A couple of tips from the teaching team for our previous milestones on Gradescope

[Arc-Celt]

• 1. Add imports in __init__.py:

  from sharpedge.reposition_image import reposition_image
  from sharpedge.frame_image import frame_image
  from sharpedge.modulate_image import modulate_image
  from sharpedge.pooling_image import pooling_image
  from sharpedge.pca_compression import pca_compression
  from sharpedge.seam_carving import seam_carve
  

  So that when someone imports them later, they can do so by using, e.g.:

  from sharpedge import reposition_image
  from sharpedge import frame_image
  

• 2. Once you have your functions implemented, I'd also add some example images under each function on the README file to better illustrate what they do.

• 3. Function "reposition_image" could be broken into the smaller tasks (rotate, flip, shift), otherwise it tries to accomplish too much, writing tests would quickly become too complex. Note that you can still have a high-level function that can call other functions that do just one job (to maintain a high level API like you currently do).

• 4. Same with "modulate_image", break it down into the different tasks (channel extraction, image conversion, channel swap)

  [Note from Archer] Since we've already done writing all the complex test cases... Can we leave the function as they are now?

• 5. Add whether we use sklearn's PCA or implement our own in the docstring.

• 6. [For future branches] Define a branch prefix or naming convention, e.g. feature/, fix/, docs/, so it's more consistent and you can clearly see what the work being done is about.

• 7. There are unnecessary files in the release packages (pycache).

• 8. Some functions are quite large (like modulate_image and frame_image). I recommend breaking them down (keep the high level API) to make testing easier.

• 9. There are some tests that tackle a lot of different errors (e.g. test_error_cases()), it might be a good idea to break it down further so you know precisely what's causing issues.