I'm trying to create a dynamically generated png of an arbitrary number of barplots. The height of the png should correspond with the number of plots that are made. In jpGraph this is very easy. All you have to do is set the height of the image to top-margin + (number_of_bars * width_of_bars) + bottom_margin. But matplotlib doesn't use pixels, so it's not as easy. Here is my code:
def output(self):
# titles = a list of the titles that get displayed down the left side
# vals = a list of values that will determine how big to make each bar
titles, vals = self.get_data()
# set the height of each bar based on how many items there are
from pylab import arange
pos = arange(len(vals))
rects = self.ax.barh(pos, vals,
height=0.8, # this doesn't seem to be in any sane unit
align='center',
color=self.bar_color)
from pylab import yticks, xticks
yticks(pos, map(self.make_ytick, titles))
xticks([], []) #remove everything from the x axis
from pylab import title
title(self.get_column_title() + " " + self.title())
self.annotate_bars(rects) # add the text to the end of each bar
return self.fig
then that fig is outputted like this:
response=HttpResponse(content_type='image/png')
fig = self.output()
# this is the magic formula that needs to be changed
# self.len is the number of bars there are
# 0.15 inches per bar + 1 inch for margins
fig.set_figheight((.15 * self.len) +1)
fig.savefig(response,
bbox_inches="tight",
pad_inches=0.5,
format='png')
return response
Here are some output examples of my code:
http://sites.google.com/site/nbvfour/homeAs you can see, there are random swaths of whitespace at the top and bottom of each, and the bard are not consistent. Hoe can I fix this? Also, how do I get those ugly tick marks to go away wile still keeping the titles of each bar?