def apply(self):
if self.data:
pb = OWGUI.ProgressBar(self, iterations=len(self.data))
self.data = orange.ExampleTable(orange.Domain(self.tmpDom), self.tmpData)
newdata = orngText.extractLetterNGram(self.data, self.size + 2, textAttribute=self.textAttributePos, callback=pb.advance)
self.lblFeatureNo.setText("\nNo. of features: \n%d" % len(newdata.domain.getmetas(orngText.TEXTMETAID)))
self.send("Example Table", newdata)
pb.finish()
else:
self.send("Example Table", None)
if __name__ == "__main__":
#from orngTextCorpus import *
import pickle, orngText
## os.chdir("/home/mkolar/Docs/Diplomski/repository/orange/OrangeWidgets/Other/")
appl = QApplication(sys.argv)
ow = OWCorrAnalysis()
#owb = OWBagofWords.OWBagofWords()
t = orngText.loadFromXML(r'c:\test\orange\msnbc.xml')
#owb.data = t
#owb.show()
stop = orngText.loadWordSet(r'C:\tmtorange\common\en_stopwords.txt')
p = orngText.Preprocess(language='hr')
print('Done with loading')
t1 = orngText.extractLetterNGram(t, 2)
#t1 = orngText.extractWordNGram(t, stopwords = stop, measure = 'MI', threshold = 7, n=2)
#t1 = orngText.extractWordNGram(t1, stopwords = stop, measure = 'MI', threshold = 10, n=3)
#t1 = orngText.extractNamedEntities(t, stopwords = stop)
#t1 = orngText.bagOfWords(t1, stopwords = stop)
print(len(t1.domain.getmetas(orngText.TEXTMETAID)))
print('Done with extracting')
#t2 = orngText.FSS(t1, 'TF', 'MIN', 0.98)
#print len(t2.domain.getmetas())
print('Done with feature selection')
appl.setMainWidget(ow)
#t3 = orngText.DSS(t2, 'WF', 'MIN', 1)
#print 'Done with document selection'
ow.dataset(t1)
print('Done')
ow.show()
if __name__=="__main__":
#from orngTextCorpus import *
import cPickle, orngText
## os.chdir("/home/mkolar/Docs/Diplomski/repository/orange/OrangeWidgets/Other/")
appl = QApplication(sys.argv)
ow = OWCorrAnalysis()
#owb = OWBagofWords.OWBagofWords()
t = orngText.loadFromXML(r'c:\test\orange\msnbc.xml')
#owb.data = t
#owb.show()
stop = orngText.loadWordSet(r'C:\tmtorange\common\en_stopwords.txt')
p = orngText.Preprocess(language = 'hr')
print 'Done with loading'
t1 = orngText.extractLetterNGram(t, 2)
#t1 = orngText.extractWordNGram(t, stopwords = stop, measure = 'MI', threshold = 7, n=2)
#t1 = orngText.extractWordNGram(t1, stopwords = stop, measure = 'MI', threshold = 10, n=3)
#t1 = orngText.extractNamedEntities(t, stopwords = stop)
#t1 = orngText.bagOfWords(t1, stopwords = stop)
print len(t1.domain.getmetas(orngText.TEXTMETAID))
print 'Done with extracting'
#t2 = orngText.FSS(t1, 'TF', 'MIN', 0.98)
#print len(t2.domain.getmetas())
print 'Done with feature selection'
appl.setMainWidget(ow)
#t3 = orngText.DSS(t2, 'WF', 'MIN', 1)
#print 'Done with document selection'
ow.dataset(t1)
print 'Done'
ow.show()
maxword = word
if freq < min:
min = freq
minword = word
sum += freq
pb.advance()
avg = sum / len(words)
if min == ():
min = 0
self.lblFeatNo.setText("No. of features: %d" % len(words))
self.lblMin.setText("Min: %d Min word = %s" % (min, minword))
self.lblMax.setText("Max: %d Max word = %s" % (max,maxword))
self.lblAvg.setText("Avg: %.3f" % avg)
pb.finish()
def selectionChanged(self):
if self.data:
self.applyButton.setDisabled(0)
if __name__ == "__main__":
a = QApplication(sys.argv)
t = orngText.loadFromXML(r'c:\test\orange\msnbc.xml')
t2 = orngText.extractLetterNGram(t)
#print t2.domain.getmetas().values()
ow = OWTextFeatureSelection()
a.setMainWidget(ow)
ow.show()
ow.dataset(t2)
a.exec_loop()
