/contrib/famzah

Enthusiasm never stops

Google Cloud API and Python

Leave a comment

Confusion is what I got the first time I wanted to automate Google Cloud using Python. While the documentation of Google Cloud is not bad, it’s far from ideal. Let me try to clarify things if you’re struggling with this like I did.

In its very core, Google Cloud API is an HTTP REST service (or gRPC for selected APIs). And you have two ways of accessing the Google Cloud API using Python:

While the “Client Libraries” is Google’s recommended option, my own experience shows that you’d better use the second option which accesses the HTTP REST API directly. That is because:

  • The Google Cloud HTTP REST API is documented better. I guess the reason for that is because it’s used by all the languages, not only Python, and because Google developers use it internally, too.
  • The official documentation of the Google Cloud services always gives an example for the REST APIs along with the other options like the web console, gcloud/gsutil, and the “Client Libraries” (where they are applicable).
  • The Google Cloud web console interface gives the resulting HTTP REST API or command line code of what you populated in the forms online. You can see this at the very bottom of the page. Therefore, you can use the web interface to enter what you indent to do and then easily see what you need to send to the HTTP REST API, in order to achieve the same as you’d do if you clicked it online.
  • Not all Google Cloud APIs are wrapped in “Client Libraries” for a specific language like Python. For example, Compute Engine is not, so you would most probably need to use the HTTP REST API directly anyway. If that’s the case, why learn two libraries when you can learn only one and use only the HTTP REST API.

The installation of the Python library for Google Cloud direct HTTP REST API access is easy. There is documentation but the process is as straightforward as “pip install google-api-python-client”. You can then access any Google Cloud HTTP API and here is an oversimplified example:

import googleapiclient.discovery
compute = googleapiclient.discovery.build('compute', 'v1')
result = compute.regions().list(project='my_project_id', maxResults=2).execute()

Before you can execute the example, you have to set up your authentication. I personally find Authenticating as a service account very convenient but you can also review the other Google Cloud API authentication options.

Please note that even the simple Python library for Google Cloud direct HTTP REST API access has its minor peculiarities and is not a 1:1 mapping of the API. For example, when you call an update() API the HTTP REST API documentation names the selector field “resourceId” while the Python library’s update() method names this field “healthCheck”, for example for the healthChecks().update() method. Therefore, you always need to consult both documentations when you develop your scripts.

Here is what I keep in my bookmarks when working with the Google Cloud API using Python:

  • Google APIs Explorer — the core documentation of every service.
  • Python library for Google Cloud direct HTTP REST API access — I consult it for the named arguments of the methods if they differ from the official API (see the example in the previous paragraph). Additionally, if you use the list() methods, the results are usually paginated and you have to call list_next() which is well documented.

Author: Ivan Zahariev

An experienced Linux & IT enthusiast, Engineer by heart, Systems architect & developer.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s