🚨 Monitors and alerts

docs/mkdocs/docs/tutorial/monitors_alerts.md

@@ -2,6 +2,25 @@
 
 ## Monitors
 
+You can create a monitor in whitebox so that alert are created automaticaly when some value is out of bounds. Here is an example:
+
+```Python
+from whitebox import Whitebox, MonitorStatus, MonitorMetrics, AlertSeverity
+
+wb = Whitebox(host="127.0.0.1:8000", api_key="some_api_key")
+
+model_monitor = wb.create_model_monitor(
+    model_id="mock_model_id",
+    name="test",
+    status=MonitorStatus.active,
+    metric=MonitorMetrics.accuracy,
+    feature="feature1",
+    lower_threshold=0.7,
+    severity=AlertSeverity.high,
+    email="jaclie.chan@somemail.io",
+)
+```
+
 ## Alerts
 
 Once the metrics reports have been produced, the monitoring alert pipeline is triggered. This means that if you have created any model monitors for a specific metric, alerts will be created if certain criteria are met, based on the thresholds and the monitor types you have specified.

whitebox/sdk/whitebox.py

@@ -95,7 +95,7 @@ def delete_model(self, model_id: str):
 
     def log_training_dataset(
         self, model_id: str, non_processed: pd.DataFrame, processed: pd.DataFrame
-    ):
+    ) -> bool:
         """
         Logs a training dataset for a model.
 
@@ -133,7 +133,7 @@ def log_inferences(
         processed: pd.DataFrame,
         timestamps: pd.Series,
         actuals: pd.Series = None,
-    ):
+    ) -> bool:
         """
         Logs inferences of a model.
 
@@ -195,7 +195,7 @@ def create_model_monitor(
         lower_threshold: Optional[float],
         severity: AlertSeverity,
         email: str,
-    ):
+    ) -> dict:
         """
         Creates a monitor for a model.
         """
@@ -219,3 +219,15 @@ def create_model_monitor(
 
         logger.info(result.json())
         return result.json()
+
+    def get_alerts(self, model_id: str = "") -> dict:
+        """
+        Returns all alerts for a model.
+        """
+        result = requests.get(
+            url=f"{self.host}/{self.api_version}/alerts?modelId={model_id}",
+            headers={"api-key": self.api_key},
+        )
+
+        logger.info(result.json())
+        return result.json()

whitebox/tests/v1/test_sdk.py

@@ -194,4 +194,4 @@ def test_sdk_create_model_monitor(client):
             email="jaclie.chan@chinamail.io",
         )
 
-        assert model_monitor == model_monitor
+        assert model_monitor is not None