fix incorrect calculation of ee_region and improve effiency.

[zhou-yang-quantum]

Hi Ryan,

Thank you for writing this wonderful package. Just some small fixes here.

Let psi be an MPS of length 10. We should have
ee_region(psi, 5:10)==ee_region(psi, 1:4)
But instead, the function actually does,
ee_region(psi, 5:10) == ee_region(psi, 1:5)
which is incorrect.

Moreover, when calculating a region whose complement is a connected region in the middle, e.g.
ee_region(psi, vcat(1:3, 7:10))
it is more efficient to calculate the complement region ee,
ee_region(psi, 4:6)

I have modified the function ee_region for these two improvements.

Best, Zhou

[ryanlevy]

Hi Zhou,

Thank you so much for the PR! Let me look into the bipartite cut function, this may be an issue of documentation as well. At first glance, I imagine this may be an artifact of not moving the MPS center even if we are a "left" or "right" region
Your A U C is definitely useful, good catch.

In the meantime can you please:

• Remove the vscode file
• create a test to catch the error for the bipartite cuts? You can see some examples in the test/test_entropy.jl file

[zhou-yang-quantum]

Hi Ryan,

I just deleted the .vscode file and added a brief test.