refactor: apply_interpolation_matrix() is now compute_stretched_components()

[john-halloran]

The original apply_interpolation_matrix() was kind of a mess. The core logic, which is honestly not that complicated, was buried under a sea of indexing and biasing hacks, many of which were unnecessary mimicry of the data format of the original MATLAB code.

It should be hopefully now be clear that this function:

1. Takes the original sample indices and stretches them.
2. Finds the sample before and after the stretched position, and linearly interpolates between them.
3. Weights the samples per component.
4. Computes derivatives of these stretched components with respect to stretch for use by solvers.

To fit this theme of simplicity, the function has been renamed to better match what it is used for.

[sbillinge]

merging to keep this going, but can we think about a test for this function? some very small matrix?