[Jboost-cvs] jboost/scripts VisualizeScores.py,1.2,1.3

[Yoav F. <yf...@us...>]

Update of /cvsroot/jboost/jboost/scripts
In directory sc8-pr-cvs17.sourceforge.net:/tmp/cvs-serv21955/scripts

Modified Files:
	VisualizeScores.py 
Log Message:
VisualizeScores.py is working.

Index: VisualizeScores.py
===================================================================
RCS file: /cvsroot/jboost/jboost/scripts/VisualizeScores.py,v
retrieving revision 1.2
retrieving revision 1.3
diff -C2 -d -r1.2 -r1.3
*** VisualizeScores.py	3 Sep 2008 18:37:32 -0000	1.2
--- VisualizeScores.py	9 Sep 2008 20:43:25 -0000	1.3
***************
*** 1,55 ****
! import sys,os,os.path,re,math
  import jarray
  
! filename = '/Users/yoavfreund/Desktop/Galit_Lahav/Raw_tifs/28/jboost/t28.0/ADD_ALL/trial0.test.boosting.info'
  
  #args = sys.argv[1:]
  #if len(args) != 1:
! #    sys.exit('Usage: ParseTree <name>.output.tree')
  #filename = args[0]
  
- infile = open(filename,'r')
- 
  header_p = re.compile('iteration=(\d+): elements=(\d+):')
  scoreline_p = re.compile('([-+\d.]+):\s+') #([+-1]+):')
  
  for line in infile:
-     #print line
      [(iter,elements)] = header_p.findall(line)
-     iter=int(iter)
      elements = int(elements)
!     # print "iter=%d elements=%d\n" % (iter,elements)
!     indexes=[]
!     scores=[]
!     labels=[]
      for count in range(elements):
          line=infile.readline()
-         #print line
-         a = scoreline_p.findall(line)
-         [index,margin,score,label]=a        
-         #[index,margin,score,x1,x2,label]=a        
-         indexes.append(int(index))
-         scores.append(float(score))
-         labels.append(int(label))
-         #print "index=%d, score=%f, label=%d\n" % (index,score,label)
-         #if index>10:
-         #    break
  
!     print len(indexes)
!     jindexes = jarray.array(indexes,'i')
!     jscores = jarray.array(scores,'d')
!     jlabels = jarray.array(labels,'i')
  
!     print jscores[3:30]
!     print jlabels[3:30]
!     if iter==20:
!         break
  
! import jboost.visualization
! d=jboost.visualization.DataSet(jindexes,jscores,jlabels)
! print d.getMin()
! v=jboost.visualization.HistogramFrame(d)
! v.show()
  
!     
!     
--- 1,80 ----
! import sys,os,os.path,re,math,glob
  import jarray
+ from java.util import Vector
  
! info_path = '/Users/yoavfreund/Desktop/Galit_Lahav/Raw_tifs/28/jboost/t28.0/ADD_ALL/'
  
  #args = sys.argv[1:]
  #if len(args) != 1:
! #    sys.exit("""
! #    Usage: VisualizeScore <info-files-path>
! #    info-file-path is a directory containing files with names like trial0.test.boosting.info, trial2.train.boosting.info ...
! #    Example: VisualizeScore /Users/yoavfreund/Desktop/Galit_Lahav/Raw_tifs/28/jboost/t28.0/ADD_ALL/'
! #    """)
  #filename = args[0]
  
  header_p = re.compile('iteration=(\d+): elements=(\d+):')
  scoreline_p = re.compile('([-+\d.]+):\s+') #([+-1]+):')
  
+ #testfiles = glob.glob(info_path+"trial*.test.boosting.info")
+ 
+ testfiles = glob.glob("/Users/yoavfreund/Downloads/dc_boost_active_output/iter?/*.test.boosting.info")
+ 
+ indexes={}
+ scores={}
+ labels={}
+ 
+ #find the iteration indices
+ 
+ infile = open(testfiles[0],'r')
+ iterList=[]
+ iterDict={}
+ iterationIndex=0
  for line in infile:
      [(iter,elements)] = header_p.findall(line)
      elements = int(elements)
! 
!     if iterDict.has_key(iter):
!         sys.exit("data file "+testfiles[0]+" has two lists corresponding to iteration "+iter)
!     iterList.append(iter)
!     iterDict[iter]=iterationIndex
!     iterationIndex=iterationIndex+1
! 
      for count in range(elements):
          line=infile.readline()
  
! from jboost.visualization import DataSet,DataElement,HistogramFrame
  
! d=DataSet(iterList)
  
! # read the data into the Dataset data structures.
! for filename in testfiles:
!     print filename
!     infile = open(filename,'r')
!     iterList=[];
!     for line in infile:
!         print line
!         [(iter,elements)] = header_p.findall(line)
!         iterationIndex=iterDict[iter]
!         elements = int(elements)
!         print "iter=%s iterationIndex=%d elements=%d\n" % (iter,iterationIndex, elements)
  
!         for count in range(elements):
!             line=infile.readline()
!             #print line
!             a = scoreline_p.findall(line)
!             index=int(a[0])
!             score=float(a[2])
!             label=int(a[-1])
!             e=DataElement(score,index,label)
!             d.addDataElement(e,iterationIndex)
! 
! print "going to preprocess\n"
! d.preProcessDataset()
! print "finisehd preprocessing\n"
! 
! print "min= %f, max=%f\n" % (d.getMin(),d.getMax())
! 
! v=HistogramFrame(d)
! 
! v.show()