# This example shows a bar graph, which starts with
# a stacked bar skipping to plot the original bar.
# This is done by using stackedbarpos before plotting
# a bar with the bar style. However you can still add
# errorbars, even to the lower bound of the bar.
# This example also shows how to handle several errorbars
# in one plot command by the usenames argument of the
# range style.

from random import random
from pyx import *

g = graph.graphxy(width=8, x=graph.axis.bar())
g.plot(graph.data.file("bar.dat", xname=0, y=2, stack=3,
                       dy="1+random()", dstack="1+random()",
                       context={"random": random}),
       [graph.style.errorbar(),
        graph.style.stackedbarpos("stack"),
        graph.style.bar([color.rgb.green]),
        graph.style.range({"y": "stack"}),
        graph.style.errorbar()])
g.writeEPSfile("errors")