Module radosgw_user doesn't allow to use cephadm shell

I have hosts where `ceph-common` package wont be installed anymore, which used to provide `radosgw-admin` command. When using `ceph.automation.radosgw_user` module, I observe the following error:

```sh
TASK [Ensure admin rgw user is present] ******************************************************************************************************
[ERROR]: Task failed: Module failed: Error executing command: [Errno 2] No such file or directory: b'radosgw-admin'
...
fatal: [...]: FAILED! => {"changed": false, "cmd": "radosgw-admin --cluster ceph user info --uid=... --format=json", "msg": "Error executing command.", "rc": 2, "stderr": "", "stderr_lines": [], "stdout": "", "stdout_lines": []}
```

Seeing [radosgw_user doc](https://galaxy.ansible.com/ui/repo/published/ceph/automation/content/module/radosgw_user/), there is no option to use `cephadm shell` instead of relying of commands installed at host level. It is an option available for some other modules.

`radosgw-admin` is available using `cephadm shell -- radosgw-admin ...`.

Could you please allow this feature ?

Start by locating the ceph.automation.radosgw_user module entry point and compare its command handling with modules that already support cephadm shell. Reproduce the documented radosgw-admin failure, then verify that the module can run the command through cephadm shell while preserving its existing user-management behavior.

Written by the indexing model from the issue text.