Delete axes interactions not working?

I'm using 2021a and it seems that deleting/disabling interactions isn't working as expected. I'm just trying to either (1) make an axes with just the pan interaction, or (2) make an axes with all interactions removed

fH = figure
ax = gca;
ax.Interactions = panInteraction % I only want the pan interaction but this line seems to do nothing

Similarly removing all via the following is ineffective:

ax.Interactions = [] % This line also seems to do nothing

I'm trying to make a safe figure that the user can't break so I really want to limit what they can do. The only thing I seem to be able to do is make the toolbar invisible via:

ax.Toolbar.Visible = false

Is this a bug? I seem to remember having the ability to manipulate ax.Interactions in earlier MATLAB versions ...

Thanks,

Sven

Answer 1

[Adam Danz]

Difference between toolbar and interactions

Interactions are the actions available by using the mouse to zoom/pan/rotate etc the axes without activing a tool in the axis toolbar. For example, after generating a figure that supports zooming you can use the mouse scroll wheel to immediately zoom into the axes.

The toolbar property is independed of the interactions property. For example, you can remove panning from the toolbar but still allow for interactive panning by mouse actions.

I agree that these differences could be clearer in the documentation.

Specify toolbar tools

To specify toolbar tools, use the axtoolbar function. This example reduces the toolbar to the pan tool.

ax = gca(); 
tb = axtoolbar(ax,'pan')

See the toolbar link I provided above for all toolbar options.

Specify axis interactions

To disable all tools except panning, set the Interactions property of the axes. This example reduces interactions to panning.

ax = gca;
ax.Interactions = panInteraction;

See the interactions link I provided above for all interaction options.

[Sven]

Thanks Adam - by the way I'm following your other thread on axes vs uiaxes (or figure vs uifigure) speed. My question here was related since I'm using a bunch of images.roi.[x] objects and have found the uifigure/uiaxes performance to be extremely sluggish, so my additional motivation for removing interactions and/or the toolbar was to potentially salvage some display speed.

Overall I'd really prefer to use uifigures for future compatibility/standardization and some of the default layout tools, but even some simple interfaces are just too unresponsive to inflict on a user. A quick example is below:

axH = uiaxes(uifigure)
for i = 1:10
    images.roi.Line(axH,'Position',rand(2,2));
end

axH = axes(figure)
for i = 1:10
    images.roi.Line(axH,'Position',rand(2,2));
end

On my machine (i7 laptop, plenty of ram) you can wave the mouse over the uifigure and hardly activate any of the MouseOver type of events that change the line color or cursor style unless you literally let the cursor rest over a line object for ~1sec for things to catch up. The performance over the java-based figure is fine though. Oh well... sticking to figure() for now...

[Adam Danz]

What mouseover events are you referencing? Ones that you've added? The only mouseOver event I see is the pointer change to a fleur rather than an arrow when covering over one of the line objects. There's a bit of a lag for the cursor change in uifigures/uiaxes but I'm talking 100 ms, ballpark.

[Sven]

Yep, that's the lag I'm talking about (line being hovered over gets highlighted and cursor changes to fleur) and the general performance if you, say, grab a line and wiggle the mouse around. Hah, ~1sec is probably an overestimation so I should have avoided the term "literally" :)

I'd say it's 250ms on my machine, particularly for the cursor change. But the lag when simply dragging those totally vanilla lines around is very, very noticable - enough so that I don't even want to consider what it would be like when conditions are non-ideal (like adding "MovingROI" listeners with non-trivial calculation time in their callbacks, or like when running on remote machines which is becoming more common these days, or when precise cursor movements are desirable).

[Adam Danz]

I see what you mean now, thanks for the usecase description. I wish I could suggest something but I have a feeling the cause of the difference in processing speed is quit deeply rooted.