Skip to content

Latest commit

 

History

History
52 lines (35 loc) · 4.04 KB

helper-class.md

File metadata and controls

52 lines (35 loc) · 4.04 KB
Raw

Python helper class: CloudSqlAdmin

Using the client libraries is definitely simpler than directly calling the REST APIs, but as we saw in the client library examples, there are still some complexities you need to manage. These range from creating a request body to dealing with pagination, authentication, and other details.

To minimize context switching and distractions, you may find it useful to put together helper functions or classes that abstract away all of those details, so that you can stay hyper-focused on your business logic. I've created a small proof of concept of such a helper class, covered below.

The CloudSqlAdmin helper class

The gcsql_admin.py file in this repo contains a CloudSqlAdmin class, which is a container for entity-specific classes that wrap the functionality of the top-level instance methods of the Cloud SQK Admin API. Here's a high-level architectural diagram:

CloudSqlAdmin architecture

I've only implemented three of the entity-specific classes so far (Instances, Databases, and Users), including methods that simplify the client library examples we looked at earlier.

The basic concept is that you instantiate CloudSqlAdmin and then call an entity-specific method. For example, here's how you'd create a new user in a Cloud SQL instance:

sql_admin = CloudSqlAdmin()
if sql_admin.users.insert(project, instance, host, username, password):
    print(f"User created: {username}")
else:
    print(f"ERROR inserting user: {sql_admin.response}")

There's less code than in the client library example, because the wrapper class has handled creating the request body and other details.

Here's another example, showing how to get all of the instances in a project without any need to deal with the details of pagination:

sql_admin = CloudSqlAdmin()
for instance in sql_admin.instances.list(project):
    print(f"Instance: {instance['name']}")

You can find examples of working with CloudSqlAdmin in the samples.py file in this repo. For convenience, those sample use settings from config.py for MY_PROJECT and MY_INSTANCE, so you'll need to set those to your own project and instance names to run the samples.

If you'd like to call Cloud SQL Admin APIs that haven't been implemented in the entity-specific classes in CloudSqlAdmin, you can use the CloudSqlAdmin.service member. The tier_list function in samples.py provides an example of how to do this:

sql_admin = CloudSqlAdmin()
request = sql_admin.service.tiers().list(project=project)
response: dict = request.execute()
print(response)

CloudSqlAdmin is just one of many approaches you could take for simplifying use of the Cloud SQL admin APIs. For example, if you prefer a functional approach you might want to simply create a service resource object (as done in the service_client() function in gcsql_admin.py) and then create helper functions that take this object as an argument in the spirit of dependency injection. Do whatever feels right for you.

A few final things to note about CloudSqlAdmin:

  • Since this is intended as a simple educational sample, it doesn't include error handling and other details you'd want to add for production use.
  • I've included Python 3 type annotations to clarify some things. These are ignored by the Python interpreter, and you can ignore them as well if you're not accustomed to Python's type annotations.
  • For authentication, there's an APP_CREDENTIALS setting in config.py that stores the name of a local app registration JSON file, and those credentials will be used if the environment variable isn't set. This essentially automates setting the environment variable, so that you don't need to do that each place where you want to run your code.