The installation process in SEAMM involves two main parts:
SEAMM core environment consisting of the GUI for creating flowcharts, submitting jobs, publishing results etc., and
Web-based dashboard responsible for monitoring jobs, managing the job directories and analyzing the results.
Both parts rely on having conda installed on the host system. Installers for Miniconda or Anaconda can be found in their hosting websites. For the minimalists, Miniconda, which can be considered as a reduced-size version of Anaconda, will be sufficient for the sole purpose of working with SEAMM.
SEAMM Core Installation
Although in principle, SEAMM would be natively installable on Windows, its key third-party dependencies could not supported by this operating system. However, the Windows Subsystem for Linux (WSL) supports SEAMM and its dependencies. The corresponding instructions for setting up a WSL environment for SEAMM installation is under preparation. While Windows 11 provides full support for the SEAMM requirements, handling Windows 10 appears to be more complicated and we are currently working on addressing this case.
A video of installing SEAMM is available in the SEAMM YouTube channel. Please note that this video is slightly out of date and there might be small differences between what you read here and watch there.
The SEAMM should be installed in the seamm conda environment. Open a terminal and run the following commands:
conda create -n seamm -c conda-forge seamm seamm-dashboard seamm-installer
Once the environment is in place, you can activate it as prompted:
conda activate seamm
You can choose a different name for your SEAMM environment by replacing
<name> argument in the
-n <name> option with your preferred title.
Now run the seamm_installer to install both the SEAMM environment and the available plug-ins:
Depending on your internet connection and processing power, the execution will probably take a while (~10-20 min) to finish. The reason is that the aforementioned command not only installs SEAMM and its core plug-ins but also additional software packages such as LAMMPS, Psi4, Packmol, and DFTB+ all withing their exclusive conda environments. As the installation of some of the large-size packages such as Psi4 can be time-consuming process, it might be a good time to sit back and take some rest, do other tasks or grab a cup of coffee!
After the seamm_installer finishes its process, you can check on the created environments using the following command:
conda info -e
which generates the proceeding output:
# conda environments: # base /Users/tester/opt/miniconda3 seamm * /Users/tester/opt/miniconda3/envs/seamm seamm-dftbplus /Users/tester/opt/miniconda3/envs/seamm-dftbplus seamm-lammps /Users/tester/opt/miniconda3/envs/seamm-lammps seamm-packmol /Users/tester/opt/miniconda3/envs/seamm-packmol seamm-psi4 /Users/tester/opt/miniconda3/envs/seamm-psi4
The GUI and the main program can be fired up by running:
On Linux and Macintosh operating systems, the seamm_installer creates a user-app called SEAMM (or whatever you named your main conda environment) that can be used for running the SEAMM. You can pin the SEAMM app icon to your desktop Dock or Taskbar (Launcher) for convenient access. On the aforementioned operating systems, the installer also installs the user services for the Dashboard and JobServer. On Linux, they should already be running behind the scene. On the Mac system, however, you will need to either log out and log back in to your machine user account to get them started or follow the instructions that the installer printed (which requires administrator access to the hosting machine you are working on). The running processes for the Jobserver and the Dashboard can be traced by using the following command:
ps -eaf | grep seamm
which yields the following output:
501 16317 1 0 8:58AM ?? 0:01.36 /Users/psaxe/opt/miniconda3/envs/seamm/bin/python3.9 /Users/psaxe/opt/miniconda3/envs/seamm/bin/jobserver 501 16325 1 0 8:58AM ?? 0:14.96 /Users/psaxe/opt/miniconda3/envs/seamm/bin/python3.9 /Users/psaxe/opt/miniconda3/envs/seamm/bin/seamm-dashboard 501 17050 16626 0 9:55AM ttys002 0:00.01 grep seamm
You should see two lines similar to those presented above; however, the paths and the versions of your python might be different.