[Jboost-cvs] jboost/scripts surfing.py,NONE,1.1

[Aaron A. <aa...@us...>]

Update of /cvsroot/jboost/jboost/scripts
In directory sc8-pr-cvs6.sourceforge.net:/tmp/cvs-serv21939/scripts

Added Files:
	surfing.py 
Log Message:
added

--- NEW FILE: surfing.py ---
#! /usr/bin/env python

import string
import getopt
import sys
import pickle
import os
import re
import math

def usage():
	print 'Usage: margin.py    '
	print '\t--surfing=filename  surfing file as output by jboost (-a -2 switch)'
	print '\t--scores=filename   scores file as output by jboost (-a -2 switch)'
	print '\t--data=filename     train/test filename'
	print '\t--spec=filename     spec filename'
	print '\t--booster=type      type={normal, brown}'
	print '\t--framerate=num     an integer specifying framerate'


def erf(z):
	t = 1.0 / (1.0 + 0.5 * abs(z));
	ans = 1 - t * math.exp( -z*z   -   1.26551223 + \
				       t * ( 1.00002368 + \
				       t * ( 0.37409196 + \
				       t * ( 0.09678418 + \
				       t * (-0.18628806 + \
				       t * ( 0.27886807 + \
				       t * (-1.13520398 + \
				       t * ( 1.48851587 + \
				       t * (-0.82215223 + \
				       t * ( 0.17087277))))))))))
	if z < 0:
		ans = - ans
	return ans;


def get_weight_line(params, start, end, max_bin):
	weights = []
	s = c= c1 = c2 = theta = 0
	if(params[0] == 'brown'):
		c = float(params[1])
		s = float(params[2]) #time remaining
	elif(params[0] == 'normal'):
		c = float(params[1])
		s = c-float(params[2]) #time played
		start = -1.0
		end = 1.0
		c1 = float(params[3])
		c2 = float(params[4])
		theta = float(params[5])
		
		
	num_steps = 100
	step = (end - start) / num_steps
	x = start
	for i in range(0,num_steps):
		mu = var = 0
		if params[0] == 'brown': 
			mu = -s
			var = c
		elif params[0] == 'normal':
			mu = theta - c1 * (math.exp(-s) - math.exp(-c))
			var = c2 * (math.exp(-2*s) - math.exp(-2*c)) + 0.05
					 
		norm = 1 / math.sqrt(var*math.pi)
		y = math.exp(-math.pow(x-mu,2)/(var)) * max_bin 
		p = (1 - erf((x-mu)/math.sqrt(var)))/2
		weights.append((x,y,p))
		x += step

	return weights


def get_margin_hist(margins, is_cum):
	num_bins = 0
	if is_cum:
		num_bins = 100
	else:
		num_bins = 30

	marg_max = max(margins)
	marg_min = min(margins)
	bin_size = (marg_max - marg_min) / num_bins
	b = marg_min
	i = 0
	hist = []
	x_axis = []
	j = 0
	EPS = bin_size / 10000
	total_seen = 0
	while b <= marg_max:
		hist.append(0)
		while i < len(margins) and (margins[i] <= b + bin_size + EPS):
			i += 1
			hist[j] += 1
		x_axis.append(b + bin_size)
		if is_cum:
			tmp = hist[j]
			hist[j] += total_seen
			total_seen += tmp 
		b += bin_size
		j += 1

	s = 0
	if is_cum:
		s = max(hist)
	else:
		s = sum(hist)
	return [(x, float(h) / s)  for (x,h) in zip(x_axis,hist)]
		


def paste_gnuplot(num_iterations, framerate): 
	print 'Starting animated gif creation'
	os.system('convert -rotate  90 -delay 5 -loop 0 surfing*.eps surfing.gif')
	print 'Finished gif creation'
	print 'See surfing.gif for animation'

line_compare = lambda x,y : cmp(float(string.split(x,'\t')[0]), float(string.split(y,'\t')[0]))

def write_gnuplot(lines, iter, datafile, params):

	lines.sort(line_compare);


	margins = [ float(string.split(x,'\t')[0])    for x in lines]
	weights = [ float(string.split(x,'\t')[1])    for x in lines]
	potentials = [ float(string.split(x,'\t')[2])    for x in lines]



	marg_max = max(margins);
	marg_min = min(margins);
	lines = get_margin_hist(margins, 1);
	max_bin = max([l[1] for l in lines])
	lines = [str(l[0]) + ' ' + str(l[1]) + '\n' for l in lines] 
	f= open('surfing_hist.dat', 'w')
	f.writelines(lines)
	f.close()

	weight_line = get_weight_line(params, marg_min, marg_max, max_bin)
#	lines = [ str(m) + ' ' + str(p) + ' ' + str(w) + '\n'  for (m,w,p) in weight_line]
	lines = [ str(m) + ' ' + str(p) + '\n'  for (m,w,p) in weight_line]
	f= open('surfing_weight.dat', 'w')
	f.writelines(lines)
	f.close()


