manager.create_droplet() instead of digitalocean.Droplet(client_id, api_key, ...) about python-digitalocean HOT 2 CLOSED

It is like this because you should use the "Droplet" class in order to create a Droplet. For me it is more logical to create a Droplet class instead of doing that inside the Manager... At the beginning I thought that the "manager" should "handle" the process of creation too, but this means that all the other classes makes no sense to exists and I should have only one class that does everything. This is against of the "rule" of writing "better" code that is not repeating and makes more sense. (and is a little diluted )

So at the end of the day you can think to put a new "create_droplet" method inside the Manager class, but if I have to follow the "guidelines" that makes the code more elegant and cleaner, it is a strange thing to me.

from python-digitalocean.

I understand your argument, but I still don't like the duplication of having to use the keys here.
What about:

    manager = digitalocean.Manager(client_id="...", api_key="...")
    droplet = digitalocean.Droplet(name='Example',
                               region_id=1,  # New York
                               image_id=2676,  # Ubuntu 12.04 x64 Server
                               size_id=66,  # 512MB
                               backup_active=False)
    manager.create_droplet(droplet)

You still have your object creation pattern, and only one place to keep the keys. It is your project, so its totally up to you. The above pattern would just suit my preferences better

from python-digitalocean.