Clarify that `pyramid_gaussian` returns a generator, not a sequence of images

[alexdesiqueira]

Description

See @grlee77's comment on #6293:

One subtle point that could perhaps be clarified more in this demo is that pyramid_gaussian returns a generator, and not a sequence of images. It is only because it is wrapped in the call to tuple that we get a sequence that we can index into (and one that we can traverse once to determine the shape and then a second time to build up the composite).

[jni]

I'm wondering if we should revisit that API decision for skimage2, too.

[alexdesiqueira]

@jni could be a good solution. We need consistency :slightly_smiling_face:

[parthgvora]

~Hi all - newbie to open source stuff looking to get involved. Would it be ok if I begin with this issue?~

[lagru]

@parthgvora, sorry for the silence on our side. Please feel free to work on this in case you are still interested!

[parthgvora]

Sorry for my silence as well - things got rather busy for me. I'll let another handle this - I don't have the free time unfortunately. Apologies!

[lagru]

No problem and thanks for letting us now! :slightly_smiling_face:

[jakeMartin1234]

Hi, I'm currently working on it and should have a Merge Request out by tomorrow!

[jakeMartin1234]

Hi, I made the required change in https://github.com/scikit-image/scikit-image/pull/6835. Please let me know if you think it needs any more changes.