Count sockets by number when balancing column heights

[BradyAJohnston]

The balancing search hashed the `Socket` dataclass for every candidate move, which was most of `arrange`'s time on the decoder benchmark. Sockets are now numbered once per search.

`test_arrange_decoder_4bit[sugiyama]` locally: 0.20 s → 0.09 s. Layouts unchanged (corpus snapshots pass). This also covers the regression in #254, which gives the search more to do.

[codspeed]

Merging this PR will improve performance by ×2.2

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 1 improved benchmark
✅ 14 untouched benchmarks

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	test_arrange_decoder_4bit[sugiyama]	5.3 s	2.4 s	×2.2

Tip

Curious why performance improved? Comment @codspeedbot explain why performance improved on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing balance-socket-ids (6e78b4d) with main (3f0481c)}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 98.29%. Comparing base (3f0481c) to head (6e78b4d).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #268   +/-   ##
=======================================
  Coverage   98.29%   98.29%           
=======================================
  Files         105      105           
  Lines       31104    31106    +2     
=======================================
+ Hits        30575    30577    +2     
  Misses        529      529           

☔ 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.