3.2.1. Binary Installation Using Anaconda for Unix-Based Systems: Linux and Mac OSX

1. Install the conda package manager, if you do not already have it (or Anaconda). Select one of the following options:

   1. Users of Fedora Linux and Red Hat derivatives (RHEL, CentOS Stream) may install from the official repositories and EPEL, respectively, with the command

      sudo dnf install conda
      

   2. All other users, download and install Miniconda.

      The download will be a .sh file with a name like Miniconda3-latest-Linux-x86_64.sh. Open a terminal in the same directory as this file, and type the following to install Conda (replace the name of your .sh file below).

      bash Miniconda3-latest-Linux-x86_64.sh
      

      When prompted to append Anaconda to your PATH, select or type Yes. Install the Conda folder inside your home directory (typically /home/YourUsername/ in Linux and /Users/YourUsername in Mac).

      Note that you should reinitialize or restart your terminal in order for the changes to take effect, as the installer will tell you.

2. Install both RMG and the RMG-database binaries through the terminal. Dependencies will be installed automatically. It is safest to make a new conda environment for RMG and its dependencies. Type the following command into the terminal to create the new environment named ‘rmg_env’ containing the latest stable version of the RMG program and its database.

   conda create -c defaults -c rmg -c rdkit -c cantera -c pytorch -c conda-forge --name rmg_env rmg rmgdatabase
   

   Whenever you wish to use it you must first activate the environment:

   conda activate rmg_env
   

3. You may now run an RMG test job. Save the Minimal Example Input File to a local directory. Use the terminal to run your RMG job inside that folder using the following command

   rmg.py input.py
   

   If you encounter the ImportError related to libmkl_rt.so.2, refer to the Fixing the ImportError related to libmkl_rt.so.2 section below to fix the error and re-run the RMG test job.

You may now use RMG-Py, Arkane, as well as any of the Standalone Modules included in the RMG-Py package.

3.2.1.1. Fixing the ImportError related to libmkl_rt.so.2

You may encounter the following ImportError when you try to run a RMG test job

Traceback (most recent call last):
  File "/PATH-TO-YOUR-ANACONDA/envs/rmg_env/bin/rmg.py", line 48, in <module>
    from rmgpy.rmg.main import RMG, initialize_log, process_profile_stats, make_profile_graph
  File "/PATH-TO-YOUR-ANACONDA/envs/rmg_env/lib/python3.7/site-packages/rmgpy/rmg/main.py", line 51, in <module>
    from cantera import ck2cti
  File "/PATH-TO-YOUR-ANACONDA/envs/rmg_env/lib/python3.7/site-packages/cantera/__init__.py", line 4, in <module>
    from ._cantera import *
ImportError: libmkl_rt.so.2: cannot open shared object file: No such file or directory

where PATH-TO-YOUR-ANACONDA is the path to the Anaconda3 directory installed on your machine. The default install location of Anaconda on Linux is /home/<your_username>/Anaconda3.

To fix this issue, simply copy libmkl_rt.so.1 as libmkl_rt.so.2 under the rmg environment library by typing the following line on your terminal:

cp /PATH-TO-YOUR-ANACONDA/envs/rmg_env/lib/libmkl_rt.so.1 /PATH-TO-YOUR-ANACONDA/envs/rmg_env/lib/libmkl_rt.so.2

Note that PATH-TO-YOUR-ANACONDA needs to be the path to your Anaconda3 directory, which mostly likely looks like /home/<your_username>/Anaconda3. If you are unable to locate the libmkl_rt.so.1 file on your computer, you can find its location with the following command:

locate libmkl_rt.so.1

After copying the file as libmkl_rt.so.2, try running the RMG test job again:

rmg.py input.py

You will see a line saying MODEL GENERATION COMPLETED on your terminal if your RMG test job ran successfully.

3.2.1.2. Updating your binary installation of RMG in Linux or Mac OSX

If you had previously installed a binary version of the RMG package, you may check and update your installation to the latest stable version available on Anaconda by typing the following command on the terminal

source activate rmg_env
conda update rmg rmgdatabase -c rmg