Registration

[fabian-s]

remaining todos:

• add landmark registration example / implementation
• implement affine registration (challenge: shorten/lengthen/shift domains, figure out best optimizer)
• include summary of results of sim. experiments in vignette
• fix figure formats/sizes in vignette
• reduce new "Suggests:"

[fabian-s]

[WIP]

[Copilot]

Pull Request Overview

This PR adds comprehensive functional data registration capabilities to the tf package by implementing elastic deformation and alignment functions. The registration functionality enables temporal alignment of functional data by separating phase (timing) variation from amplitude (shape) variation.

• Implements core registration functionality using SRVF and FDA methods
• Adds comprehensive documentation with real-world examples using growth and pinch datasets
• Provides integration capabilities with the mlr3 framework

Reviewed Changes

Copilot reviewed 26 out of 29 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
vignettes/registration.Rmd	Comprehensive tutorial on curve registration with real-world examples
vignettes/mlr3-integration.Rmd	Integration example with mlr3 framework for FDA
tests/testthat/test-vec-cast.R	Removes debug print statements from test helper function
tests/testthat/test-register.R	Comprehensive test suite for registration functionality
tests/testthat/test-assertions.R	Tests for warping function validation
man/*.Rd	Documentation files for registration functions and datasets
data-raw/*.R	Data processing scripts for example datasets
attic/*.R	Development and troubleshooting scripts
_pkgdown.yml	Updates package documentation structure
R/register.R	Core registration implementation with warping/unwarping functions
R/data.R	Documentation for example datasets
R/calculus.R	Function inversion implementation
Multiple other R files	Supporting utilities and improvements

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

There's a spelling error: 'tpyically' should be 'typically'.

Suggested change

	1. Estimating "warping functions" using `tf_register()` - these capture the time scale deformations of the observed functions relative to a (tpyically iteratively estimated) common template.
	1. Estimating "warping functions" using `tf_register()` - these capture the time scale deformations of the observed functions relative to a (typically iteratively estimated) common template.

[Copilot]

The placeholder '!!TODO!!' should be replaced with an actual section reference or removed if the section doesn't exist yet.

Suggested change

- alignments can be fairly sensitive to the specifics of data representation (i.e., what kind of basis representation is being used, how many gridpoints the functions are being evaluated on, how to interpolate between grid points, whether/how much the data was smoothed before registration). Even for a simple looking example like the "Berkeley Growth" data shown below, you can get very different alignment results from the same method, depending on the specific pre-processing - see section !!TODO!!.

- alignments can be fairly sensitive to the specifics of data representation (i.e., what kind of basis representation is being used, how many gridpoints the functions are being evaluated on, how to interpolate between grid points, whether/how much the data was smoothed before registration). Even for a simple looking example like the "Berkeley Growth" data shown below, you can get very different alignment results from the same method, depending on the specific pre-processing.

[Copilot]

There's a spelling error: 'lenghtening' should be 'lengthening'.

Suggested change

	The graph of the original growth trajectory itself is compressed by the alignment early on (slope of unwarping function < 1 for age 2-7, arrows to the left lenghtening), then dilated (steep unwarping function slope for age 7-10, arrows to the left shortening).
	The graph of the original growth trajectory itself is compressed by the alignment early on (slope of unwarping function < 1 for age 2-7, arrows to the left lengthening), then dilated (steep unwarping function slope for age 7-10, arrows to the left shortening).

[Copilot]

The regular expression '$^' will never match any string because it looks for end-of-string followed by start-of-string, which is impossible. This should likely be '^$' to match empty strings.

Suggested change

	if (is.character(x)) {
	x <- sub("$^", "NA", x)
	x <- sub("^$", "NA", x)