Improve efficiency of bond truncation

[Yue-Zhengyuan]

This PR improves the efficiency of bond_truncate (for both ALSTruncation and FullEnvTruncation) by reusing intermediate partial contraction results.

The test test/bondenv/bond_truncate.jl is rewritten to simulate the more realistic scenario of applying a 2-site gate on an iPEPS with physical dimension d and virtual dimension D. It can reveal that FET is much slower when d is large (due to physical leg not transferred to the reduced bond tensors to be optimized).

[Yue-Zhengyuan]

@sanderdemeyer I recommend you (squash-)merge this PR to the NTU branch (#144) when trying it.

[codecov]

Codecov Report

❌ Patch coverage is 97.14286% with 2 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/algorithms/contractions/bondenv/als_solve.jl	95.00%	2 Missing ⚠️

Files with missing lines	Coverage Δ
src/algorithms/truncation/bond_truncation.jl	94.87% <100.00%> (+0.35%)	⬆️
src/algorithms/truncation/fullenv_truncation.jl	96.82% <100.00%> (+0.05%)	⬆️
src/algorithms/contractions/bondenv/als_solve.jl	96.36% <95.00%> (-3.64%)	⬇️

🚀 New features to boost your workflow:

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

[leburgel]

Just a few questions, but otherwise looks good to me.

Test failures and timeouts are unrelated, and will be solved by #364.

[Yue-Zhengyuan]

I dropped the ::Val{i} dispatch on _als_tensor_R, _als_tensor_S so that I can reuse the underlying contraction without having to pass the entire xs vector all the time.