pyLBPM

A python-based management interface for the LBPM software

Installation

To install pyLBPM with pip (note Python installs scripts to $HOME/.local/.bin)

git clone git@github.com:digital-porous-media/pyLBPM.git
cd pyLBPM
pip install ./
export PATH=$HOME/.local/bin:$PATH

⚠️ Note It is best to run the following lines from nodes with access to a GPU. On Texas Advanced Computing Center (TACC) systems, this can be done using an interactive development session. The total installation time takes 1-2 hours.

Accessing idev on TACC Vista

idev -p gh-dev -t 2:00:00
module load gcc cuda python3

Accessing idev on TACC Lonestar6

idev -p gpu-a100-small -t 2:00:00
module load gcc/9.4.0 cuda

To download LBPM dependencies, launch a Python command prompt and run the following from the Python command line.

from pyLBPM import lbpm_setup
lbpm_install_path="~/local"
lbpm_config_setup=lbpm_setup.download_dependencies(lbpm_install_path)
lbpm_config_setup=lbpm_setup.install_dependencies(lbpm_install_path)

To install LBPM from the python command line

from pyLBPM import lbpm_setup
lbpm_install_path="~/local"
lbpm_config_setup=lbpm_setup.install_lbpm(lbpm_install_path)

Launching simulations

Step 1. Set up a working directory for your simulation

SIMDIR=/path/to/my/simulation
mkdir -p $SIMDIR

Step 2. Run example scripts to setup a simple simulation (modify as needed to suit your case)

 python example/setup_color_simulation.py $SIMDIR

Step 3. Launch the simulation (should capture local build environment from config files)

python example/launch_color.py $SIMDIR

Managing simulation data

From the python command line

sdir="/path/to/my/simulation"
from pyLBPM import lbpm_color_model
timelog=lbpm_color_model.read_timelog(simulation_directory=sdir, plot_data=False)
print(timelog)

Visualizing simulation data

To do: write a python interface to read LBPM HDF5 files