Using Python with the Pencil Code

The postprocessing in the Pencil Code (hereafter PC) is historically done under IDL (Interactive Data Language; see e.g. its wikipedia webpage). However, some PC developers recently adopted the public-domain software Python for the processing and we will use it during this school for the plotting.

What do you need for running Python? Python is now installed by default in any major Linux distribution (Fedora, Ubuntu, Mandriva, etc...). But some additional scientific libraries/tools are usally required:

numpy	the scientific library for manipulating arrays [strongly needed]
matplotlib	the plotting library, mainly for 2-D plots [strongly needed]
ipython	the interactive Python shell [strongly needed]
python-tk	[needed by default on Ubuntu]
basemap	a plotting library that allows to plot on a sphere, to do maps, etc... [optional]

All these things are already installed on each ICTP PC in the computing room (whom the hostname is of the form infolab-*.ictp.it).

the PYTHONPATH variable is initialized when sourcing the PC (if it already exists, the PC Python directory $PENCIL_HOME/python is simply added).
a good idea is to directly load all the PC Python subroutines when launching ipython; to do that, open your $HOME/.ipython/ipythonrc file and add the pencil directory at the line 537 (in the 'import_all' section):
```
import_all pencil
```
It means that all the Python subroutines located in the $PENCIL_HOME/python directory are loaded each time you start a ipython shell. They will be accessible under the name 'pc.*'.
Note: before editing you should have started ipython at least once in order for the file to be created.
Example 1: how to plot the time_series.dat file under Python.
Example 2: how to plot a snapshot of entropy under Python.

Numerical Experiments homepage

$Date: 2009/09/09 12:17:32 $, $Author: dintrans $, $Revision: 1.11 $