[Jboost-cvs] jboost/scripts error.py, NONE, 1.1 margin.py, 1.1.1.1, 1.2

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

Update of /cvsroot/jboost/jboost/scripts
In directory sc8-pr-cvs6.sourceforge.net:/tmp/cvs-serv29800/scripts

Modified Files:
	margin.py 
Added Files:
	error.py 
Log Message:
Putting in error.py.  This file is 

Updated build to be more platform independent

Controller now does numRounds+1 iterations

Monitor now conforms to error.py output format

jboost.config has been moved to demo.

Aaron



Index: margin.py
===================================================================
RCS file: /cvsroot/jboost/jboost/scripts/margin.py,v
retrieving revision 1.1.1.1
retrieving revision 1.2
diff -C2 -d -r1.1.1.1 -r1.2
*** margin.py	16 May 2007 04:06:02 -0000	1.1.1.1
--- margin.py	16 May 2007 21:16:29 -0000	1.2
***************
*** 31,34 ****
--- 31,76 ----
  
  
+ def write_files(examples, datafile, sample, margins, falsepos, truepos, falseneg, trueneg):
+ 
+     f= open(datafile + '.margin.report', 'w')
+     f.write('total\t tp\t tn\t fp\t fn\n')
+     f.write(str(len(margins)) + \
+ 	    '\t ' + str(len(truepos)) + \
+ 	    '\t ' + str(len(trueneg)) + \
+ 	    '\t ' + str(len(falsepos)) + \
+ 	    '\t ' + str(len(falseneg)) + \
+ 	    '\n'
+ 	    )
+     f.write('Sensitivity: ' + str(float(len(truepos))/(len(truepos)+len(falseneg))) + '\n');
+     f.write('False Pos Rate: ' + str(float(len(falsepos))/(len(trueneg)+len(falsepos))) + '\n');
+ 
+ 
+     list= [[margins[w],examples[w]] for w in falseneg.keys()]
+     sortByColumns(list, [0], True)
+     writeList(list, datafile + '.false.neg')
+ 
+     if (sample!=None):
+ 	    sample= [list[x] for x in range(0, len(list), 50)]
+ 	    f.write('Sampling '+ str(len(sample)) + ' false negative examples.\n')
+ 	    writeList(sample, datafile + '.false.neg.samples')
+ 
+ 
+ 
+     list= [[margins[w],examples[w]] for w in falsepos.keys()]
+     sortByColumns(list, [0], True)
+     writeList(list, datafile+'.false.pos')
+ 
+     if (sample!=None):
+ 	    sample= [list[x] for x in range(0, len(list), 5)]
+ 	    f.write('Sampling '+ str(len(sample)) + ' false positive examples.\n')
+ 	    writeList(sample, datafile + '.false.pos.samples')
+ 
+ 
+     f.close()
+ 
+     os.system('cat ' + datafile +'.margin.report')
+ 
+ 
+ 
  
      
***************
*** 112,118 ****
--- 154,167 ----
      
      margins= map(computeMargin, scores, labels)
+ 
+     if len(examples) != len(margins):
+ 	print "Don't have margin for each example"
+ 	sys.exit(2)
      
      print 'Matching margins with data'
  
+ 
+     pos_label, neg_label = getLabels()
+ 
      falsepos = dict([(w,[margins[w],examples[w],labels[w]])
  		for w in range(len(margins))
***************
*** 131,170 ****
  
