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 type in types:
optype, pmml_type, data_type = type
clf = PMMLBaseEstimator(pmml=StringIO(template.format(optype, pmml_type)))
data_dictionary = clf.find(clf.root, "DataDictionary")
data_field = clf.find(data_dictionary, "DataField")
result = clf.parse_type(value, data_field)
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():
clf = PMMLBaseEstimator(pmml=StringIO(template.format('continuous', 'boolean')))
data_dictionary = clf.find(clf.root, "DataDictionary")
data_field = clf.find(data_dictionary, "DataField")
result = clf.parse_type(value, data_field)
assert isinstance(result, Boolean)
assert result == expected
def test_parse_type_value_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.find(clf.root, "DataDictionary")
data_field = clf.find(data_dictionary, "DataField")
with self.assertRaises(Exception) as cm: clf.parse_type("not_in_category", data_field)
assert str(cm.exception) == "Value does not match any category."
assert clf.parse_type("setosa", data_field) == "setosa"
assert clf.parse_type("versicolor", data_field) == "versicolor"
assert clf.parse_type("virginica", data_field) == "virginica"
def test_parse_type_interval(self):
template = """
<PMML xmlns="http://www.dmg.org/PMML-4_3" version="4.3">
<DataDictionary>
<DataField name="test" optype="ordinal" dataType="float">
<Interval closure="openOpen" rightMargin="1"/>
<Interval closure="openClosed" leftMargin="1" rightMargin="1.5"/>
<Interval closure="openOpen" leftMargin="1.5" rightMargin="2.5"/>
<Interval closure="closedOpen" leftMargin="2.5" rightMargin="3.5"/>
<Interval closure="closedClosed" leftMargin="3.5" />
</DataField>
</DataDictionary>
</PMML>"""
clf = PMMLBaseEstimator(pmml=StringIO(template))
data_dictionary = clf.find(clf.root, "DataDictionary")
data_field = clf.find(data_dictionary, "DataField")
assert clf.parse_type(-1, data_field) == Interval(-1, rightMargin=1, closure='openOpen')
with self.assertRaises(Exception): clf.parse_type(1, data_field)
assert clf.parse_type(2, data_field) == Interval(2, leftMargin=1.5, rightMargin=2.5, closure='openOpen')
assert clf.parse_type(2.5, data_field) == Interval(2.5, leftMargin=2.5, rightMargin=3.5, closure='closedOpen')
assert clf.parse_type(3.5, data_field) == Interval(3.5, leftMargin=3.5, closure='closedClosed')
def test_parse_type_value_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
clf = PMMLBaseEstimator(pmml=StringIO(template.format("continuous", "does_not_exist")))
data_dictionary = clf.find(clf.root, "DataDictionary")
data_field = clf.find(data_dictionary, "DataField")
with self.assertRaises(Exception) as cm: clf.parse_type("test", data_field)
assert str(cm.exception) == "Unsupported data type."
# Test invalid operation type
clf = PMMLBaseEstimator(pmml=StringIO(template.format("does_not_exist", "string")))
data_dictionary = clf.find(clf.root, "DataDictionary")
data_field = clf.find(data_dictionary, "DataField")
with self.assertRaises(Exception) as cm: clf.parse_type("test", data_field)
assert str(cm.exception) == "Unsupported operation type."