VibIR-Parallel-Compute

This packages is primarily about doing Vibration and Infrared computations and analysis. It is designed to be conducted in parallel and take advantage of multiple nodes available in High Performance Computing (HPC) environments. Aside from the primary computations, additional tools are provided to help with the analysis and post-processing.

Getting started

Installation

via direct Pip

Activate your python environment (conda or virtualenv) and install using pip:

pip install git+https://gitlab.com/morikawa-lab-osakau/vibir-parallel-compute.git

This will install the latest version

via Git

Download the git repository and install:

git clone https://gitlab.com/morikawa-lab-osakau/vibir-parallel-compute.git
pip install vibir-parallel-compute/

This method allows you to retain a copy of the code and also to select a specific version you need.

Developer mode

If you want to add new features or need to edit some portion of the code. You can install it using developer mode to save time on frequent reinstalls.

git clone https://gitlab.com/morikawa-lab-osakau/vibir-parallel-compute.git
pip install -e vibir-parallel-compute/

Usage

A web manual is made available in https://morikawa-lab-osakau.gitlab.io/vibir-parallel-compute. Tutorials and examples are provided to help you get started.

Developer notes

To succesfully run the pytest tests, there are a few prerequisites:

1. Python environment
2. Quantum Espresso executables (pw.x) and the pseudopotential file

Common environment via Docker

Note: The docker container creation can be skipped if you already have an existing QE installation and virtual environment. Just change the necessary variables to the correct paths in your system. Also, we're using the GBRV pseudopotentials in the input files so that need to be specifically available. See the necessary environmental variables in the ("Environment preparation")[#environment-preparation] section.

For the same environment, you can use the docker image provided in the vibir-parallel-compute/docker/Dockerfile:

# Create the docker image (imagename = vibirimage, about 1.6GB) 
docker build -t vibirimage .

# Create and run container (containername=vibircontainer)
docker run -it --name vibircontainer vibirimage

# (OPTIONAL) if exited and want to return to the session/container
# docker start vibircontainer
# docker exec -it vibircontainer /bin/bash

with ls you can see two directories, the pyenv containing the python virtual environment, and the qe containing the QE and pseudopotentials.

Environment preparation

Here, we prepare the environment variables and the python environment for the tests.

# Download repository
git clone -b JOSS-branch https://gitlab.com/morikawa-lab-osakau/vibir-parallel-compute.git

# Package installation (change "pyenv" if you'd like to use your own venv)
source ~/pyenv/bin/activate
cd vibir-parallel-compute/
pip install .

# Export the necessary QE paths (adjust the paths accordingly, if you'll use your own QE installation)
export ESPRESSO_PSEUDO="/home/testuser/qe/pseudo"
export ESPRESSO_COMMAND="pw.x"
export ESPRESSO_TMPDIR="TMPDIR"
export ESPRESSO_BIN="/home/testuser/qe/bin"
export PATH=/home/testuser/qe/bin:$PATH

Running pytest

# Run pytest
pytest -rP -vvv --durations=0  ./vibir_parallel_compute  

Contributing

Please read CONTRIBUTING.md for details on our code of conduct, and the process for submitting pull requests to us.