[WIP] Allow atom selection via VR controllers

[Nargaruga]

This allows the user to select atoms by pointing the laser and pressing the button mapped to Action1 (typically the trigger), as requested in #378.

Some problems are still present:

• there is a considerable offset between the laser pointer and the atom that is actually selected. Reducing cRange in layer1/ScenePicking.cpp does not seem to improve the situation by much.

Possible improvements:

• add the ability to select whole residues;
• change the laser color when a residue is being pointed at (similar to how the laser becomes blue when aiming at the menu).

Additional notes:

• this requires openvr_cut_laser to be set to true in layer1/SettingInfo.h;
• I was not able to compile PyMOL with OpenVR until I added the #include <thread> to contrib/vr/OpenVRControllerModel.cpp (change not included in the commit). Not sure if it's a configuration issue on my end.

[Nargaruga]

there is a considerable offset between the laser pointer and the atom that is actually selected. Reducing cRange in layer1/ScenePicking.cpp does not seem to improve the situation by much.

Applying a vertical offset to the picking projection parameters seems to solve the issue (see commit 43ab48e). The change does not seem to influence other VR interactions (i.e. with the menu or the sequence viewer).

Before the change (the purple color of the laser indicates that an atom is being pointed at, which shouldn't be the case here)

After the change

This is a band-aid solution, I assume the correct approach would be to update the following parameters:

pymol-open-source/contrib/vr/OpenVRStub.cpp

Lines 137 to 150 in 9d3061c

	void VRSystemStub::GetProjectionRaw(EVREye eEye, float *pfLeft, float *pfRight, float *pfTop, float *pfBottom)
	{
	if (eEye == Eye_Left) {
	*pfLeft = -1.39666343f;
	*pfRight = 1.23994398f;
	*pfTop = -1.47110546f;
	*pfBottom = 1.45868433f;
	} else {
	*pfLeft = -1.24060011f;
	*pfRight = 1.39432383f;
	*pfTop = -1.46788049f;
	*pfBottom = 1.46032500f;
	}
	}

But I do not know how they were first extrapolated, nor if they are tied to a specific VR headset (I'm testing the changes on a Meta Quest 3).

[JarrettSJohnson]

Thanks for submitting this. I see that this is still WIP. Would you still want to work on this, or have the changes committed? We could have some value that we pass in at compile time to tweak this offset. I haven't looked too much into it, but I feel like projection values from VR is probably meant to come from a calibrated headset.

[Nargaruga]

Hi, I haven't looked into this anymore after my last comment, but I would still be available to work on it.
Perhaps the projection values (or the offset) could be made available as parameters, so that the user might edit them from the advanced settings window?