Responds to #245, and an alternative to #246

This PR adds functionality for making it easier to plot a model over a broader frequency range (across a frequency range beyond the fit range). It does so by making it easier to pass in frequency and power vectors to plot with a model. This strategy means that the model object does not store the broader range, but during exploration phases this can be visualized.

Notes:

• This strategy generally means the broader range spectrum needs to be "around" (in memory) to be passed in explicitly.
  • A limitation of this is that if a fooof object is reloaded, the broader range spectrum while not de facto be available for this.
  • The upshot of this is trying to find a balance between supporting evaluation without increasing the amount of data stored in the object and/or saved out, which could otherwise lead to managing a lot of extra data that is rarely used.
• The above points are in relation to the alternate strategy explored in [ENH] Enable full plot range for fm.plot() #246 (I commented in that PR why I personally do not prefer that option).
• Below is a fairly comprehensive exploration of this functionality, with the idea that this can / should be turned into a docsite example to illustrate the underlying points here, and show the functionality

For reviewing, @ryanhammonds and @voytek (if you have time) - before a code deep-dive, I would appreciate your thoughts on strategy here (comparing this to #246). If we like this approach then code should be reviewed as normal - I think it should be pretty solid, although there could be some tweaks on how to manage the API for FOOOF.report if y'all have any thoughts on that.

Demonstration

Code for running following examples:

from fooof import FOOOF
from fooof.sim import gen_power_spectrum
freqs, powers = gen_power_spectrum([1, 150], [1, 100, 1], [[10, 0.15, 0.5], [25, 0.1, 2.5]])

Using plot_fm to plot models on top of broader range power spectra

# With this PR, can do the following to plot the model on top of a broader frequency range
fm.plot(freqs=freqs, power_spectrum=powers, plt_log=True, freq_range=[3, 75])

Which gives the following output:

Exploring frequency ranges using fm.report

fm = FOOOF()
fm.report(freqs, powers, freq_range=[3, 30], plt_log=True, plot_full_range=True)

This gives the following output:

In addition, these reports can be saved out.

# Save out model report, adding in a broader range frequency spectrum
fm.save_report('test_fm.pdf', freqs=freqs, power_spectrum=powers, plt_log=True, freq_range=[3, 75])

Showing how this works with FOOOFGroup

This process can also be used from a FOOOFGroup object, to examine individual model fits:

from fooof.sim import gen_group_power_spectra
from fooof import FOOOFGroup

freqs, spectra = gen_group_power_spectra(10, [1, 150], [1, 100, 1], [[10, 0.15, 0.5], [25, 0.1, 2.5]])
fg = FOOOFGroup()
fg.fit(freqs, spectra, [3, 30])

# Then, plot the model over the broad range (can also make report / save, etc)
fg.get_fooof(ind).plot(freqs=freqs, power_spectrum=spectra[ind, :])

Output: