Data Api plus Aurora Serverless v2

python/cross_service/aurora_item_tracker/README.md

@@ -44,9 +44,15 @@ python -m pip install -r requirements.txt
 
 ### Aurora Serverless DB cluster and Secrets Manager secret
 
-This example requires an Aurora Serverless DB cluster that contains a MySQL database. The
-database must be configured to use credentials that are contained in a Secrets Manager
-secret. 
+This example requires an Aurora DB cluster with the Data API feature enabled. As of March 2024,
+the available choices are:
+* Aurora PostgreSQL using Aurora Serverless v2 or provisioned instances in the cluster.
+* Aurora MySQL or Aurora PostgreSQL using a Serverless v1 cluster.
+
+For this example, we assume that the Aurora cluster uses a combination of Aurora PostgreSQL
+and Aurora Serverless v2.
+
+The database must be configured to use credentials that are contained in a Secrets Manager secret. 
 
 Follow the instructions in the 
 [README for the Aurora Serverless application](/resources/cdk/aurora_serverless_app/README.md) 
@@ -71,16 +77,16 @@ CloudFormation setup script:
 credentials, such as `arn:aws:secretsmanager:us-west-2:123456789012:secret:docexampleauroraappsecret8B-xI1R8EXAMPLE-hfDaaj`.
 * **DATABASE** — Replace with the name of the database, such as `auroraappdb`.  
 
-*Tip:* The caret `^` is the line continuation character for a Windows command prompt.
-If you run this command on another platform, replace the caret with the line continuation
+*Tip:* The caret `\` is the line continuation character for a Linux or Mac command prompt.
+If you run this command on another platform, replace the backslash with the line continuation
 character for that platform.
 
 ```
-aws rds-data execute-statement ^
-    --resource-arn "CLUSTER_ARN" ^
-    --database "DATABASE" ^
-    --secret-arn "SECRET_ARN" ^
-    --sql "create table work_items (iditem INT AUTO_INCREMENT PRIMARY KEY, description TEXT, guide VARCHAR(45), status TEXT, username VARCHAR(45), archived BOOL DEFAULT 0);"
+aws rds-data execute-statement \
+    --resource-arn "CLUSTER_ARN" \
+    --database "DATABASE" \
+    --secret-arn "SECRET_ARN" \
+    --sql "create table work_items (iditem SERIAL PRIMARY KEY, description TEXT, guide VARCHAR(45), status TEXT, username VARCHAR(45), archived BOOL DEFAULT false);"
 ```
 
 #### AWS Management Console
@@ -102,6 +108,19 @@ as `auroraappdb`.
 This opens a SQL query console. You can run any SQL queries here that you want. Run the 
 following to create the `work_items` table:
 
