In my Pulumi Python code, I have the following

input_type

classes:

@pulumi.input_type
class SftpHomeDirectoryArgs:
    def __init__(
        self,
        *,
        bucket_name: str,
        bucket_path: str,
        entrypoint: str = "/",
        bucket_kms: Optional[pulumi.Input[str]] = None,
    ):
        ...

@pulumi.input_type
class SftpUserArgs:
    def __init__(
        self,
        *,
        username: str,
        home_directory: SftpHomeDirectoryArgs,
        ssh_key: Optional[str] = None,
        password_login: bool = False,
    ):
        ...

Since

bucket_kms

in

SftpHomeDirectoryArgs

is defined as

pulumi.Input[str]

, should

home_directory

in

SftpUserArgs

also be declared as

pulumi.Input[SftpHomeDirectoryArgs]

? Or is it fine as it is? Thanks!

I think either should work

Both approaches work, but is there a best practice or recommendation for cases like this? Specifically, in my example, is there any difference between declaring:

home_directory: SftpHomeDirectoryArgs

vs

home_directory: pulumi.Input[SftpHomeDirectoryArgs]

Would explicitly wrapping it in

pulumi.Input

provide any benefits, such as better handling of dependencies or asynchronous values?

it would mean that the other parts of home_directory (i.e. name, path, entrypoint) could be defined with input values, but we don't generally nest input values We generally don't nest input values, so if the leaf field has it you probably don't need the parent field to have it as well

Thanks!