The offical documentation for LWSSPY can be found here: Documentation

Hopefully, it works using conda install lwsspy will work in the future.

However,

git clone [email protected]:lsawade/lwsspy.git
cd lwsspy
pip install .

Should work. Use pip install -e . for development mode.

It's a bit more convoluted since we will have to install some things using, conda, then install parallel HDF5, mpi4py, and h5py. And then, we can install the rest of the functions via pip.

1. conda env create -f summitenv.yml
2. Check the documentation for the installation of Parallel HDF5 on the cluster, and subsequent installations of mpi4py, and h5py.
3. pip install summitreq.yml

This repo contains a startup.py file that can be called when loading the python shell. If following line

export PYTHONSTARTUP=path/to/repo/startupfiles/python.py

is added to the ~/.bashrc file, Python will use the environment variable to load up the script. The script right now is set to load all of pyplot's and numpy's functions without prefix as well as all of lwsspy's functions.

This makes it possible to simply do small commands in Matlab style such as help(fakerelation) or plot(x,y,'o'), etc.

In addition to the Python startup file. Ipython has the ability to reload modified modules on the fly. This is extremely convenient:

Simply run the line:

cp path/to/repo/startupfiles/ipython.ipy ~/.ipython/profile_default/startup/

To run the lines in ipython.ipy. The lines are the following:

# Activate autoreload
%load_ext autoreload
%autoreload 2