fix: SGConv layer fixes after implementing support for GNNHeteroGraph

[askorupka]

This PR introduces some optimizations and fixes for SGConv.

Fix:

• previously the loop for propagation involved x = x .* c' and it was the only place where x was referred to after expansion to xi and xj. in the propagation process xj was passed which made this line of code unnecessary and potentially problematic (remember that for GNNHeteroGraph x contains both xi and xj)

Optimizations:

• reduces number of ifs as definition for edge_t not necessary anymore (after changing add_self_loop behavior it was used only once in degree)
• introduces optimization for multiplying weights before convolution if Dout < Din (but only when g is not GNNHeteroGraph

[askorupka]

closing as per above discussion with @CarloLucibello, SGConv layer GNNHeteroGraph support to be removed by me.