def test_get_type_exception(self):
template = """
<PMML xmlns="http://www.dmg.org/PMML-4_3" version="4.3">
<DataDictionary>
<DataField name="test" optype="{}" dataType="{}"/>
</DataDictionary>
</PMML>"""
# Test invalid data type
pmml = StringIO(template.format("continuous", "does_not_exist"))
clf = PMMLBaseEstimator(pmml)
data_dictionary = clf.root.find("DataDictionary")
data_field = data_dictionary.find("DataField")
with self.assertRaises(Exception) as cm:
get_type(data_field)
assert str(cm.exception) == "Unsupported data type."
# Test invalid operation type
pmml = StringIO(template.format("does_not_exist", "string"))
clf = PMMLBaseEstimator(pmml)
data_dictionary = clf.root.find("DataDictionary")
data_field = data_dictionary.find("DataField")
with self.assertRaises(Exception) as cm:
get_type(data_field)
assert str(cm.exception) == "Unsupported operation type."
def test_parse_type_value_continuous(self):
template = """
<PMML xmlns="http://www.dmg.org/PMML-4_3" version="4.3">
<DataDictionary>
<DataField name="test" optype="{}" dataType="{}"/>
</DataDictionary>
</PMML>"""
values = ["1234", 1234, 12.34, True]
types = [
['continuous', 'integer', int],
['continuous', 'float', float],
['continuous', 'double', float],
]
for value in values:
for field_type in types:
op_type, pmml_type, data_type = field_type
clf = PMMLBaseEstimator(StringIO(template.format(op_type, pmml_type)))
data_dictionary = clf.root.find("DataDictionary")
data_field = data_dictionary.find("DataField")
result = get_type(data_field)(value)
assert isinstance(result, data_type)
def test_parse_type_value_continuous_boolean(self):
template = """
<PMML xmlns="http://www.dmg.org/PMML-4_3" version="4.3">
<DataDictionary>
<DataField name="test" optype="{}" dataType="{}"/>
</DataDictionary>
</PMML>"""
tests = {
"1": True,
"True": True,
"YES": True,
1: True,
True: True,
"0": False,
"False": False,
0: False
}
for value, expected in tests.items():
pmml = StringIO(template.format('continuous', 'boolean'))
clf = PMMLBaseEstimator(pmml)
data_dictionary = clf.root.find("DataDictionary")
data_field = data_dictionary.find("DataField")
result = get_type(data_field)(value)
assert isinstance(result, bool)
assert result == expected
def test_get_type_ordinal(self):
template = """
<PMML xmlns="http://www.dmg.org/PMML-4_3" version="4.3">
<DataDictionary>
<DataField name="Volume" optype="ordinal" dataType="string">
<Value value="loud"/>
<Value value="louder"/>
<Value value="loudest"/>
</DataField>
</DataDictionary>
</PMML>"""
clf = PMMLBaseEstimator(pmml=StringIO(template))
data_dictionary = clf.root.find("DataDictionary")
data_field = data_dictionary.find("DataField")
data_type: Category = get_type(data_field)
assert data_type.categories == ['loud', 'louder', 'loudest']
assert data_type.ordered
def test_get_type_categorical(self):
template = """
<PMML xmlns="http://www.dmg.org/PMML-4_3" version="4.3">
<DataDictionary>
<DataField name="Class" optype="categorical" dataType="string">
<Value value="setosa"/>
<Value value="versicolor"/>
<Value value="virginica"/>
</DataField>
</DataDictionary>
</PMML>"""
clf = PMMLBaseEstimator(pmml=StringIO(template))
data_dictionary = clf.root.find("DataDictionary")
data_field = data_dictionary.find("DataField")
data_type: Category = get_type(data_field)
assert data_type.categories == ['setosa', 'versicolor', 'virginica']
assert not data_type.ordered