NuGridPy

NuGridPy is a Python package containing tools to access and analyse (e.g. plot) various output formats (including the hdf5-based se format) from NuGrid codes (mppnp and ppn) and from the MESA stellar evolution code. In principle the NuGridPy package can be used on any stellar evolution code output if the tools to write se hdf5 files available on the NuGrid web page are used. The mesa.py module will work with MESA ASCII output in the 'LOGS' directory.

Previous home of NuGridPy

MOST RECENT NuGrid SVN REVISION STILL CONTAINING NUGRIDPY: r6619

NuGridPy has been migrated to this github repository from the svn repository. All components present here have now been removed from the svn.

Introduction

The NuGridPy modules were written with an interactive work mode in mind, in particular taking advantage of the interactive ipython session that we usually start with 'ipython --pylab' or inside a jupyter ipython notebook. Once your session starts import the mesa or nugridse or ppn module (depending on which type of data you are working with), for example:

	from NuGridPy import mesa as ms

and read the docstring: help(ms)

There are reasonable doc strings in the modules. If you have made tested and debugged improvements we are happy to know about them and we may add them to the release available on the web page. The tools provided here are useful to us, but of course there are still many things that need attention and improvement. We have a good list of scheduled improvements, let us know if you want to help with these. Pull requests and new issues are most welcome!

Installation

pip install git+https://github.com/NuGrid/NuGridPy.git

Or clone this repo and point the PYTHONPATH variable to the NuGridPy repo directoy.

Required packages

All modules should work with the recommended NuGridDoc python distribution, with one additional package, the future package that needs to be installed additionally.

If you prefer installing packages individually, here are the dependencies explicitely:

• standard:
  • xlrd
  • numpy
  • matplotlib
  • h5py
  • unittes
• python2/3 compatibility
  • future (provides 'past' and 'builtin' modules)

Some of the packages are standard. But you may need:

	sudo pip(3) install -U xlrd future h5py

or, if you want to install in the user home directory,

	pip(3) install --user -U xlrd future h5py

• on Fedora python2:

	dnf install -y h5py python2-future

If h5py complains about missing hdf5.h or hdf5_hl.h:

	sudo apt-get install libhdf5-dev

Mac installation with macports

1. Install ipython and pip:

	port install py34-ipython py34-pip
	port select --set ipython py34-ipython
	port select --set pip py34-pip

2. Install standard prerequisites with pip:

	pip install -U --user numpy scipy xlrd matplotlib future unittest

3. Install h5py:

	port install openmpi-gcc5	
	port install hdf5 +hl +openmpi +cxx +fortran +gcc5
	HDF5_DIR=~/macports CFLAGS="-I$HOME/macports/include/openmpi-gcc5" pip install -U --user h5py

Errors about failure to rebuild netcdf package can be ignored.

4. Set up PYTHONPATH environment variable to point to the

parent directory of NuGridPy, e.g.:

	export PYTHONPATH="$HOME/NuGrid"

TESTING

The following command runs a suite of test cases for the NuGridPy package:

	python -m NuGridPy.selftest