Implementing `ot.emd()` to get the optimal transportation required

[axiomcura]

The Python POTlibrary provides two Earth Mover’s Distance methods: ot.emd() and ot.emd2().

Currently, buscar uses ot.emd2(), which returns a single scalar metric representing the minimal amount of “work” required to transform one population of cells into another.

In contrast, ot.emd() returns the optimal transport plan, which is a matrix describing how each treated cell maps to one or more healthy cells. This matrix can be interpreted as a potential reprogramming map, showing the most efficient way to transform the treated cell population into the healthy population.

This richer output from ot.emd() provides more opportunities for interpretability, enabling deeper insights into the underlying cellular transitions.