def sklearn_predict_for_handler(handler):
try:
if handler.request.body:
data = json.loads(handler.request.body)
object_id = data.get('objectID', None)
if object_id == None:
raise Exception("please input handlerID")
X_data_id = data.get('XDataID', None)
if X_data_id == None:
raise Exception("please input X_data_id")
x_data_obj = DataStorage.get_data_obj_by_data_id(X_data_id)
ml_obj, clf = MlObject.get_MlObject_by_obj(object_id)
predict_y = clf.predict(x_data_obj.pandas_data.values)
predict_y = pd.DataFrame(predict_y, columns=['predict'])
MlObject.save_MlObject_by_obj_id(clf, object_id)
data_obj_predict_y = DataStorage.create_data_obj_by_pandas_data(
predict_y)
if data_obj_predict_y:
result = {}
result['dataID'] = data_obj_predict_y.data_id
result['columnNames'] = data_obj_predict_y.column_names
handler.write(json.dumps(result))
else:
raise Exception("please input arguments")
return
except Exception as e:
handler.write(str(e))
def sklearn_fit_for_handler(handler, fit_arg_array):
try:
if handler.request.body:
data = json.loads(handler.request.body)
object_id = data.get('objectID', None)
if object_id == None:
raise Exception("please input handlerID")
X_data_id = data.get('XDataID', None)
if X_data_id == None:
raise Exception("please input XDataID")
y_data_id = data.get('yDataID', None)
if y_data_id == None:
raise Exception("please input yDataID")
x_data_obj = DataStorage.get_data_obj_by_data_id(X_data_id)
y_data_obj = DataStorage.get_data_obj_by_data_id(y_data_id)
ml_obj, clf = MlObject.get_MlObject_by_obj(object_id)
sklearn = data.get('sklearn', None)
print(y_data_obj.pandas_data)
if sklearn:
sklearn_arg = regqeust_arg_to_sklearn_arg(
sklearn, fit_arg_array)
clf.fit(x_data_obj.pandas_data.values,
y_data_obj.pandas_data.values, **sklearn_arg)
else:
clf.fit(x_data_obj.pandas_data.values,
y_data_obj.pandas_data.values)
MlObject.save_MlObject_by_obj_id(clf, object_id)
else:
raise Exception("please input arguments")
return
except Exception as e:
handler.write(str(e))
def get(self):
try:
object_id = self.get_argument('objectID', None)
if object_id == None:
raise Exception("please input objectID")
ml_obj, clf = MlObject.get_MlObject_by_obj(object_id)
sklearn_arg = self.get_argument('sklearn', None)
print(sklearn_arg)
if not sklearn_arg:
raise Exception("please input sklearn arg")
sklearn_arg = sklearn_arg.split(",")
result = {}
if 'coef_' in sklearn_arg:
result['coef_'] = clf.coef_.tolist()
if 'intercept_' in sklearn_arg:
result['intercept_'] = clf.intercept_.tolist()
if 'sparse_coef_' in sklearn_arg:
result['sparse_coef_'] = clf.intercept_.tolist()
if 'n_iter_' in sklearn_arg:
result['n_iter_'] = clf.n_iter_.tolist()
if not result:
raise Exception("please input valid sklearn arg")
self.write(json.dumps(result))
return
except Exception as e:
self.write(str(e))
def put(self):
try:
if self.request.body:
data = json.loads(self.request.body)
object_id = data.get('objectID', None)
if object_id == None:
raise Exception("please input handlerID")
sklearn_arg = data.get('sklearn', None)
if not sklearn_arg:
raise Exception("please input sklearn arg")
ml_obj, clf = MlObject.get_MlObject_by_obj(object_id)
set_parameter_from_sklearn_object(clf, ['alpha', 'fit_intercept', 'normalize', 'copy_X', 'max_iter', 'tol', 'solver', 'random_state'], sklearn_arg)
MlObject.save_MlObject_by_obj_id(clf, object_id)
else:
raise Exception("please input arguments")
return
except Exception as e:
self.write(str(e))
def post(self):
try:
if self.request.body:
json_data = json.loads(self.request.body)
Ridge_argstr = ['alpha','fit_intercept', 'normalize', 'copy_X', 'max_iter', 'tol', 'solver', 'random_state']
sklearn_arg = regqeust_arg_to_sklearn_arg(json_data['sklearn'], Ridge_argstr)
ridgeHander = Ridge(**sklearn_arg)
else:
ridgeHander = Ridge()
mlobj = MlObject.create_MlObject_by_obj(ridgeHander)
result = {}
result['objectID'] = mlobj.object_id
self.write(json.dumps(result))
except Exception as e:
self.write(str(e))
def get(self):
try:
object_id = self.get_argument('objectID', None)
if object_id == None:
raise Exception("please input objectID")
ml_obj, clf = MlObject.get_MlObject_by_obj(object_id)
sklearn_arg = self.get_argument('sklearn', None)
if not sklearn_arg:
raise Exception("please input sklearn arg")
sklearn_arg = sklearn_arg.split(",")
result = get_parameter_from_sklearn_object(clf, ['alpha', 'fit_intercept', 'normalize', 'copy_X', 'max_iter', 'tol', 'solver', 'random_state'], sklearn_arg)
self.write(json.dumps(result))
return
except Exception as e:
self.write(str(e))