Contents
1. Introduction
2. Dependencies
3. Installation
4. Configuration file
5. License
6. Credits
7. ContactSciTools is a Python package containing many useful tools for scientific computing in Python. The package is built on top of other widely used packages such as NumPy, SciPy, ScientificPython, Gnuplot, etc.
SciTools only requirements are Python (see http://python.org) and Numerical Python (see http://numpy.org). In the past, SciTools worked transparently with NumPy, Numeric, and Numarray, through its numpytools module. Although this module still works, we recommend users to use NumPy only.
There are several different ways of installing SciTools:
Normal install ([, ... ] means optional)
python setup.py install [,--prefix=$PREFIX]
This install makes SciTools use Matplotlib as plotting engine, if a a working Matplotlib is found. Otherwise Gnuplot is used (which requires both the Gnuplot program in C and the Python interface Gnuplot.py to be installed). If neither Matplotlib nor Gnuplot is found, Matplotlib is still used as default plotting engine. (The rest of SciTools will work well, of course.)
Installing Matplotlib is easy on most platforms: just download the tarball, pack it out, and run python setup.py install.
You can explicitly specify the desired plotting engine on the command line, using the --easyviz_backend option:
python setup.py install --easyviz_backend gnuplot [,--prefix=$PREFIX]
The default plotting engine is specified in the file scitools.cfg in the scitools package directory, and the --easyviz_backend option leads to an automatic edit of the "backend" line in the [easyviz] section of that configuration file.
Don't edit the scitools.cfg file manually before installation (use the --easyviz_backend option) because setup.py will override your edits of the easyviz backend. However, all other edits of the configuration file can be edited. For example, the Matplotlib GUI is set to TkAgg (i.e., using Tkinter, which is standard in most Python installations). The GUI can be set to other values, such as Qt4Agg (provided you have Qt4 and a Python interface to it).
After SciTools is installed, you may edit the installed version of scitools.cfg, or (better) have your own .scitools.cfg file in your home directory.
Setuptools using eggs
First build the egg with the following command
python setupegg.py [, egg_info --tag-svn-revision ] bdist_egg
The --easyviz_backend option can be used, as described above, to change the default plotting engine.
Then install the created egg using easy_install
easy_install [, --prefix=$PREFIX] dist/Scitools-0.7-py2.6.egg
(Version numbers of SciTools and Python in this filename may vary.)
Manual installation
Put the lib subdirectory in a directory contained in PYTHONPATH, and the bin subdirectory in a directory contained in PATH. (This gives access to the directories for scitools, easyviz, etc.) Edit the scitools.cfg file to set the desired default plotting engine (look for the "backend =" line in the [easyviz] section).
Debian/Ubuntu packages
If you have Debian, or a Debian based platform like Ubuntu, you can do
sudo apt-get install python-scitools
as SciTools is now a part of Debian. You can also find newer Debian packages for SciTools at googlecode.com.
Extended (not supported anymore)
Run build_scitools++.py. This makes a new directory build/scitools++ that you can move to any convenient place. scitools++ contains scitools plus some standard Python modules like Gnuplot, IPython, epydoc, etc. Put scitools++/lib in PYTHONPATH and scitools++/bin in PATH. Or you may run scitools++/setup.py to install scitools++ in the official installation directories on your system.
The behavior of many parts of SciTools and in particular the subpackage Easyviz (for plotting) can be controlled in a configuration file. Please read the subsection "Setting Parameters in the Configuration File" under the section "Advanced Easyviz Topics" in the Easyviz tutorial (pydoc scitools.easyviz will show the tutorial). In particular, if you use Matplotlib as the default plotting engine, you may want to turn on the use of LaTeX for rendering legends, titles, and numbers. By default, LaTeX is turned off when SciTools is installed. It is easy to change this by locating the scitools.cfg file in the folder where the SciTools package is installed and edit the line with text.usetex in the [matplotlib] section of this file. A better solution is to copy the system scitools.cfg file to .scitools.cfg in your home folder and edit that file. You can also change the GUI used by Matplotlib. Users who do not apply Matplotlib for plotting will seldom need to edit the configuration file.
SciTools is licensed under the new BSD license, see the LICENSE file.
Lumpy.py and Gui.py are licensed under GPL, however, permission is granted by Allen Downey to include these under a BSD license.
SciTools was initially mainly developed by Hans Petter Langtangen hpl@simula.no for his book "Python Scripting for Computational Science" (Springer, 1st edition 2003, 3rd edition 2009). The Easyviz package was mainly developed by Johannes H. Ring johannr@simula.no. Johannes H. Ring has been the principal maintainer of SciTools.
Some modules included in SciTools are written by others:
Allen B. Downey downey@allendowney.com wrote Lumpy.py and Gui.py Imri Goldberg lorgandon@gmail.com wrote aplotter.py Fred L. Drake, Jr. fdrake@acm.org wrote pprint2.py Gael Varoquaux gael.varoquaux@normalesup.org wrote pyreport
Code contributors include:
Rolv E. Bredesen rolveb@simula.no Joachim B. Haga jobh@simula.no Mario Pernici Mario.Pernici@mi.infn.it Ilmar Wilbers ilmarw@simula.no Arve Knudsen arvenk@simula.no
Comments, suggestions, bug fixes should be sent to
scitools@googlegroups.comThe latest release of SciTools can be obtained from
http://scitools.googlecode.comwhile the latest Subversion revision can be obtained either anonymously (for non-members) by
svn co http://scitools.googlecode.com/svn/trunk/ scitoolswhile SciTools members should use
svn co https://scitools.googlecode.com/svn/trunk/ scitools --username USERNAMESee http://code.google.com/p/scitools/source/checkout for details.