	xrange = max([abs(marg_min), abs(marg_max)])
	yrange = max_bin

	epsoutlines = []
	epsoutlines.append('set terminal post\n')
	epsoutlines.append('set output \'surfing%05d.eps\'\n' % (iter))
	epsoutlines.append('set title "'+  params[0] +  'boost ' + datafile + ' Surfing: Iteration ' + str(iter) + '" font "Times,20"\n')
	epsoutlines.append('set key left top\n')
	epsoutlines.append('set yzeroaxis lt -1\n')
	if params[0]=='normal':
		epsoutlines.append('set xrange [-1:1]\n')
		epsoutlines.append('set yrange [0:1]\n')
	else:
		epsoutlines.append('set xrange [%0.2f:%0.2f]\n' % (-xrange, xrange))
		epsoutlines.append('set yrange [0:1]\n')
	epsoutlines.append('set xlabel "Margin" font "Times,20"\n')
	epsoutlines.append('set ylabel "Cumulative Distribution" font "Times,20"\n')
	epsoutlines.append('plot "surfing_hist.dat" using 1:2 title "Margin CDF" with lines, ' + \
				'"surfing_weight.dat" using 1:2 title "Potential" with lines \n')
#	if params[0]=='normal':
#		epsoutlines.append('set parametric\n')
#		epsoutlines.append('set trange [0:1]\n')
#		epsoutlines.append('replot ' + str(params[5]) + ', t \n')

	pngoutlines = epsoutlines[:]
	pngoutlines[0] = 'set terminal png notransparent small\n'
	pngoutlines[1] = ('set output \'surfing%05d.png\'\n' % (iter))

	gifoutlines = epsoutlines[:]
	gifoutlines[0] = 'set terminal gif notransparent\n'
	gifoutlines[1] = ('set output \'surfing%05d.gif\'\n' % (iter))


	f = open('surfing.png.gnuplot', 'w')
	f.writelines(pngoutlines)
	f.close()

	f = open('surfing.eps.gnuplot', 'w')
	f.writelines(epsoutlines)
	f.close()

	f = open('surfing.gif.gnuplot', 'w')
	f.writelines(gifoutlines)
	f.close()

	os.system('gnuplot surfing.eps.gnuplot')
	print 'Finished with iteration ' + str(iter)




def main():
    # Usage: see usage()
    # Looks at all the examples that have negative margins
    # the output can be used to find the examples that are probably mislabeled
    # and also the examples that might need more features

    try:
        opts, args= getopt.getopt(sys.argv[1:], '', ['surfing=', 'scores=','data=','spec=','labels=', 'iteration=', 'framerate=', 'booster=', 'sample'])
    except getopt.GetoptError, inst:
        print 'Received an illegal argument:', inst
        usage()
        sys.exit(2)

    surffile = scoresfile = datafile = framerate = specfile = labelsfile = sample = iteration = booster = None
    for opt,arg in opts:
	    if (opt == '--surfing'):
		    surffile = arg
	    if (opt == '--booster'):
		    booster = arg
	    if (opt == '--framerate'):
		    framerate = int(arg)
	    if (opt == '--scores'):
		    scoresfile = arg
	    if (opt == '--spec'):
		    specfile = arg
	    if (opt == '--data'):
		    datafile = arg


	

    if(surffile == None or scoresfile == None or datafile == None or specfile == None):
	    print 'Need score, data, and spec file.'
	    usage()
	    sys.exit(2)
    if(framerate==None):
	    print 'Need frame rate.'
	    usage()
	    sys.exit(2)
    if(booster==None or (booster!='normal' and booster!='brown')):
	    print 'Only accepts `normal` and `brown` boost.'
	    usage()
	    sys.exit(2)
	    
    print 'Reading surfing file'
    f= open(surffile,'r')
    data= f.readlines()
    f.close()
    num_elements= int((string.split(data[0],'='))[2])
    num_iterations = int(string.split(string.split(data[-num_elements-2], ',')[0], '=')[1])

    os.system('rm surfing*.eps surfing*.png')

    for iter in range(0, num_iterations+1, framerate):
	    if (booster=='brown'):
		    total_time = string.split(data[iter*(num_elements+2)+1], ' ')[2]
		    time_left  = string.split(data[iter*(num_elements+2)+1], ' ')[3]
		    params = ('brown', total_time, time_left)
	    if (booster=='normal'):
		    total_time = string.split(data[iter*(num_elements+2)+1], ' ')[2]
		    time_left  = string.split(data[iter*(num_elements+2)+1], ' ')[3]
		    c1 = 1.818
		    c2 = 1
		    theta = 0.1
		    #c1  = string.split(data[iter*(num_elements+2)+1], ' ')[4]
		    #c2  = string.split(data[iter*(num_elements+2)+1], ' ')[5]
		    #theta  = string.split(data[iter*(num_elements+2)+1], ' ')[6]
		    params = ('normal', total_time, time_left, c1, c2, theta)
	    lines = [x   for x in data[iter*(num_elements+2)+2:(iter+1)*(num_elements+2)]]
	    write_gnuplot(lines, iter, datafile, params)

    paste_gnuplot(num_iterations, framerate)
	    
	    

if __name__ == "__main__":
    main()