Fix align1D

[ericpre]

As reported by @mwalls on gitter, the align1D is not working properly. The reason is that the data is not normalised before computing the cross correlation. I am surprised that it was not reported before!

Progress of the PR

• Normalise by subtracting the mean before calculating the cross correlation,
• add tests,
• add entry to CHANGES.rst ,
• ready for review.

Minimal example of the bug fix

scale = 0.1
g = hs.model.components1D.Gaussian(sigma=scale*5)
x = np.stack([np.linspace(-5, 5, 100)]*5)
s = hs.signals.Signal1D(g.function(x) + 1E5)
s.axes_manager[-1].scale = scale
shifts = np.random.random(len(s.axes_manager[0].axis)) * 2
shifts[0] = 0
s.shift1D(-shifts, show_progressbar=False)
shifts2 = s.estimate_shift1D(show_progressbar=False)
np.testing.assert_allclose(shifts, shifts2, rtol=0.3)

[codecov]

Codecov Report

Merging #2675 (65958dd) into RELEASE_next_patch (8363b6e) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@                 Coverage Diff                 @@
##           RELEASE_next_patch    #2675   +/-   ##
===================================================
  Coverage               76.62%   76.62%           
===================================================
  Files                     201      201           
  Lines                   29684    29681    -3     
  Branches                 6500     6500           
===================================================
- Hits                    22744    22743    -1     
  Misses                   5177     5177           
+ Partials                 1763     1761    -2     

Impacted Files	Coverage Δ
hyperspy/_signals/signal1d.py	72.95% <100.00%> (+0.24%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 8363b6e...65958dd. Read the comment docs.

[jlaehne]

Looks good and works on both the test provided above and on the example provided from @mwalls on gitter.