def polynomial_expansion_test(self):
polynomial_exp = PolynomialFeatures(degree=2, include_bias=False)
polynomial_exp.mlinit(input_features='a',
output_features='poly')
Xres = polynomial_exp.fit_transform(self.df[['a']])
self.assertEqual(Xres[0][1], Xres[0][0] * Xres[0][0])
polynomial_exp.serialize_to_bundle(self.tmp_dir, polynomial_exp.name)
expected_model = {
"op": "polynomial_expansion",
"attributes": {
"degree": {
"type": "double",
"value": 2
}
}
}
# Test model.json
with open("{}/{}.node/model.json".format(self.tmp_dir, polynomial_exp.name)) as json_data:
model = json.load(json_data)
self.assertEqual(expected_model['attributes']['degree']['value'], model['attributes']['degree']['value'])
def polynomial_expansion_test(self):
extract_features = ['a']
feature_extractor = FeatureExtractor(
input_scalars=['a'],
output_vector='extracted_a_output',
output_vector_items=["{}_out".format(x) for x in extract_features])
polynomial_exp = PolynomialFeatures(degree=2, include_bias=False)
polynomial_exp.mlinit(prior_tf=feature_extractor,
output_features='poly')
Xres = polynomial_exp.fit_transform(self.df[['a']])
self.assertEqual(Xres[0][1], Xres[0][0] * Xres[0][0])
polynomial_exp.serialize_to_bundle(self.tmp_dir, polynomial_exp.name)
expected_model = {
"op": "sklearn_polynomial_expansion",
"attributes": {
"combinations": {
"string": "[x0,x0^2]"
}
}
}
# Test model.json
with open("{}/{}.node/model.json".format(
self.tmp_dir, polynomial_exp.name)) as json_data:
model = json.load(json_data)
self.assertEqual(expected_model['op'], model['op'])
self.assertEqual(
expected_model['attributes']['combinations']['string'],
model['attributes']['combinations']['string'])
# Test node.json
with open("{}/{}.node/node.json".format(
self.tmp_dir, polynomial_exp.name)) as json_data:
node = json.load(json_data)
self.assertEqual(polynomial_exp.name, node['name'])
self.assertEqual(polynomial_exp.input_features,
node['shape']['inputs'][0]['name'])
self.assertEqual(polynomial_exp.output_features,
node['shape']['outputs'][0]['name'])
def polynomial_expansion_deserializer_test(self):
polynomial_exp = PolynomialFeatures(degree=2, include_bias=False)
polynomial_exp.mlinit(
input_features=['a', 'b'],
output_features=['a', 'b', 'a_sqd', 'a_mult_b', 'b_sqd'])
Xres = polynomial_exp.fit_transform(self.df[['a', 'b']])
self.assertEqual(Xres[0][2], Xres[0][0] * Xres[0][0])
polynomial_exp.serialize_to_bundle(self.tmp_dir, polynomial_exp.name)
# Deserialize the PolynomialExpansion
node_name = "{}.node".format(polynomial_exp.name)
poly_tf_ds = PolynomialFeatures()
poly_tf_ds.deserialize_from_bundle(self.tmp_dir, node_name)
# Transform some sample data
res_a = polynomial_exp.transform(self.df[['a', 'b']])
res_b = poly_tf_ds.transform(self.df[['a', 'b']])
self.assertEqual(res_a[0][0], res_b[0][0])
self.assertEqual(res_a[1][0], res_b[1][0])
self.assertEqual(res_a[2][0], res_b[2][0])
self.assertEqual(res_a[3][0], res_b[3][0])
self.assertEqual(res_a[0][1], res_b[0][1])
self.assertEqual(res_a[1][1], res_b[1][1])
self.assertEqual(res_a[2][1], res_b[2][1])
self.assertEqual(res_a[3][1], res_b[3][1])