Allow passing extra arguments to test runs

[pepisg]

Hi! Thanks a lot for open sourcing this package!

I was setting up some tests involving nav2 costmaps and wanted to expose an argument to chose wether to launch rviz on nav2's launch. This is useful in my case because I want to debug the tests while i create them, but don't want to use rviz when running on my CI system.

I modified the cli to support this.

Sorry in advance if there's already a way of doing this that I missed.

[troygibb]

Hey @pepisg thanks for the contribution!! Can you provide an example of how you would set this up within a replay test? Off the bat, I'd like to nest these args somewhere specific to the test. But an example would help clarify the correct usage pattern

[pepisg]

It would look something like this:

class Run:
    def generate_launch_description(
        self, replay_run_params: ReplayRunParams
    ) -> LaunchDescription:

        launch_rviz = "True" if _args.launch_rviz else "False"
        launch_args = {
            "launch_rviz": launch_rviz,
        }
        return LaunchDescription(
            [
                IncludeLaunchDescription(
                    PythonLaunchDescriptionSource(str(nav2_launch_file)),
                    launch_arguments=launch_args.items(),
                ),
            ]
        )

which then I'd launch as:

ros2 run replay_testing replay_test my_test.py --launch-rviz

[troygibb]

@pepisg sorry for the late reply here!

Yea I like this. I can't decide if it should be nested under something like --launch-args --launch-rviz or should be totally generic. I could see the same args being applied to the fixture/analyze step as well. Any thoughts?