Add FiniteMPS getindex

[VictorVanthilt]

This PR implements getindex for FiniteMPS using both Integers and AbstractUnitRanges.

[VictorVanthilt]

@lkdvos can you look at the implementation as well as if the Base.@propagate_inbounds magic would work with this? I haven't worked with this before.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Files with missing lines	Coverage Δ
src/states/finitemps.jl	79.55% <100.00%> (+1.03%)	⬆️

... and 1 file with indirect coverage changes

[lkdvos]

Looks okay to me!

For the boundscheck business: the long story short (ish) is that @inbounds will bypass code that is marked within a @boundscheck block, on the condition that it is inlined, or passed through a function barrier with @propagate_inbounds.

Boundschecking itself happens by the function checkbounds, which has two signatures: checkbounds(Bool, args...) and checkbounds(args...). The first outputs a bool to check the bounds, the second calls the first and throws if necessary.

In general, the boundschecking really only matters for performance-critical sections, or expensive boundschecks. I like putting an extra check in the function, as this gives clearer error messages, but putting @inbounds should only be used as a last resort, since that can lead to segfaults.

Could you try and add some tests for this, testing inferred, proper error throwing etc? Nothing too major, but just a few lines to ensure everything is (and keeps) functioning property would be great.