from reverend.thomas import Bayes

import cPickle


data = cPickle.load(open('IssueTrackerTexts2.pickle', 'rb'))
guesser = Bayes()

from random import randint

def result_sorter(seq):
    seq.sort(lambda x, y: cmp(x[1], y[1]))
    

test_samples = []
for issueid, item in data.items():
    sections = item['sections']
    if 'General' in sections:
        sections.remove('General')
    
    if not sections:
        continue
    
    if randint(1,10)==1:
        test_samples.append(item)
        continue
    
    for section in sections:
        guesser.train(section, item['description'])
        guesser.train(section, item['title'])
    
        
for test_sample in test_samples:
    text = test_sample['description']
    sections = test_sample['sections']
    print sections
    
    result = guesser.guess(text)[:5]
    
    result.sort(lambda x, y: cmp(y[1], x[1])) # x,y : y,x reverses sort
    
    for name, num in result:
        print "%.5f\t%s" % (num, name)
    
    
    print