  
! 
!     f= open(datafile + '.margin.report', 'w')
!     f.write('total\t tp\t tn\t fp\t fn\n')
!     f.write(str(len(examples)) + \
! 	    '\t ' + str(len(truepos)) + \
! 	    '\t ' + str(len(trueneg)) + \
! 	    '\t ' + str(len(falsepos)) + \
! 	    '\t ' + str(len(falseneg)) + \
! 	    '\n'
! 	    )
!     f.write('Sensitivity: ' + str(float(len(truepos))/(len(truepos)+len(falseneg))) + '\n');
!     f.write('False Pos Rate: ' + str(float(len(falsepos))/(len(trueneg)+len(falsepos))) + '\n');
! 
! 
!     list= [[margins[w],examples[w]] for w in falseneg.keys()]
!     sortByColumns(list, [0], True)
!     writeList(list, datafile + '.false.neg')
! 
!     if (sample!=None):
! 	    sample= [list[x] for x in range(0, len(list), 50)]
! 	    f.write('Sampling '+ str(len(sample)) + ' false negative examples.\n')
! 	    writeList(sample, datafile + '.false.neg.samples')
! 
!     list= [[margins[w],examples[w]] for w in falsepos.keys()]
!     sortByColumns(list, [0], True)
!     writeList(list, datafile+'.false.pos')
! 
!     if (sample!=None):
! 	    sample= [list[x] for x in range(0, len(list), 5)]
! 	    f.write('Sampling '+ str(len(sample)) + ' false positive examples.\n')
! 	    writeList(sample, datafile + '.false.pos.samples')
  
  
-     f.close()
- 
-     os.system('cat ' + datafile +'.margin.report')
      
      
--- 180,186 ----
  
  
!     write_files(examples, datafile, sample, margins, falsepos, truepos, falseneg, trueneg)
  
  
      
      

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

import string
import getopt
import sys
import os

def usage():
	print 'Usage: margin.py    '
	print '--info=info_file    scores file as output by jboost'
	print '--logaxis           should the axis be log-scaled (default: false)'
	print '--bound             show the bound on training error'

def main():
    # Usage: see usage()

    try:
	    opts, args= getopt.getopt(sys.argv[1:], '', ['info=','logaxis','bound'])
    except getopt.GetoptError:
	    usage()
	    sys.exit(2)
	    
    infofilename = logaxis = showbound = None
    for opt,arg in opts:
	    if (opt == '--info'):
		    infofilename = arg
	    elif (opt == '--logaxis'):
		    logaxis = True
	    elif (opt == '--bound'):
		    showbound = True
	

    if(infofilename == None):
	usage()
	sys.exit(2)

    print 'Reading info file'
    f = open(infofilename,'r')
    lines = f.readlines()
    f.close()

    def get_data(line):
	    list = line.split('\t')
	    ret = {}
	    ret["iter"] = list[0]
	    ret["bound"] = list[1]
	    ret["train"] = list[2]
	    ret["test"] = list[3]
	    return ret

    read_iter = False
    data = []
    for line in lines:
	    line = line.strip()
	    #print line
	    if (line[:8] =='End time'):
		    break
	    if (read_iter):
		    data.append(get_data(line))
	    if (line=='iter \tbound \ttrain \ttest'):
		    read_iter = True

    outlines = []
    for line in data:
	    out = ''
	    out += line['iter'] + ' '
	    out += line['train'] + ' '
	    out += line['test'] + ' '
	    if (showbound):
		    out += line['bound'] + ' '
	    outlines.append(out+'\n')
	    
    f = open('error.dat', 'w')
    f.writelines(outlines)
    f.close()

    outlines = []

    outlines.append('set terminal png notransparent medium\n')
    outlines.append('set output \'error.png\'\n')
    outlines.append('set xlabel "Iteration"\n')
    outlines.append('set ylabel "Error"\n')
    outlines.append('set title ""\n')
    if (logaxis):
	    outlines.append('set logscale x 10\n')
    out = ''
    out += 'plot "error.dat" using 1:2 title \'train\' with lines'
    out += ', "error.dat" using 1:3 title \'test\' with lines'
    if showbound:
	    out += ', "error.dat" using 1:4 title \'bound\' with lines'
    out += '\n'
    outlines.append(out)


    f = open('error.gnuplot', 'w')
    f.writelines(outlines)
    f.close()

if __name__ == "__main__":
    main()