Add unit tests to `fit_ellipse_to_points`

[lauraporta]

Is your feature request related to a problem? Please describe.

The function fit_ellipse_to_points currently lacks dedicated unit tests. While it is covered indirectly through integration tests, we need explicit verification that the fitted parameters are meaningful and that the resulting ellipse accurately represents the provided cloud of points. In particular, we are most interested in ensuring the correctness of the ellipse center.

Describe the solution you'd like

Add at least one unit test to validate that fit_ellipse_to_points correctly estimates the center of the fitted ellipse when given a well-defined set of points.

Describe alternatives you've considered

Currently, the function is tested indirectly through integration tests within the broader pipeline. However, this approach does not provide fine-grained validation of its correctness.

Additional context

The function fit_ellipse_to_points takes as input a cloud of points (centers of detected blobs computed elsewhere) and fits an ellipse to them. The primary parameters of interest are the ellipse’s center, eccentricity, and orientation.

[Prabhu6283]

@lauraporta I would like to address this issue .

[Prabhu6283]

@lauraporta My approach to resolve this issue is like - 1) Making a set of points forming a well-defined ellipse with a known center. and 2)Using fit_ellipse_to_points to fit an ellipse to the points. How about this approach let me know.

[onemriganka]

/attempted #36

[Snehallaldas]

To address the need for dedicated unit tests for the fit_ellipse_to_points function, I propose creating at least one comprehensive unit test that specifically validates the ellipse center estimation, which is your primary parameter of interest.

The solution would involve:

Generating synthetic point data with known ellipse parameters (center, axes lengths, orientation)
Feeding these points into fit_ellipse_to_points
Comparing the estimated center coordinates against the known ground truth
Specifically, the test should:

Create a well-defined set of points along an ellipse with precisely known parameters
Add realistic noise to simulate real-world data
Validate that the function recovers the true center within an acceptable tolerance
Include edge cases such as circles (equal axes), sparse point distributions, or partial ellipses
This approach provides direct verification that the function correctly estimates the center coordinates, regardless of how it performs in the broader integration tests. It would give you confidence that the core functionality works as expected before it's used within your larger pipeline.

The unit test would serve as a regression check to ensure that future code changes don't adversely affect the center estimation accuracy, which appears to be critical for application.

What do you think?

[lauraporta]

Hi @Snehallaldas, thanks for your interest in derotation! Welcome to submit a pull request in which you implement your tests ideas. You can have a look at the one made by @onemriganka, and work on it directly if you wish as it is not complete.

[AlgoFoe]

Hi @lauraporta, I am working on adding dedicated unit tests for fit_ellipse_to_points that validate center recovery using synthetic ellipses with known parameters.
I will open a PR shortly, and happy to adjust tolerance or coverage.