Fix 1d plots y-axis limits when toggling log-scale with buttons and fix 2d plot colorbar when Inf values

[nvaytet]

1d plots:

Toggling to log scale using the toolbar button when the current axis limit goes below 0 (which is quite common because we add a little padding around the range of values) would cause the axis to completely mess up.
This is due to a behaviour in matplotlib that when axis limits are not explicitly set, the change to log scale correctly figrues out the limits, but when the limits are set (and we always set the limits), then it tries to use the number that were set and it gets a negative number as the starting point for the log scale.

We now use rescale_to_data to set the limits on the axis before changing the axis scaling to log.

Example:

import scipp as sc
sc.plot(sc.arange('x', 100.))

Note that the issue only appeared when using the toolbar button. Doing sc.plot(..., norm='log') found the correct limits.

Before:

After:

2d plots:

Issue when trying to find colorbar limits when Inf values present (which happens often when you are normalizing and you divide by 0)

import numpy as np
import scipp as sc

a = np.arange(100.0)
a[0] = np.nan
a[-1] = np.inf
b = sc.array(dims=['x'], values=a, unit='K').fold('x', sizes={'x': 10, 'y': -1})
sc.plot(b)

Before:

After:

[SimonHeybrock]

Isn't this just x.values?