+For a PostgreSQL-compatible database:
+```sql
+create table work_items ( 
+  iditem SERIAL PRIMARY KEY, 
+  description TEXT, 
+  guide VARCHAR(45), 
+  status TEXT, 
+  username VARCHAR(45), 
+  archived BOOL DEFAULT false
+);
+
+```
+For a MySQL-compatible database:
 ```sql
 create table work_items ( 
   iditem INT AUTO_INCREMENT PRIMARY KEY, 

python/cross_service/aurora_item_tracker/storage.py

@@ -112,14 +112,26 @@ def add_work_item(self, work_item):
         """
         Adds a work item to the database.
 
+        Note: Wrapping the INSERT statement in a WITH clause and querying the auto-generated
+              ID value is required by a change in the Data API for Aurora Serverless v2.
+              The "generatedFields" fields in the return value for DML statements are all blank.
+              That's why the RETURNING clause is needed to specify the columns to return, and
+              the entire statement is structured as a query so that the returned value can
+              be retrieved from the "records" result set.
+
+              This limitation might not be permanent; the DML statement might be simplified
+              in future.
+
         :param work_item: The work item to add to the database. Because the ID
                           and archive fields are auto-generated,
                           you don't need to specify them when creating a new item.
         :return: The generated ID of the new work item.
         """
         sql = (
+            f"WITH t1 AS ( "
             f"INSERT INTO {self._table_name} (description, guide, status, username) "
-            f" VALUES (:description, :guide, :status, :username)"
+            f" VALUES (:description, :guide, :status, :username) RETURNING iditem "
+            f") SELECT iditem FROM t1"
         )
         sql_params = [
             {"name": "description", "value": {"stringValue": work_item["description"]}},
@@ -128,7 +140,10 @@ def add_work_item(self, work_item):
             {"name": "username", "value": {"stringValue": work_item["username"]}},
         ]
         results = self._run_statement(sql, sql_params=sql_params)
-        work_item_id = results["generatedFields"][0]["longValue"]
+        # Old style, for Serverless v1:
+        #        work_item_id = results["generatedFields"][0]["longValue"]
+        # New style, for Serverless v2:
+        work_item_id = results["records"][0][0]["longValue"]
         return work_item_id
 
     def archive_work_item(self, iditem):

python/cross_service/aurora_rest_lending_library/README.md

@@ -7,7 +7,7 @@ Service (Amazon RDS) API and AWS Chalice to create a REST API backed by an
 Amazon Aurora database. The web service is fully serverless and represents
 a simple lending library where patrons can borrow and return books. Learn how to:
 
-* Create and manage a serverless Amazon Aurora database cluster.
+* Create and manage a serverless Amazon Aurora database. This example uses Aurora Serverless v2.
 * Use AWS Secrets Manager to manage database credentials.
 * Implement a data storage layer that uses Amazon RDS Data Service to move data into
 and out of the database.  
@@ -47,29 +47,15 @@ all must be run separately.
 
 ---
 
-### 1. Database deployment
-
-This database setup includes:
-* an Amazon Aurora serverless data cluster
-* an AWS Secrets Manager secret to hold the database user credentials.
-
-This infrastructure is defined in a [setup.ts](resources/cdk/aurora_serverless_app/setup.ts) (see [README.md](resources/cdk/aurora_serverless_app/README.md)).
-
-To execute this CDK stack, run:
-```
-cdk deploy
-```
-The output will show `ClusterArn`, `DbName`, and `SecretArn`. Make sure these values are reflected in the [config for this project](config.yml).
-
----
-
-### 2. Populate database
+    Creates an Amazon Aurora cluster and associated Aurora Serverless v2 database instance.
+    Also creates an AWS Secrets Manager secret to hold the database user credentials.
+    It fills the database with example data pulled from the [Internet Archive's Open Library](https://openlibrary.org). 
 
 Fill the database with example data pulled from the [Internet Archive's Open Library](https://openlibrary.org). 
 
 ```
-python library_demo.py populate_data
-```
+python library_demo.py deploy_database
+``` 
 
 The database is now ready and can be accessed through the 
 [AWS Console Query Editor](https://console.aws.amazon.com/rds/home?#query-editor:) 

python/cross_service/aurora_rest_lending_library/library_api/app.py

@@ -50,7 +50,7 @@ def index():
     """Briefly describes the REST API."""
     return {
         "description": "A simple lending library REST API that runs entirely on "
-        "serverless components."
+        "serverless components: Aurora Serverless v2, API Gateway"
     }
 
 
@@ -131,8 +131,15 @@ def add_patron():
 
     :return: The ID of the added patron.
     """
-    patron_id = get_storage().add_patron(app.current_request.json_body)
-    return {"Patrons.PatronID": patron_id}
+    try:
+        patron_id = get_storage().add_patron(app.current_request.json_body)
+        return {"Patrons.PatronID": patron_id}
+    except Exception as err:
+        logger.exception(
+            f"Got exception in add_patron() inside library_api/app.py: {str(err)}"
+        )
+        logger.exception(f"Returning None instead of patron_id.")
+        return None
 
 
 @app.route("/patrons/{patron_id}", methods=["DELETE"])
@@ -155,7 +162,15 @@ def list_borrowed_books():
 
     :return: The list of currently borrowed books.
     """
-    return {"books": get_storage().get_borrowed_books()}
+    try:
+        json_doc = {"books": get_storage().get_borrowed_books()}
+        return json_doc
+    except Exception as err:
+        logger.exception(
+            f"Exception while calling get_storage().get_borrowed_books(): {str(err)}"
+        )
+        logger.exception(f"Continuing with blank list of borrowed books...")
+        return {"books": []}
 
 
 @app.route("/lending/{book_id}/{patron_id}", methods=["PUT", "DELETE"])