Remove duplicate objects without relying on Hashing

[MicahGale]

Right now duplicates are removed by using hashes and the set operator. However these objects are mutable and these values changing can lead to unpredictable results. See: https://eng.lyft.com/hashing-and-equality-in-python-2ea8c738fb9d.

This affects most NumberdObjectCollection children objects.

[MicahGale]

Thanks to @tjlaboss for the references as well:

https://github.com/openmc-dev/openmc/blob/develop/openmc/mixin.py#L9
openmc-dev/openmc#872
openmc-dev/openmc#946

[MicahGale]

marked this issue as related to #97

[MicahGale]

Without hashing the best we can do is: O(nlog(n)): https://stackoverflow.com/questions/46554866/efficiently-finding-duplicates-in-a-list.

However, we have a hash-like cheat: numbers.

So I think we should:

1. find number collisions with NumberedObjectCollection
2. Then do a O(nlogn) - O(n2) comparison of all of the collisions.