Use 0.01 increment for channel sliders

[will-moore]

Fixes #467.

To test - see https://merge-ci.openmicroscopy.org/web/figure/file/95252 for example figure and images (user-3)

• Channel sliders should now have an increment of 0.01 allowing greater control, particularly with floating-point images with small ranges
• When you type values into the start/end input fields, they are no-longer forced to be Integers:

[dominikl]

It works. But a downside is that it also applies to int type images.

[will-moore]

Since we don't store pixel-type info for images in figure, I've used the maximum pixel value of the range slider to determine if the slider increments should be small 0.01 or integers. The current threshold is 100. Below this we use 0.01 for slider increment.
I think this mostly behaves OK. It means you can have different behaviour for different channels of an Image, but if one channel is quite dark, that's probably OK.

[dominikl]

👍 Looks good to me.

[jburel]

@will-moore is it possible to make the textbox a bit bigger to avoid the truncation?

[will-moore]

@jburel When are you seeing this truncation? I can't reproduce, since we only allow non-integer values in the text boxes if the max intensity is less than 100, and then it's formatted to 2 decimal places, so a max of 4 characters, which looks OK:

[will-moore]

In discussion with @jburel, the truncation he was seeing was in the first screenshot above, before 5f0bf0f added the formatting of those text inputs. So I think this is good to merge?