Plot csv files for ps

ps_plot.py

import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
import numpy as np
import csv

files=[ "l1dsize.csv","l1dassoc.csv", "l2uassoc.csv", "l2usize.csv", "l3uassoc.csv", "l3usize.csv", "lsize.csv"]

for filename in files:
	with open(filename, 'r') as csvfile:
		csvreader = csv.reader(csvfile, delimiter=';')
		values=list(csvreader)	

	# assocs or sizes
	datapoints=[]
	#for example 500, 1024, 3000
	problemsizes=[]

	#offset of 2 to skip headlines
	for i in range(2, len(values)):
		if not values[i][0] in problemsizes:
			problemsizes.append(values[i][0])
		if not values[i][1] in datapoints:
			datapoints.append(values[i][1])

	relcycles = []
	tmp=[]
	foo=1
	for i in range(2, len(problemsizes)*len(datapoints)+2):
		tmp.append(values[i][10])
		if (foo % len(datapoints))==0:
			relcycles.append(tmp)
			tmp=[]
		foo+=1	
		


	#replace ',' with '.' in relcycles for matplotlib to interpret string as float
	for i, cyclevalues in enumerate(relcycles):
		for j, value in enumerate(cyclevalues):
			relcycles[i][j]=value.replace(',','.')


	
	#let the plotting begin
	canvas = plt.subplot(1,1,1)

	xrange = np.arange(0,len(datapoints),1)
	for  i,dataset in enumerate(relcycles):
		canvas.plot(xrange, dataset, label='Psize='+problemsizes[i])
		plt.xticks(np.arange(len(datapoints)), datapoints)

	#shrink canvas box
	box = canvas.get_position()
	canvas.set_position([box.x0, box.y0, box.width * 0.8, box.height])
	#put legend box right
	canvas.legend(loc='upper left', bbox_to_anchor=(1, 0.5))

	plt.title(values[0][0])
	plt.ylabel(values[1][10])
	plt.xlabel(values[1][1])

	plt.savefig(filename.replace('.csv',''))
	plt.close()