Some corrections for MRAF in the package

[ichristen]

Discussed in #64

^{Originally posted by Shuyun1247 October 27, 2023}
Hello,

I've been using your package, and it's been working great for me. However, I encountered some issues when using the MRAF method in your code, specifically when setting the signal region be a rectangular "null_region". I've identified some areas in algorithm.py that, upon modification, resolve these issues. Here are the changes I suggest:

1. Modification in ijcam_to_knmslm function:
   In algorithms.py, find the definition of the function ijcam_to_knmslm. Before the affine transformation, add the line b = b.reshape(-1) to ensure the correct shape for b.

   Original:

   cp_img = cp.array(img, dtype=self.dtype)
   cp.abs(cp_img, out=cp_img)
   
   # Perform affine.
   ...

   Modified:

   cp_img = cp.array(img, dtype=self.dtype)
   cp.abs(cp_img, out=cp_img)
   
   b = b.reshape(-1)
   
   # Perform affine.
   ...

2. Modification related to null_knm and null_region_knm:
   Find the following section in algorithms.py and modify it as follows to account for both null_knm and null_region_knm.

Original:

if self.null_knm is None:
 ...

Modified:

if self.null_knm is None and self.null_region_knm is None:
 ...
...

...

I'm not entirely sure if I was using the SpotHologram incorrectly, but after making these corrections, the method worked properly for both "null_vector" and "null_region" for us. I hope these suggestions are helpful. Let me know if you need any more details or if there are any concerns.

Thank you!