def test_normalizer_normalizes(self):
res = [
ParseResult('Test', 'Test', 100),
ParseResult('Test', 'Test', 0)
]
conf = TestConfig()
conf.enabled_confidence_normalizers.append(NormalizerStub)
hnc = HierarchicalNormalizerChain(conf, [], [])
results = hnc.normalize(res)
self.assertEqual(1, len(results))
self.assertIsInstance(results[0], ParseResult)
def test_normalizer_with_pc_conf_under_70(self):
date_result = ParseResult('Date', None, 10)
pc_result = ParseResult('Postal Code', None, 40)
results = DateWithPostalCode.normalize([date_result, pc_result])
count = 0
for res in results:
if res.type == 'Date':
count += 1
self.assertEqual(res.confidence, 44)
elif res.type == "Postal Code":
count += 1
self.assertEqual(count, len(results))
def test_normalizer(self):
char_result = ParseResult('Character', None, 25)
pc_result = ParseResult('Postal Code', None, 80)
results = CharacterWithoutBoolean.normalize([char_result, pc_result])
count = 0
for res in results:
if res.type == 'Character':
count += 1
self.assertEqual(res.confidence, 100)
elif res.type == "Postal Code":
count += 1
self.assertEqual(res.confidence, 80)
self.assertEqual(count, len(results))
def test_normalizer(self):
eq_result = ParseResult('Phone', None, 90)
uri_result = ParseResult('URI', None, 80)
results = PhoneWithUri.normalize([eq_result, uri_result])
count = 0
for res in results:
if res.type == 'Phone':
count += 1
self.assertEqual(res.confidence, 65)
elif res.type == "URI":
count += 1
self.assertEqual(res.confidence, 80)
self.assertEqual(count, len(results))
def test_normalizer(self):
eq_result = ParseResult('Equation', None, 90)
pc_result = ParseResult('Postal Code', None, 80)
results = EquationWithPhonePostalCode.normalize([eq_result, pc_result])
count = 0
for res in results:
if res.type == 'Equation':
count += 1
self.assertEqual(res.confidence, 75)
elif res.type == "Postal Code":
count += 1
self.assertEqual(res.confidence, 80)
self.assertEqual(count, len(results))
def result(self, subtype="Unknown", confidence=0, value=None, data=None):
"""
Returns a ParseResult object detailing the results of parsing
:param subtype: parse result subtype
:type subtype: str
:param confidence: how confident we are in this result (1-100)
:type confidence: int
:param value: representation of the parsed data
:type value: mixed
:param additional_data: any additional data a parser wants to provide
:type additional_data: mixed
:return: parse result
:rtype: ParseResult
"""
if confidence == 0:
confidence = self.confidence
if data is None:
data = {}
return ParseResult(p_type=self.type,
subtype=subtype,
confidence=confidence,
value=value,
additional_data=data)
def test_normalize(self):
results = [
ParseResult('Number', 'Integer', 100),
ParseResult('Number', 'Octal', 25),
ParseResult('Date', 'Date', 50)
]
results = IntOctWithPhoneDatePostalCode.normalize(results)
count = 0
for res in results:
count += 1
if res.subtype == 'Integer':
self.assertEqual(10, res.confidence)
elif res.subtype == 'Octal':
self.assertEqual(5, res.confidence)
else:
self.assertEqual(50, res.confidence)
self.assertEqual(count, 3)
def test_normalize(self):
result = ParseResult('Number', None, 70)
results = NumberWithNonNumbers.normalize([result])
count = 0
for res in results:
count += 1
self.assertEqual(res.confidence, 35)
self.assertEqual(count, len(results))