Fix stackoverflow in renormalize_southwest_corner

[leburgel]

Bumped into this by accident: using renormalize_southwest_corner directly on an enlarged corner and its projectors gave a stackoverflow, must have been a typo.

[Yue-Zhengyuan]

Looking back on this, it seems that all four renormalize_xxxx_corner(quadrant, P_left, P_right) are simply wrappers to the same renormalize_corner(quadrant, P_left, P_right). How about we directly write, for example,

function renormalize_northwest_corner((row, col), enlarged_env, P_left, P_right)
    return renormalize_corner(
        enlarged_env[NORTHWEST, row, col],
        P_left[NORTH, row, col],
        P_right[WEST, _next(row, end), col],
    )
end

to reduce one layer of wrapping?

[leburgel]

I don't think this works, because the enlarged_env can also contain a sparse representation of the enlarged corners, as EnlargedCorner instances. So the extra layer is required to dispatch to the EnlargedCorner implementations like

function renormalize_southwest_corner(ec::EnlargedCorner, P_left, P_right)
    return renormalize_southwest_corner(ec.E_1, ec.C, ec.E_2, P_left, P_right, ec.A)
end

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

Files with missing lines	Coverage Δ
...lgorithms/contractions/ctmrg/renormalize_corner.jl	52.63% <100.00%> (+8.77%)	⬆️

... and 10 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Yue-Zhengyuan]

Is it convenient to add some test coverage here? In #326, I observed that tests only cover half of the CTMRG code.

[leburgel]

Yes, we should add some coverage here before merging, seems like a good opportunity.

[leburgel]

I added a small test to unblock this for now. We can add to this test stub afterwards until all contractions are covered by just comparing dense and sparse versions, even if there are some that aren't actually used in any of the algorithms yet.

[Yue-Zhengyuan]

Good to go for me. Though we need to reduce duplication of tests some time in the future.