add monitor subcommands that work from file

[Hefeweizen]

These subcommands read a filename from the command line, and then use
specified file for all values. These features allow a workflow of dog monitor show <id> > monitor_id.json && dog monitor fupdate monitor_id.json. This workflow allows version control of all monitors,
as well as repo-driven updates.

[yannmh]

This is very helpful. Thank you @Hefeweizen

[Hefeweizen]

@yannmh Thanks; I appreciate the review.

I did mvp for my needs and only got this working with monitors. I think it'd be nice if this capability (load from file) were available for all object types (e.g. dashboards, &c). Should I piecemeal this in on each subcommand, or is there something architectural that could be done?

[yannmh]

I agree this would also be great for other resources.
Adding a fpost and fupdate to all of them doesn't feel right ... instead I'd like to explore the possibility of re-using the existing "post" and "update" subcommands and maybe some extra parameter to specific that the we intend to pass content as json, e.g. something like

cat my_monitor.json | dog monitor update --json

[Hefeweizen]

@yannmh I chose fupdate/fpost because I just wanted traction on the idea and to get something started without having to build too much argument handling into the existing _post. For better integration, it seems something like this may work:

    def _post(cls, args):
       if args.json:
          _post_stdin()
       else:
         _post_from_args()

    def _post_from_args(cls, args):
        """
        previously, just `_post()`
        """
        api._timeout = args.timeout
        format = args.format
        options = None
        ....

    def _post_stdin(cls, args):
       monitor = json.load(sys.stdin)
       ....

This provides the capability locally, but we'd have to make the changes across every subcommand (monitors, dashboards, &c). Is there anything we can do structurally to provide the capability across all of them in one swoop?