Special case compute_identity edges to forward ref types

[ffreyer]

Description

When passing a node of graph A as an input to graph B a new node is created in graph B which connects to the node in A with compute_identity. This was done to avoid sharing nodes between multiple graphs so that nodes are easier to manage. Since this was just using the normal machinery, the new node would pin whatever type it first gets.

This pr changes that behavior by special casing compute_identity. It now reuses the Refs in input nodes directly, which preserves their potentially abstract eltypes and also allows us to skip a bunch of update code. So maybe this is also slightly faster.

This is in relation to JuliaDataCubes/PyramidScheme.jl#78.

Type of change

• Bug fix (non-breaking change which fixes an issue)

Checklist

• Added an entry in CHANGELOG.md (for new features and breaking changes)
• Added or changed relevant sections in the documentation
• Added unit tests for new algorithms, conversion methods, etc.
• Added reference image tests for new plotting functions, recipes, visual options, etc.

[MakieBot]

Benchmark Results

SHA: bd67d09aa162fe2b3b1b4df191afe3e3fdd54f06

Warning

These results are subject to substantial noise because GitHub's CI runs on shared machines that are not ideally suited for benchmarking.

[SimonDanisch]

Seems like a good idea and ready to merge if tests pass.