
  git clone https://github.com/ReactionMechanismGenerator/PyDQED.git
  cd PyDQED
  export LIBRARY_PATH=$(dirname $(gfortran -print-file-name=libgfortran.a))
  make
  make install
  

• Get and build PyDAS:

  cd ~/Code
  git clone https://github.com/ReactionMechanismGenerator/PyDAS.git
  cd PyDAS
  export LIBRARY_PATH=$(dirname $(gfortran -print-file-name=libgfortran.a))
  make
  make install
  

• Build RMG-Py:

  cd ~/Code/RMG-Py
  make -j4
  

• Run an example:

  cd ~/Code/RMG-Py/
  python rmg.py examples/rmg/minimal/input.py
  

  Verify your installation by opening the resulting output.html file under the “examples/rmg/minimal” directory.

  You can also use the Makefile targets to test and run examples:

  cd ~/Code/RMG-Py/
  make test
  make eg1
  make eg2
  

To run with on-the-fly Quantum Mechanics calculations, you will also need to install MOPAC or Gaussian, then run make QM.