Relabel dimension names to integers when contracting

[mtfishman]

Summary

Relabels the dimension names to integers before contracting in mul_nameddims, so the contraction-label bookkeeping in TensorAlgebra.contract runs on Ints rather than on the dimension-name type. A contraction's structure depends only on the equality pattern of the names, so the names are matched to integers once (shared names get the same integer), the bookkeeping runs on the integers, and the result names are recovered by position.

This is aimed at IndexName, which carries a UUID and a tag dictionary and is comparatively expensive to compare and store. The previous path handed the IndexNames straight to TensorAlgebra.contract, which then re-compared them across every setdiff and findfirst pass. Matching once up front collapses that to a single name-matching pass. The speedup grows with the number of indices, from roughly 7% for two rank-2 tensors to roughly 34% for two rank-5 tensors, in the small-dimension regime where this bookkeeping is the dominant cost. The saving is a fixed per-call amount, so it fades to nothing at larger dimensions once the matrix multiply dominates.

Builds directly on the type-stable contraction-label derivation in ITensor/TensorAlgebra.jl#192, which is what makes Int labels cheap to run through contract. Confined to mul_nameddims, with no change to IndexName. The in-place mul! paths are left as-is.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 73.99%. Comparing base (bde999b) to head (704b10d).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #196      +/-   ##
==========================================
+ Coverage   73.77%   73.99%   +0.21%     
==========================================
  Files          29       29              
  Lines        1548     1561      +13     
==========================================
+ Hits         1142     1155      +13     
  Misses        406      406              

Flag	Coverage Δ
docs	24.90% <100.00%> (+0.54%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

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