Allow submodules to be imported in python module

python/seldon_core/microservice.py

@@ -184,8 +184,15 @@ def main():
     annotations = load_annotations()
     logger.info("Annotations: %s", annotations)
 
-    interface_file = importlib.import_module(args.interface_name)
-    user_class = getattr(interface_file, args.interface_name)
+    parts = args.interface_name.rsplit(".", 1)
+    if len(parts) == 1:
+        logger.info("Importing ",args.interface_name)
+        interface_file = importlib.import_module(args.interface_name)
+        user_class = getattr(interface_file, args.interface_name)
+    else:
+        logger.info("Importing submodule %s",parts)
+        interface_file = importlib.import_module(args.interface_name)
+        user_class = getattr(interface_file, parts[1])
 
     if args.persistence:
         logger.info('Restoring persisted component')

python/tests/model-template-app2/.s2i/environment

@@ -0,0 +1,4 @@
+MODEL_NAME=mymodule.MyModel
+API_TYPE=REST
+SERVICE_TYPE=MODEL
+PERSISTENCE=0

python/tests/model-template-app2/mymodule/MyModel.py

@@ -0,0 +1,39 @@
+import logging
+logger = logging.getLogger(__name__)
+
+class MyModel(object):
+    """
+    Model template. You can load your model parameters in __init__ from a location accessible at runtime
+    """
+
+    def __init__(self):
+        """
+        Add any initialization parameters. These will be passed at runtime from the graph definition parameters defined in your seldondeployment kubernetes resource manifest.
+        """
+        logger.info("Initializing model")
+
+    def predict(self, X, features_names):
+        """
+        Return a prediction.
+
+        Parameters
+        ----------
+        X : array-like
+        feature_names : array of feature names (optional)
+        """
+        logger.info("Predict called - will run idenity function")
+        return X
+
+    def send_feedback(self, features, feature_names, reward, truth, routing=None):
+        """
+        Handle feedback
+
+        Parameters
+        ----------
+        features : array - the features sent in the original predict request
+        feature_names : array of feature names. May be None if not available.
+        reward : float - the reward
+        truth : array with correct value (optional)
+        """
+        logger.info("Send feedback called")
+        return []

python/tests/test_microservice.py

@@ -87,6 +87,28 @@ def test_model_template_app_rest(tracing):
         response.raise_for_status()
         assert response.json() == {'data': {'ndarray': []}, 'meta': {}}
 
+
+@pytest.mark.parametrize(
+        'tracing', [(False), (True)]
+    )
+def test_model_template_app_rest_submodule(tracing):
+    with start_microservice(join(dirname(__file__), "model-template-app2"),tracing=tracing):
+        data = '{"data":{"names":["a","b"],"ndarray":[[1.0,2.0]]}}'
+        response = requests.get(
+            "http://127.0.0.1:5000/predict", params="json=%s" % data)
+        response.raise_for_status()
+        assert response.json() == {
+            'data': {'names': ['t:0', 't:1'], 'ndarray': [[1.0, 2.0]]}, 'meta': {}}
+
+        data = ('{"request":{"data":{"names":["a","b"],"ndarray":[[1.0,2.0]]}},'
+                '"response":{"meta":{"routing":{"router":0}},"data":{"names":["a","b"],'
+                '"ndarray":[[1.0,2.0]]}},"reward":1}')
+        response = requests.get(
+            "http://127.0.0.1:5000/send-feedback", params="json=%s" % data)
+        response.raise_for_status()
+        assert response.json() == {'data': {'ndarray': []}, 'meta': {}}
+
+
 @pytest.mark.parametrize(
         'tracing', [(False), (True)]
     )