def label_from_string(text):
text = text.strip()
prio = persist.get_classifier('prio').classify(ngen(text))
text = ' ' + text + ' '
out, text, date = parse_date(text)
ttype = {
Task.PRIORITY_HIGH: 'Important task due',
Task.PRIORITY_MEDIUM: 'Assignment due',
Task.PRIORITY_LOW: 'Task for',
Task.PRIORITY_ADMIN: 'Appointment on',
}[prio]
return (" %s %s" % (ttype, out)), text.strip(), date, prio
def prio_changed(self, renderer, path, iter):
if self.concurrent_modification:
return
miter = self.model.iter_nth_child(None, int(path))
task = self.model.get_value(miter, 0)
task.prio = task.prio_match(self.prio_store.get_value(iter, 0))
self.remove(miter)
self.add(task)
self.persist.sync()
mc = persist.get_classifier('prio')
mc.update(task.prio, ngen(task.text), task.text)
mc.save()
def report(self):
print "num labels", len(self.weights)
print "dimensions", map(len, self.weights.values())
if __name__ == '__main__':
classifier = MiraClassifier()
total, ok = 0.0, 0.0
while True:
print
print "Enter a training line:",
try:
line = raw_input()
except EOFError:
break
ngrams = ngen(line)
label = classifier.classify(ngrams)
print "---> the right label [guessed %s]:" % label,
line = raw_input()
if line:
classifier.update(line, ngrams)
else:
ok += 1
total += 1
classifier.report()
print "cumulative accuracy:", ok/total
for k, weights in classifier.weights.items():
if k is None:
continue
print
print k, '-', '.'.join(map(
