[cmd] Add deadband trigger methods to CommandGenericHID

[narmstro2020]

This should make creating triggers from axis values being non-zero a little more ergonomic without having to write an inline supplier. First drafted with Java. C++ coming soon

[github-actions]

This PR modifies commands. Please open a corresponding PR in Python Commands and include a link to this PR.

[narmstro2020]

/format

[narmstro2020]

/format

[KangarooKoala]

Could you give an example of a specific control binding that would use this? (e.g., when trigger is pressed more than a certain threshold, shoot) axisGreaterThan() and axisLessThan() already exist, and I'm having trouble thinking of a reasonably common use case that can't be solved by those.

Reading the initial comment for this PR again, I see that the provided example was "axis values being non-negative". Is there a reason why axisGreaterThan(deadband) wouldn't work?

[narmstro2020]

Reading the initial comment for this PR again, I see that the provided example was "axis values being non-negative". Is there a reason why axisGreaterThan(deadband) wouldn't work?

Unless I'm misreading axisGreatherThan would only check if the value is above a threshold, not non-negative. (EDIT: non-zero) So for example. Let's say I want to Trigger manual control when a joystick enters a non-negative (EDIT: non-zero) value (after the deadband).

You would have to do. new Trigger(() -> MathUtil.applydeadband(controller.getLeftY(), deadband) != 0) and then check an onTrue, whileTrue, etcetera.

EDIT: The original post had a typo. non-negative should've read non-zero. Fixed in the top level PR comment

[KangarooKoala]

Could you describe to me what you mean by non-negative? To me, the meaning of non-negative means >=0, which means axisGreaterThan(0) (or axisGreaterThan(deadband)) will produce a trigger that's true when the axis is non-negative.

[narmstro2020]

Could you describe to me what you mean by non-negative? To me, the meaning of non-negative means >=0, which means axisGreaterThan(0) (or axisGreaterThan(deadband)) will produce a trigger that's true when the axis is non-negative.

I meant to say non-zero. Not non-negative. Fixing the initial comment. Sorry about that.

[KangarooKoala]

No worries! The question still remains, though- What's a concrete example of a binding to an axis being non-zero?

[narmstro2020]

No worries! The question still remains, though- What's a concrete example of a binding to an axis being non-zero?

So activate manual control of a device (drive, elevator, arm, shooter, etc) via

new Trigger(() -> MathUtil.applydeadband(controller.getLeftY(), deadband) != 0).onTrue(device.drive(MathUtil.applydeadband(controller.getLeftY(), deadband);

Excuse the crudity of this model. There are more elegant ways to pass the deadbanded joystick into the device.drive method.

device.drive would be a Command factory of some sort.

This would reduce it down to

controller.axisActive(0, 0.1).onTrue(... command binding)

[KangarooKoala]

That example makes sense! Give me a moment to write up a review with some things I noticed.

[narmstro2020]

That example makes sense! Give me a moment to write up a review with some things I noticed.

No problem. That'll get me a chance to write the C++.

[narmstro2020]

/format

[narmstro2020]

Is this tools build related to my PR

[KangarooKoala]

Based on the errors ("Could not resolve ..."), it seems to be a CI network issue. (Which is unrelated to the PR)

[narmstro2020]

/format

[narmstro2020]

Do I need to do the python item mentioned in Github-actions above?

[WispySparks]

Do I need to do the python item mentioned in Github-actions above?

I would say yes, https://github.com/robotpy/robotpy-commands-v2/blob/main/commands2/button/commandgenerichid.py

[narmstro2020]

anything left to do on this one?

[KangarooKoala]

Looks good to me, for what it's worth.