def normaliser():
"""
Returns default normaliser, already configured.
"""
n = NormalisationRuleEngine()
n.load_rules(os.path.join(os.path.split(__file__)[0], 'rules', 'normalisation'))
return n
def normaliser():
"""
Returns default normaliser, already configured.
"""
n = NormalisationRuleEngine()
n.load_rules(
os.path.join(os.path.split(__file__)[0], 'rules', 'normalisation'))
return n
def testBadErrors(self):
r = NormalisationRuleEngine()
try:
r.load_rules(os.path.join(os.path.dirname(__file__), 'test_normalisation_rules_malformed/'))
except RuleLoadErrors as e:
self.assertEqual(len(e.errors), 12, "These errors were raised: " + str(e))
else:
self.fail('No exceptions were raised/caught')
def testBadErrors(self):
r = NormalisationRuleEngine()
try:
r.load_rules(os.path.join(os.path.dirname(__file__), 'test_normalisation_rules_malformed/'))
except RuleLoadErrors as e:
self.assertEquals(len(e.errors), 12, "These errors were raised: " + str(e))
else:
self.fail('No exceptions were raised/caught')
def testTag(self):
e = NormalisationRuleEngine()
e.load_rules(os.path.join(os.path.dirname(__file__), 'test_normalisation_rules'))
t = Timex(type='date')
e.annotate([[('We', 'POS', set()),
('took', 'POS', set()),
('a', 'POS', set()),
('plane', 'POS', set()),
('on', 'POS', set()),
('the', 'POS', set()),
('06', 'POS', {t}),
('th', 'POS', {t}),
('January', 'POS', {t}),
('1996', 'POS', {t}),
('to', 'POS', set()),
('Atlanta', 'POS', set())]], '')
self.assertEquals(t.value, '19960106')
def testTag(self):
e = NormalisationRuleEngine()
e.load_rules(os.path.join(os.path.dirname(__file__), 'test_normalisation_rules'))
t = Timex(type='date')
e.annotate([[('We', 'POS', set()),
('took', 'POS', set()),
('a', 'POS', set()),
('plane', 'POS', set()),
('on', 'POS', set()),
('the', 'POS', set()),
('06', 'POS', {t}),
('th', 'POS', {t}),
('January', 'POS', {t}),
('1996', 'POS', {t}),
('to', 'POS', set()),
('Atlanta', 'POS', set())]], '')
self.assertEqual(t.value, '19960106')
