Simplify and optimize some GraphNode find functions

[NewboO]

Follow-up of #4603

This PR simplifies GraphNode.find and GraphNode.findOne.
It also avoids creating and merging lots of arrays in GraphNode.find which means less garbage.
In addition, GraphNode.findByName now simply calls GraphNode.findOne so the traversing logic isn't duplicated anymore.

I confirm I have read the contributing guidelines and signed the Contributor License Agreement.

[Maksims]

Do you have a benchmarks with various hierarchy cases?

[NewboO]

Do you have a benchmarks with various hierarchy cases?

I ran somes on one of my projects which has a hierarchy of ~800 nodes.
Benchmarked from root and from a deeper node (only 80 descendants), finding by name both existing (100 occurrences to return from root, 5 from the deeper node) and non-existing:

GraphNode.find, existing name

                       |  old  |  new
--------------------------------------
800+ nodes (100k runs) |  5.3s |  2.5s
--------------------------------------
80 nodes (1M runs)     | 0.98s | 0.75s


GraphNode.find, non-existing name

                       |  old  |  new
--------------------------------------
800+ nodes (100k runs) |  4.6s |  2.5s
--------------------------------------
80 nodes (1M runs)     | 0.63s | 0.70s


GraphNode.findOne, existing name relatively deep

                       |   old  |   new
----------------------------------------
800+ nodes (100k runs) |  0.29s |  0.17s
----------------------------------------
80 nodes (1M runs)     | 0.027s | 0.036s


GraphNode.findOne, non-existing name

                       |  old  |  new
--------------------------------------
800+ nodes (100k runs) |  4.5s |  2.7s
--------------------------------------
80 nodes (1M runs)     | 0.56s | 0.69s

So it may be sometimes a bit slower when only a few nodes need to be checked, but it gets way faster on bigger hierarchies.

[slimbuck]

This PR seems to me a wonderful example of how simplifying code can make it more readable, maintainable and run faster all at the same time. Great job @NewboO !!

[mvaligursky]

Great work! I'm happy to merge this. @slimbuck ?

[willeastcott]

LGTM! Thanks. 🙏