Return None When There Isn't Meaningful Letterboxing

[mjs2600]

The current implementation will return the original image if there isn't any meaningful letterboxing around the image. This PR changes that so that the caller knows if an image was modified or not.

[Copilot]

Pull request overview

This PR changes the behavior of the unletterbox function to return None when no meaningful letterboxing is detected, instead of returning the full image bounds. This allows callers to distinguish between images that were actually modified versus those that didn't need processing.

Changes:

• Modified unletterbox to return None when letterboxing is not detected or not meaningful
• Updated load_and_preprocess to handle the new None return value correctly
• Updated tests to expect None instead of full image bounds for unchanged images

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

File	Description
perception/hashers/tools.py	Changed three return statements from returning full image bounds (0, w), (0, h) to returning None, and updated documentation
perception/local_descriptor_deduplication.py	Updated load_and_preprocess to check if unletterbox returns None and only crop if bounds are returned; removed unnecessary parentheses in lambda
tests/test_tools.py	Updated tests to expect None instead of full image bounds; added formatting improvements; contains a critical bug in test_ffmpeg_video

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.