Custom distance function in KNN

## 🚀 Feature

Use of custom distance function while calculating k nearest neighbour.

NOTE: Please look at the existing list of Issues tagged with the label ['enhancement`](https://github.com/facebookresearch/pytorch3d/issues?q=label%3Aenhancement). **Only open a new issue if you do not see your feature request there**.

## Motivation

It would be great if we can use custom distance function while calculating k nearest neighbour instead of euclidean distance.

## Pitch

In Molecular dynamics simulations, the simulations are carried out in a fixed size box so while calculating k nearest neighbour we have to use periodic boundary condition and minimum image convention so it will be convenient if we can pass custom distance function for calculating the distance.

For example:

Box boundaries (3,3,3)

Point 1 coordiante: (0.2,0.2,0.2)
Point 2 coordinate: (2.7,2.7,2.7)

Euclidean distance: 4.33
Pbc distance: 0.866

NOTE: we only consider adding new features if they are useful for many users.