Mode Enum (#119)

* add `Mode` Enum to `pymilo_client.py`

* update testcases accordingly

* updaet according to Enum refactorings

* add port to PyMiloServer init params

* Update pymilo_client.py

Author: AHReccese

pymilo/streaming/pymilo_client.py

@@ -1,14 +1,24 @@
+from enum import Enum
 from .encryptor import DummyEncryptor
 from .compressor import DummyCompressor
 from ..pymilo_obj import Export, Import
 from .communicator import RESTClientCommunicator
 from ..transporters.general_data_structure_transporter import GeneralDataStructureTransporter
+
+
+class Mode(Enum):
+    """fallback state of the PyMiloClient."""
+
+    LOCAL = 1
+    DELEGATE = 2
+
+
 class PymiloClient:
 
     def __init__(
             self,
             model=None,
-            mode="LOCAL",
+            mode=Mode.LOCAL,
             server="http://127.0.0.1",
             port= 8000
             ):
@@ -22,9 +32,8 @@ def __init__(
             server_url="{}:{}".format(server, port)
         )
 
-    def toggle_mode(self, mode="LOCAL"):
-        mode = mode.upper()
-        if mode not in ["LOCAL", "DELEGATE"]:
+    def toggle_mode(self, mode=Mode.LOCAL):
+        if mode not in Mode.__members__.values():
             raise Exception("Invalid mode, the given mode should be either `LOCAL`[default] or `DELEGATE`.")
         self._mode = mode
 
@@ -52,12 +61,12 @@ def upload(self):
             print("Local model upload failed.")
 
     def __getattr__(self, attribute):
-        if self._mode == "LOCAL":
+        if self._mode == Mode.LOCAL:
             if attribute in dir(self._model):
                 return getattr(self._model, attribute)
             else:
                 raise AttributeError("This attribute doesn't exist either in PymiloClient or the inner ML model.")
-        elif self._mode == "DELEGATE":
+        elif self._mode == Mode.DELEGATE:
             gdst = GeneralDataStructureTransporter()
             def relayer(*args, **kwargs):
                 print(f"Method '{attribute}' called with args: {args} and kwargs: {kwargs}")

pymilo/streaming/pymilo_server.py

@@ -7,11 +7,11 @@
 
 class PymiloServer:
 
-    def __init__(self):
+    def __init__(self, port=8000):
         self._model = None
         self._compressor = DummyCompressor()
         self._encryptor = DummyEncryptor()
-        self._communicator = RESTServerCommunicator(ps=self)
+        self._communicator = RESTServerCommunicator(ps=self, port=port)
         self._communicator.run()
 
     def export_model(self):

tests/test_ml_streaming/scenarios/scenario1.py

@@ -1,7 +1,7 @@
 import numpy as np
 from sklearn.metrics import mean_squared_error
 from sklearn.linear_model import LinearRegression
-from pymilo.streaming.pymilo_client import PymiloClient
+from pymilo.streaming.pymilo_client import PymiloClient, Mode
 from pymilo.utils.data_exporter import prepare_simple_regression_datasets
 
 
@@ -20,7 +20,7 @@ def scenario1():
 
     # 2.
     linear_regression.fit(x_train, y_train)
-    client = PymiloClient(model=linear_regression, mode="LOCAL")
+    client = PymiloClient(model=linear_regression, mode=Mode.LOCAL)
 
     # 3.
     result = client.predict(x_test)

tests/test_ml_streaming/scenarios/scenario2.py

@@ -1,7 +1,7 @@
 import numpy as np
 from sklearn.metrics import mean_squared_error
 from sklearn.linear_model import LinearRegression
-from pymilo.streaming.pymilo_client import PymiloClient
+from pymilo.streaming.pymilo_client import PymiloClient, Mode
 from pymilo.utils.data_exporter import prepare_simple_regression_datasets
 
 
@@ -17,13 +17,13 @@ def scenario2():
     # 1.
     x_train, y_train, x_test, y_test = prepare_simple_regression_datasets()
     linear_regression = LinearRegression()
-    client = PymiloClient(model=linear_regression, mode="LOCAL")
+    client = PymiloClient(model=linear_regression, mode=Mode.LOCAL)
 
     # 2.
     client.upload()
 
     # 3.
-    client.toggle_mode(mode="DELEGATE")
+    client.toggle_mode(Mode.DELEGATE)
     client.fit(x_train, y_train)
 
     # 4.
@@ -34,7 +34,7 @@ def scenario2():
     client.download()
 
     # 6.
-    client.toggle_mode(mode="LOCAL")
+    client.toggle_mode(mode=Mode.LOCAL)
     result = client.predict(x_test)
     mse_local = mean_squared_error(y_test, result)
     return np.abs(mse_server-mse_local)