strdelta.append('-%s' % (f,))
elif ';' in f1 or ';' in f2:
f1 = f1.split(';')
f2 = f2.split(';')
minus = ['-%s' % (e,) for e in f1 if e != '' and e not in f2]
plus = ['+%s' % (e,) for e in f2 if e != '' and e not in f1]
if minus or plus:
strdelta.append('%s(%s%s%s)' % (f, ','.join(minus), (' | ' if (plus and minus) else ''), ','.join(plus)))
else:
strdelta.append('%s(%s->%s)' % (f, f1[:4], f2[:4]))
return ', '.join(strdelta)
if __name__ == "__main__":
qlist, questions = answerfv.load_questions(sys.argv[1])
for qid in qlist:
print("[%s] %s (%s)" % (questions[qid]['id'], questions[qid]['text'], questions[qid]['anspattern']))
alist1, ans1 = answerfv.load_answers(sys.argv[2], qid)
alist2, ans2 = answerfv.load_answers(sys.argv[3], qid)
# List answers in the order of confidence in set #2
for a in alist2:
if a not in ans1:
continue # Not common answer
a1 = ans1[a]
a2 = ans2[a]
delta = compare_fv(ans2['_header'], a1, a2)
if delta == []:
continue
print("\t%-38.38s\t%s %.3f\t%s" % (a, a2['iM'], float(a2['confidence']), delta2str(delta)))
print('')
print("%38.38s" % ('', ) + "\tin %Q\tin %A\tavg%A\tavg%CA\tavg%CAp\tRR B")
for field in qacounter.columns:
print("%38.38s\t%.3f\t%.3f\t%.3f\t%.3f\t%.3f\t%c%c %c" % (
field,
qacounter.portion_questions(field),
qacounter.portion_answers(field),
qacounter.mean_portion_all_answers(field),
qacounter.mean_portion_correct_answers(field),
qacounter.mean_portion_answers_correct(field),
'!' if qacounter.policy_too_rare(field) else '.',
'!' if qacounter.policy_extremely_rare(field) else '.',
'!' if qacounter.policy_bad_predictor(field) else '.',
))
if __name__ == "__main__":
""" Gather data """
qlist, questions = answerfv.load_questions(sys.argv[1])
qacounter = QACounter()
for qid in qlist:
alist, ans = answerfv.load_answers(sys.argv[2], qid)
acounter = AnswerCounter()
acounter.from_alist(qid, alist, ans)
qacounter.add_question(acounter)
""" Report data """
stats_report(qacounter)
plus = ['+%s' % (e, ) for e in f2 if e != '' and e not in f1]
if minus or plus:
strdelta.append('%s(%s%s%s)' %
(f, ','.join(minus),
(' | ' if
(plus and minus) else ''), ','.join(plus)))
else:
strdelta.append('%s(%s->%s)' % (f, f1[:4], f2[:4]))
return ', '.join(strdelta)
if __name__ == "__main__":
qlist, questions = answerfv.load_questions(sys.argv[1])
for qid in qlist:
print("[%s] %s (%s)" % (questions[qid]['id'], questions[qid]['text'],
questions[qid]['anspattern']))
alist1, ans1 = answerfv.load_answers(sys.argv[2], qid)
alist2, ans2 = answerfv.load_answers(sys.argv[3], qid)
# List answers in the order of confidence in set #2
for a in alist2:
if a not in ans1:
continue # Not common answer
a1 = ans1[a]
a2 = ans2[a]
delta = compare_fv(ans2['_header'], a1, a2)
if delta == []:
continue
print("\t%-38.38s\t%s %.3f\t%s" %
(a, a2['iM'], float(a2['confidence']), delta2str(delta)))
print('')
for field in qacounter.columns:
print("%38.38s\t%.3f\t%.3f\t%.3f\t%.3f\t%.3f\t%c%c %c" %
(field,
qacounter.portion_questions(field),
qacounter.portion_answers(field),
qacounter.mean_portion_all_answers(field),
qacounter.mean_portion_correct_answers(field),
qacounter.mean_portion_answers_correct(field),
'!' if qacounter.policy_too_rare(field) else '.',
'!' if qacounter.policy_extremely_rare(field) else '.',
'!' if qacounter.policy_bad_predictor(field) else '.',
))
if __name__ == "__main__":
""" Gather data """
qlist, questions = answerfv.load_questions(sys.argv[1])
qacounter = QACounter()
for qid in qlist:
alist, ans = answerfv.load_answers(sys.argv[2], qid)
acounter = AnswerCounter()
acounter.from_alist(qid, alist, ans)
qacounter.add_question(acounter)
""" Report data """
stats_report(qacounter)
