Spectral curve and Lorentz correction support

[jglaser]

This PR is a proof-of-concept for how to integrate physics-based corrections as a deterministic scaling function, allowing for a single trainable, and multiplicative intensity scale. See issue #194 for a discussion of the conceptual dissonance between this physics-based approach and the standard data-driven one. We should refine this PR as necessary.

Example usage:

Apply the wavelength-dependent instrument correction, stored in a two-column text file called spectral_curve.txt, and the analytical Lorentz correction, during harmonic deconvolution:

careless poly \
  --iterations=30_000 \
  --merge-half-datasets \
  --separate-files \
  --wavelength-key='WAVEL' \
  --spectral-file spectral_curve.txt \
  --lorentz-correction \
  --trainable-spectral-scale \
  --wavelength-range 2 4 \
  "WAVEL" \
  uncorrected_data_in_asu.mtz \
  my_output_prefix

[jglaser]

This PR should be ready for review and CI

[kmdalton]

This looks good for the most part. I highlighted some places where the documentation could be improved (but it is pretty good already). I asked for two small refactors to your Scaler.

Finally, please figure out whether your Scaler is compatible with --scale-file (see test). If so, let's add an explicit test for it. Otherwise, we should warn users that it isn't supported. With at most one trainable variable, my suspicion is it won't be a heavily a used feature, but I'd just like to make sure.

[kmdalton]

hi @jglaser -- just checking in. let me know if you need any help with this or if my review wasn't clear.

[jglaser]

Finally, please figure out whether your Scaler is compatible with --scale-file (see test). If so, let's add an explicit test for it. Otherwise, we should warn users that it isn't supported. With at most one trainable variable, my suspicion is it won't be a heavily a used feature, but I'd just like to make sure.

done

[jglaser]

hi @jglaser -- just checking in. let me know if you need any help with this or if my review wasn't clear.

Hi @kmdalton .... thank you for your detailed review. I applied the changes (after successfully returning from vacation :). The PR should be complete now.