Generalized, polarization aware anisotropic ratio #147
Conversation
…tions-reflections' of https://github.com/usnistgov/PyHyperScattering into 133-feat-generalize-ar-calculations-chi-center-polarizations-reflections
cbishop4
left a comment
cbishop4
left a comment
There was a problem hiding this comment.
All looks solid; on to testing
| img (xarray): xarray to work on | ||
| chi (numeric): q about which slice should be centered, in deg | ||
| chi_width (numeric): width of slice in each direction, in deg | ||
| def slice_chi(self, chi, chi_width=5, do_avg: bool = True): |
There was a problem hiding this comment.
Couldn't the handling on this be greatly simplified by making use of the _reRange_chi function later? Or is there a reason we don't want to do that?
| print(f' Pol 0: {pol0}') | ||
| print(f' Pol 90: {pol90}')''' | ||
|
|
||
| def _reRange_chi(self, chi): |
There was a problem hiding this comment.
As somewhat noted above, I think this could maybe be applied more to other functions to greatly simplify them. @pbeaucage , are you married to the "nshift" handling within the slice_chi function? It might be easier to just take everything and anytime you're trying to index chi, you just change those things you're trying to index with to valid values
| if (AR_para < AR_perp).all() or (AR_perp < AR_para).all(): | ||
| warnings.warn('One polarization has a systematically higher/lower AR than the other. Typically this indicates bad intensity values.',stacklevel=2) | ||
| # User wants to infer chi centers from beam polarization metadata | ||
| if infer_chi_from_pol == True: |
There was a problem hiding this comment.
Elegant, I like it
cbishop4
left a comment
cbishop4
left a comment
There was a problem hiding this comment.
All looks good with tests
|
Had a conversation with @pbeaucage about the failing test_AR_unity - looking to refactor that now. |
pbeaucage
changed the title
…er-polarizations-reflections
…er-polarizations-reflections
|
@delongchamp I'm going thru stale PRs and I think (think) the only thing this is hung on is a validation that we expect A of a perfect sine wave to be 0.31478. I feel like you came to a conclusion on an analytical solution to this during the code camp; any chance this survived? Or if enough of us bless that as a heuristic solution, then we can just do that... I'm not picky. |
|
@pbeaucage I'm sorry to be the holdup here but I really want to give this one a more thorough review. IT's difficult to strike a balance between flexibility - calculating A with different strategies such as single, multiple polarization and potentially accomodating biaxiality, and transparency, letting the user know how A is/was calculated. |
Hackathon final push for Bijal's AR work.