PyQwt

Qwt and PyQwt exists for both Linux/Unix and Windows. Qwt is a set of plotting widgets. Installing these is currently not quite as comfortable as installing numpy, but the instructions in the package are excellent:

boud@calcifer:~/src > tar -xzf qwt-0.3.0.tar.gz
boud@calcifer:~/src > cd qwt-0.3.0
boud@calcifer:~/src/qwt-0.3.0 > ls
CHANGES  COPYING  CVS  doc  Doxyfile  Doxyfile.users  examples
      include  INSTALL  makefiles  propagate  qwt.pro  README
      README.QT  src
boud@calcifer:~/src/qwt-0.3.0 > cp makefiles/linux-g++ ./Makefile
boud@calcifer:~/src/qwt-0.3.0 > make
...
ln -s libqwt.so.0.3.0 libqwt.so
ln -s libqwt.so.0.3.0 libqwt.so.0
ln -s libqwt.so.0.3.0 libqwt.so.0.3
boud@calcifer:~/src/qwt-0.3.0 > nc README
boud@calcifer:~/src/qwt-0.3.0 > su
Password:
root@calcifer:/home/boud/src/qwt-0.3.0 > mv libqwt.* /usr/local/lib
root@calcifer:/home/boud/src/qwt-0.3.0 > mkdir /usr/local/include/qwt
    

This did the trick—at least on my system! Now you just have to instal the Python bindings. This is even easier, since PyQwt now uses Distutils to get itself installed. However, note that you need to have a source installation of PyQt if you intend to build PyQwt from source. There are currently binary packages for Windows and some versions of Linux, like Mandrake.

PyQwt has a number of illustrative demo scripts—here, I picked one to demonstrate to you the way it works.

#!/usr/bin/env python

#
# qwtdemo.py
#
# Demonstrates that you can plot NumPy arrays and lists of Python floats.
# NumPy arrays are more elegant and more than 20 times faster than lists.

import sys

from qt import *
from qwt import *
from Numeric import *

def drange(start, stop, step):
    start, stop, step = float(start), float(stop), float(step)
    size = int(round((stop-start)/step))
    result = [start]*size
    for i in xrange(size):
        result[i] += i*step
    return result
        
def lorentzian(x):
    return 1.0/(1.0+(x-5.0)**2)


class ListArrayDemo(QWidget):
    def __init__(self, *args):
        apply(QWidget.__init__, (self,) + args)

        # create a plot widget for NumPy arrays
        self.aplot = QwtPlot('Plot -- NumPy arrays', self)
        # calculate 2 NumPy arrays
        xa = arange(0.0, 10.0, 0.01)
        ya = lorentzian(xa)
        # insert a curve, make it red and copy the arrays
        ca = self.aplot.insertCurve('y = lorentzian(x)')
        self.aplot.setCurvePen(ca, QPen(Qt.red))
        self.aplot.setCurveData(ca, xa, ya)

        # create a plot widget for lists of Python floats
        self.lplot = QwtPlot('Plot -- List of Python floats', self)
        # calculate 2 lists of Python floats
        xl = drange(0.0, 10.0, 0.01)
        yl = map(lorentzian, xl)
        # insert a curve, make it blue and copy the lists
        cl = self.lplot.insertCurve('y = lorentzian(x)')
        self.lplot.setCurvePen(cl, QPen(Qt.blue))
        self.lplot.setCurveData(cl, xl, yl)

    def resizeEvent(self, e):
        x = e.size().width()
        y = e.size().height()/2
    	self.aplot.resize(x, y)
	self.aplot.move(0, 0)
        self.lplot.resize(x, y)
        self.lplot.move(0, y)

# admire	
app = QApplication(sys.argv)
demo = ListArrayDemo()
app.setMainWidget(demo)
demo.resize(400, 600)
demo.show()
app.exec_loop()
    

As you can see, the core of the Qwt library is the QwtPlot widget - an object that knows how to plot, but can be used as any other Qt widget.