Generic alignment adapter interface and transform helpers

This file defines a minimal, method-agnostic interface for lifting pairwise manifold alignment algorithms to the multi-domain setting. Adapters for concrete methods (e.g., GRASP, PARROT, GPA, KEMA) should implement these generics for their aligner classes. Create a simple aligner descriptor

Usage

new_aligner(name, group, supports_multi = FALSE, ...)

Arguments

name

Character name of the algorithm

group

One of "O" (orthogonal), "GL" (linear), "perm" (assignment), or "kernel" (embedding-only)

supports_multi

Logical; if TRUE, adapter should also implement a native multi-domain entry point via fit_many()

...

Additional fields stored on the object

Value

An object of class c("<name>_aligner", "aligner")

Examples

algo <- new_aligner("my_method", group = "O")