Derotate with shifts and ellipses

[lauraporta]

Description

Type of PR

• Bug fix
• New feature
• Other

Why is this PR needed?

This PR enables the derotation of movies where the transformation is elliptical, leveraging homography and shear principles. It also addresses key bugs in the Rotator class.

Summary of Changes

Main Updates:

1. Fixed multiple bugs in the Rotator class.
2. Integrated apply_homography into the main derotation function.
3. Improved docstrings throughout the codebase.
4. Added logic for calculating orientation and rotation angle in the fit_ellipse module, ensuring a fully integrated test pipeline.

Refactoring:

5. Split the large integration test (tests/test_integration/test_finding_center_of_rotation.py) into two separate components:
   • derotation/simulate/synthetic_data.py: Contains logic to generate synthetic rotated and derotated movies for tests and examples.
   • tests/test_integration/test_derotation_with_simulated_data.py: An integration test using synthetic data.
   • This test can also be run as a script to generate figures, aiding in the development of new derotation features.
6. Conceptual change in testing approach: Instead of feeding a fixed center of rotation, we now specify the shift from the expected center based on image shape. This ensures compatibility with images of various sizes.
7. Moved blob detection logic: Previously in incremental_derotation_pipeline, now extracted into a dedicated blob_detection module. Its functionalities help in the following:
   • Identify cell positions post-rotation.
   • Fit ellipses.
   • Assess derotation quality—if the blob centers remain stable, the derotation is considered successful.

Testing

• Added test cases covering various critical parameters.
• Tests generate debugging plots to aid analysis.

Breaking Changes

• No breaking changes introduced.

Documentation Updates

• Updated relevant docstrings.

Checklist

• Code has been tested locally.
• Added tests for all new functionality.
• Updated documentation to reflect changes.
• Code formatted using pre-commit.

[codecov]

Codecov Report

Attention: Patch coverage is 67.04981% with 86 lines in your changes missing coverage. Please review.

Project coverage is 51.56%. Comparing base (5d95792) to head (66b5353).
Report is 24 commits behind head on main.

Files with missing lines	Patch %	Lines
derotation/simulate/synthetic_data.py	59.28%	68 Missing ⚠️
derotation/analysis/blob_detection.py	57.89%	16 Missing ⚠️
derotation/derotate_by_line.py	90.90%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #35      +/-   ##
==========================================
+ Coverage   47.26%   51.56%   +4.29%     
==========================================
  Files           7        9       +2     
  Lines         677      894     +217     
==========================================
+ Hits          320      461     +141     
- Misses        357      433      +76     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[IgorTatarnikov]

Looks good, just a few comments here and there

[lauraporta]

Thank you, I've committed a few of your suggestions