Added README and sample code for feature group.

Author: Ubuntu

samples/feature_group/README.md

@@ -0,0 +1,103 @@
+# Feature Group Sample
+
+This sample demonstrates how to create a feature group using the Amazon AWS Controllers for Kubernetes (ACK) service controller for Amazon SageMaker.
+
+## Prerequisites
+
+This sample assumes that you have completed the [common prerequisites](https://github.com/aws-controllers-k8s/sagemaker-controller/blob/main/samples/README.md).
+
+### Create an S3 bucket:
+
+Since we are using the offline store in this example, you need to set up an s3 bucket. [Here are directions](https://docs.aws.amazon.com/AmazonS3/latest/userguide/create-bucket-overview.html) to set up your s3 bucket through the S3 Console, AWS SDK, or AWS CLI.
+
+### Updating the Feature Group Specification:
+
+In the `my-feature-group.yaml` file, modify the placeholder values with those associated with your account and feature group.
+
+## Creating your Feature Group
+
+### Create a Feature Group:
+
+To submit your prepared feature group specification, apply the specification to your Kubernetes cluster as such:
+
+```
+$ kubectl apply -f my-feature-group.yaml
+featuregroup.sagemaker.services.k8s.aws/my-feature-group created
+```
+
+### List Feature Groups:
+
+To list all feature groups created using the ACK controller use the following command:
+
+```
+$ kubectl get featuregroup
+```
+
+### Describe a Feature Group:
+
+To get more details about the feature group once it's submitted, like checking the status, errors or parameters of the feature group use the following command:
+
+```
+$ kubectl describe featuregroup my-feature-group
+```
+
+## Ingesting Data into your Feature Group
+
+Note that ingestion is **not** supported in the controller.
+The following assumes creation of a feature group with its name stored in `feature_group_name`
+
+```
+### Sample CSV data file for Ingestion Example:
+#TransactionID,EventTime
+#1,1623434915
+#2,1623435267
+#3,1623435284
+
+###Example boto3 ingestion of feature group:
+
+import boto3
+import csv
+
+sagemaker_featurestore_runtime_client = boto3.Session().client(
+  service_name="sagemaker-featurestore-runtime")
+
+### OPTION 1: To Download all records at once and upload records sequentially
+with open('./Downloads/Sample_data.csv') as file_handle:
+  records =[
+             [
+               {'FeatureName':featureName,
+                'ValueAsString':valueAsString} 
+             for featureName, valueAsString in row.items()]
+           for row in csv.DictReader(file_handle, skipinitialspace=True)]
+        
+for record in records:
+  sagemaker_featurestore_runtime_client.put_record(
+    FeatureGroupName=feature_group_name,
+    Record=record)
+
+### OPTION 2: To Download records sequentially and upload records sequentially
+with open('./Downloads/Sample_data.csv') as file_handle:
+  for row in csv.DictReader(file_handle, skipinitialspace=True):
+    record =[
+              {'FeatureName':featureName,
+               'ValueAsString':valueAsString} 
+            for featureName, valueAsString in row.items()]
+    sagemaker_featurestore_runtime_client.put_record(
+      FeatureGroupName=feature_group_name,
+      Record=record)       
+       
+# To Check that the records are retrievable       
+for recordIdentifierValue in range(1,len(records) + 1):
+  sagemaker_featurestore_runtime_client.get_record(
+    FeatureGroupName=feature_group_name,
+    RecordIdentifierValueAsString=str(recordIdentifierValue))
+```
+
+## Deleting your Feature Group
+
+To delete the feature group, use the following command:
+
+```
+$ kubectl delete featuregroup my-feature-group
+featuregroup.sagemaker.services.k8s.aws "my-feature-group" deleted
+```

samples/feature_group/my-feature-group.yaml

@@ -0,0 +1,19 @@
+apiVersion: sagemaker.services.k8s.aws/v1alpha1
+kind: FeatureGroup
+metadata:
+  name: <YOUR FEATURE GROUP NAME>
+  spec:
+    eventTimeFeatureName: EventTime
+    featureDefinitions:
+      - featureName: TransactionID
+        featureType: Integral
+      - featureName: EventTime
+        featureType: Fractional
+    featureGroupName: <YOUR FEATURE GROUP NAME>
+    recordIdentifierFeatureName: TransactionID
+    offlineStoreConfig:
+      s3StorageConfig:
+        s3URI: s3://<YOUR BUCKET>/feature-group-data
+    onlineStoreConfig:
+enableOnlineStore: True
+    roleARN: <YOUR SAGEMAKER ROLE ARN>