python_tutorial

Python tutorial for Space@VT Grad seminar

Authors: AJ Ribeiro
Nathaniel A. Frissell
Sebastien de Larquier

Content

• Repository content
• Python install instructions
• Running this code
• See the wiki for more stuff

Repository content

This repository contains IPython notebooks (*.ipynb) covering:

• Python basics
• Plotting library: matplotlib
• Scientific/Numerical library: SciPy & Numpy
• Code wrapping (C/Fortran)
• SpacePy
• A space science exemple use of Python

These notebooks mostly contain all the code needed, but some examples are directly runfrom stand-alone python scripts (*.py).

The "data" directory contains data files used as examples in the notebooks.

The "examples" directory contains more python examples (from the matplotlib gallery).

Python install instructions

In order to run the examples provided in this repository, you will need:

• Python 2.7
• IPython
• matplotlib
• SciPy
• Numpy
• h5py
• idlsave
• SpacePy

Linux

Ubuntu

apt-get install -y python2.7 python-pip ipython
pip install --upgrade ipython
apt-get install -y ipython-notebook ipython-qtconsole
apt-get install -y python-scipy python-matplotlib python-mpltoolkits.basemap python-h5py

Note: aptitude does not install the latest version of basemap yet. For a more up-to-date version, go on the basemap website and follow install instructions.

SpacePy on Ubuntu

Take care of dependencies:

apt-get install -y easy_install
easy_install networkx
apt-get install -y gfortran
apt-get install -y python-numpy-dbg # (Maybe needed, maybe not)
easy_install ffnet
apt-get install libncurses5-dev # (Needed for NASA CDF)

You also need the NASA CDF library: Download tar.gz file from http://cdf.gsfc.nasa.gov/ Follow instructions in README.install. In summary:

make OS=linux ENV=gnu all
make test # (Optional step... just makes sure everything is working!)
make INSTALLDIR=/usr/local/cdf install # (Feel free to change the install directory if you like.)
# Source /usr/local/cdf/bin/definitions.B in your .bashrc file!!

Now go download SpacePy from http://spacepy.lanl.gov/. They have good instructions, but if you did everything above, you should be able to just run their install script and it should be good to go:

python setup.py install

Other distributions

Most linux distributions come with package managers (like aptitude for Ubuntu, or Yum for Fedora). Use it to install python.

MacOS

We strongly recommend using MacPorts. Using MacPorts you can run the following lines to install python and useful modules.

ver=27
port install python${ver}
port install py${ver}-ipython
port install py${ver}-numpy
port install py${ver}-matplotlib
port install py${ver}-matplotlib-basemap
port install py${ver}-scipy
port install py${ver}-h5py
port install py${ver}-tornado
port install py${ver}-zmq
port install py${ver}-pil

Windows

Unfortunately, although python runs on windows, it is poorly supported and requires a lot of patience to install...

1. Download and install Python 2.7 from http://www.python.org/download/

2. Add c:\Python and c:\Python27\Scripts (or wherever you installed Python) to your environment variables

3. Follow the instructions to install setuptools from http://pypi.python.org/pypi/setuptools

4. the rest of the packages can be installed using easy_install.exe from the windows command prompt, E.G. easy_install.exe ipython

   if at any point you are told that another package is needed, use easy_install.exe to install it

NOTES:

• If your install fails, look into manually installing the binaries (http://www.lfd.uci.edu/~gohlke/pythonlibs/)
• In order to build and install the C extension library, you must install Microsoft Visual C (MSVC)

Running this code

To open IPython notebooks (*.ipynb), simply start the notebook interface in this directory, or start it anywhere then drag and drop the notebooks in your browser.

To start IPython notebook, run

$ ipython notebook --pylab=inline

To run *.py files, execute

$ python your_script.py

or from an IPython console

$ run your_script.py