def runTest(self):
# empty list
ds = sentdoc.PottsDocSentiScore()
L = sentlex.MobyLexicon()
ds.verbose = True
ds.set_lexicon(L)
negated_sent = 'not/DT good/JJ'
ds.set_active_pos(True, False, False, False)
ds.set_parameters(negation=True, negation_window=15)
(dpos, dneg) = ds.classify_document(negated_sent)
self.assertTrue(dneg > dpos, 'Negation did not invert scores.')
ds.set_parameters(negation=True,
negation_window=15,
atenuation=True,
at_pos=0.5,
at_neg=0.5)
(dpos, dneg) = ds.classify_document(negated_sent)
ds.set_parameters(negation=True,
negation_window=15,
atenuation=True,
at_pos=1.0,
at_neg=1.0)
(dposfull, dnegfull) = ds.classify_document(negated_sent)
self.assertTrue(dpos > dneg, 'Negation did not atenuate scores.')
self.assertTrue(dposfull > dpos, 'Negation did not atenuate scores.')
def ds(moby):
ds = sentdoc.PottsDocSentiScore()
ds.verbose = False
ds.set_lexicon(moby)
ds.set_parameters(a=True,
v=False,
n=False,
r=False,
negation=True,
negation_window=15,
negation_adjustment=0.5)
return ds
def runTest(self):
# load lexicon
L = sentlex.MobyLexicon()
self.assertTrue(L.is_loaded, 'Test lexicon did not load correctly')
# create a class that scores only adjectives
ds = sentdoc.PottsDocSentiScore()
ds.verbose = True
ds.set_active_pos(True, False, False, False)
ds.set_parameters(score_mode=ds.SCOREALL,
score_freq=False,
negation=True,
negation_adjustment=0.5)
ds.set_lexicon(L)
# separator ok?
self.assertEqual(ds._detect_tag(TESTDOC_ADJ), '/',
'Unable to detect correct separator')
# now score!
(dpos, dneg) = ds.classify_document(TESTDOC_ADJ, verbose=True)
self.assertTrue(ds.resultdata and ds.resultdata.has_key('doc') and ds.resultdata.has_key('annotated_doc')\
and ds.resultdata.has_key('resultpos') and ds.resultdata.has_key('resultneg'), 'Did not populate resultdata after scoring doc')
self.assertTrue(dpos > 0.0,
'Did not find positive words on positive doc')
print 'TESTDOC_ADJ (pos,neg): %2.2f %2.2f' % (dpos, dneg)
# again, for negative text
(dpos, dneg) = ds.classify_document(TESTDOC_BADADJ, verbose=True)
self.assertTrue(dneg > 0.0,
'Did not find negative words on negative doc')
print 'TESTDOC_BADADJ (pos,neg): %2.2f %2.2f' % (dpos, dneg)
# negated text
ds.set_neg_detection(True, 5)
(dpos, dneg) = ds.classify_document(TESTDOC_NEGATED, verbose=True)
self.assertTrue(dpos > 0.0,
'Did not find positive words on TESTDOC_NEGATED')
print 'TESTDOC_NEGATED (pos,neg): %2.2f %2.2f' % (dpos, dneg)
# currupt data - should still work
(dpos, dneg) = ds.classify_document(TESTDOC_CORRUPT, verbose=True)
self.assertTrue(dpos > dneg,
'Did not process corrupt document correctly')
def runTest(self):
# empty list
ds = sentdoc.PottsDocSentiScore()
ds.verbose = False
ds.set_active_pos(True, False, False, False)
ds.set_parameters(negation_adjustment=0.5,
negation=True,
negation_window=15)
self.assertEqual((ds.a, ds.v, ds.n, ds.r), (True, False, False, False),
'Failed set POS parameters')
self.assertEqual((ds.negation, ds.negation_window), (True, 15),
'Failed set negation')
ds.set_parameters(score_mode=ds.SCOREONCE, score_freq=True)
self.assertEqual(ds.score_mode, ds.SCOREONCE,
'Unable to set parameters via kwards')
self.assertEqual(ds.score_freq, True,
'Unable to set parameters via kwards')