class TestLinearRegressionIntegration(TestCase):
def setUp(self):
df = pd.read_csv(path.join(BASE_DIR, '../models/categorical-test.csv'))
Xte = df.iloc[:, 1:]
yte = df.iloc[:, 0]
self.test = (Xte, yte)
pmml = path.join(BASE_DIR, '../models/linear-regression.pmml')
self.clf = PMMLLinearRegression(pmml)
def test_predict(self):
Xte, _ = self.test
ref = np.array([0.959047661, 0.306635481, 0.651029985, 0.954874880, 0.889268596, 0.874413539, 0.552911965, 0.793971929, 0.567604727, 0.694441984, 0.977588079, 1.020076443, 0.938209348, 0.741296266, 0.785681506, 0.783314305, 0.147203243, 0.953499858, 0.861694209, 0.818535888, 1.054586791, 0.508564304, 0.490740907, 0.692194962, 0.546339084, 0.584074930, 0.817451147, 0.007120341, -0.023103301, 0.354232979, 0.452602313, -0.232817829, 0.289612034, 0.241502904, 0.098388728, 0.413283786, 0.349024715, 0.315999598, 0.656973238, 0.525739661, 0.243258999, 0.128203855, 0.151826018, 0.357043960, 0.647876971, 0.405659892, 0.264334997, 0.280004394, 0.948749766, -0.028252457, 0.415301011, 0.509803923])
assert np.allclose(ref, self.clf.predict(Xte))
def test_score(self):
Xte, yte = self.test
ref = 0.409378064635437
assert ref == self.clf.score(Xte, yte == 'Yes')
def test_fit_exception(self):
with self.assertRaises(Exception) as cm:
self.clf.fit(np.array([[]]), np.array([]))
assert str(cm.exception) == 'Not supported.'
def setUp(self):
df = pd.read_csv(path.join(BASE_DIR, '../models/categorical-test.csv'))
Xte = df.iloc[:, 1:]
yte = df.iloc[:, 0]
self.test = (Xte, yte)
pmml = path.join(BASE_DIR, '../models/linear-regression.pmml')
self.clf = PMMLLinearRegression(pmml)
def get_model(self, model_url):
print("model_url was: {0}".format(model_url))
if not model_url:
self.model = "No Model To Use"
if model_url[-3:] == "pkl":
print("found a pickle")
with open(model_url, "rb") as ff:
self.model = pickle.load(ff)
if model_url[-4:] == "pnnm":
self.model = PMMLLinearRegression(model_url)
def test_nonlinear_model(self):
with self.assertRaises(Exception) as cm:
PMMLLinearRegression(pmml=StringIO("""
<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>
<DataField name="a" optype="continuous" dataType="double"/>
</DataDictionary>
<RegressionModel>
<MiningSchema>
<MiningField name="Class" usageType="target"/>
</MiningSchema>
<RegressionTable>
<NumericPredictor name="a" exponent="1" coefficient="1"/>
<NumericPredictor name="a" exponent="1" coefficient="1"/>
</RegressionTable>
</RegressionModel>
</PMML>
"""))
assert str(cm.exception) == 'PMML model is not linear.'
def test_invalid_model(self):
with self.assertRaises(Exception) as cm:
PMMLLinearRegression(pmml=StringIO("""
<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>
<MiningSchema>
<MiningField name="Class" usageType="target"/>
</MiningSchema>
</PMML>
"""))
assert str(cm.exception) == 'PMML model does not contain RegressionModel.'
class CTModel():
def __init__(self):
self.TMP_MODEL_HOME = "/tmp"
self.model = "I wish I was a model"
def get_model(self, model_url):
print("model_url was: {0}".format(model_url))
if not model_url:
self.model = "No Model To Use"
if model_url[-3:] == "pkl":
print("found a pickle")
with open(model_url, "rb") as ff:
self.model = pickle.load(ff)
if model_url[-4:] == "pnnm":
self.model = PMMLLinearRegression(model_url)
#if not model_url.contains("."):
# wget.download(model_url, self.TMP_MODEL_HOME)
# model = h2o.loadModel(TMP_MODEL_HOME)
def show_model(self):
if self.model.type == "h2o":
model_summary = self.model.model_performance()
if self.model.type == "pickle":
model_summary = self.model.do_something()
if self.model.type == "pnnm":
model_summary = self.model.do_something()
return model_summary
def regression_results(self, y_true, y_pred):
# Regression metrics
explained_variance = metrics.explained_variance_score(y_true, y_pred)
mean_absolute_error = metrics.mean_absolute_error(y_true, y_pred)
mse = metrics.mean_squared_error(y_true, y_pred)
mean_squared_log_error = metrics.mean_squared_log_error(y_true, y_pred)
median_absolute_error = metrics.median_absolute_error(y_true, y_pred)
r2 = metrics.r2_score(y_true, y_pred)
fpr, tpr, thresholds = metrics.roc_curve(y_true, y_pred, pos_label=2)
auc = metrics.auc(fpr, tpr)
model_scores = {
'explained_variance: ': round(explained_variance, 4),
'AUC: ': auc,
'R2: ': round(r2, 4),
'MAE: ': round(mean_absolute_error, 4),
'MSE: ': round(mse, 4),
'RMSE: ': round(np.sqrt(mse), 4),
'mean_squared_log_error: ': round(mean_squared_log_error, 4)
}
return model_scores
def fit_predict_and_score(self, X, Y):
self.model.fit(X, Y)
Y_pred = self.model.predict(X)
scores = self.regression_results(Y, Y_pred)
self.scores = scores
return scores
def make_graph(self):
df = pd.DataFrame(
dict(r=list(self.scores.values()), theta=list(self.scores.keys())))
fig = px.line_polar(df, r='r', theta='theta', line_close=True)
fig.update_traces(fill='toself')
figfile = io.BytesIO()
fig.write_image(figfile, format='png')
figfile.seek(0) # rewind to beginning of file
figbuffer = b''.join(figfile)
figdata_png = base64.b64encode(figbuffer)
print("im in the graph_maker")
return figdata_png.decode('utf8')