Named axes

[psobolewskiPhD]

From OMERO we should know the OME axes TCZYX so it would be nice to pass the names to napari.

• Related to: Consider squeezing out singleton dimensions  #62

[jo-mueller]

You mean the channel names? Aren't these already passed on to napari? If I load an image from OMERO with axes names set, I receive as many layers in the napari browser as there are channels with the layer names corresponding to the channel names in napari.

[psobolewskiPhD]

No no, I meant the names of the coordinate axes. By default they are 0, 1, 2, 3, ... in napari because it's totally dimensionally agnostic/ambivalent. But with the data coming from OMERO we can know that we have for example a T-stack. So, we could tell napari the axes are T, Y, X. For a Z-stack, we could have the axes as Z, Y, X and so on for higher dim data. This would b especially handy if squeezing out empty dims (#62), because otherwise you'd have know way of knowing of you original had a T-stack or Z-stack.

[will-moore]

I think I was asking how to do that a while back in napari, either napari-omero or for napari-ome-zarr? But I forget why that didn't get done? Maybe it's easier now?

[psobolewskiPhD]

@will-moore
Yes and no? 😬
In napari/napari#6979 layers got Layer.axis_labels, which is a tuple of str.
But the viewer.dims does not (yet?) automatically use this information.
But we could do viewer.dims.axis_labels = viewer.layers[-1].axis_labels or something similar after adding the layer?
cc: @Czaki

Edit:
Of course there can be issues if layers of different dims are loaded simultaneously.
I think we should consider reverting the ability to have multiple OMERO images open at the same time? (#76)