Support decomposition of Controlled ops to GraphSolver

[maliasadi]

Context:
The decomposition graph solver could not decompose Controlled ops. This PR adds this support to the graph solver to handle them, mirroring the two decomposition families used by PL:

1. Decompose C(Op) directly: Rules registered against the controlled operator itself: e.g. a dedicated CRX rule, C(SWAP), C(PhaseShift), etc. in PL. These were already supported by the graph. A RuleNode whose output has numControlWires > 0 is registered and solved like any other rule; so no new logic was needed to the solver; they just need to be provided as rules. I'll take care of this in a follow-up PR to integrate the support with the graph-decomposition pass.

2. Decompose Op, then control each produced gate: From each base rule Op, the graph builder should synthesize C(Op) with the same k control wires applied to every produced gate and the same multiplicities. I assumed that only the cost of controlling each produced gate matters here; the actual control-each-gate wrapping happens in the follow-up PR when integrating with the pass.

Both pathways coexist in the graph, and the solver picks the cheapest decomp pathway! :)

Unlike the adjoint case #3001, there is no build-time short-circuit that suppresses one pathway in general; both are always added and left to compete. The one exception mirrors PL's control-specific special cases (e.g. GlobalPhase, ChangeOpBasis): controlling their decomposition is not semantically valid, so I added a static method to the graph to suppress them (isCtrlRuleRequired) and they must be decomposed by a dedicated controlled rule. Since the graph reasons about resource cost and not the semantics of rules, this suppression is required to keep the solver from picking a cheap but invalid controlled decomposition.

Benefits:

• Supporting C(Op) decomposition and resource estimation
• Composes with existing operator identity (numControlWires is matched exactly, like adjoint), so multi-controlled operators is also supported in the graph.

Possible Drawbacks:

• For circuits with many Ctrl ops, the graph grows as we add synthesized controlled rules for each C(Op) and track them across multiple decomp pathways.
• I'm using explicit rules for the special ops (GlobalPhase, ...), and the suppression list is currently hard-coded. This can be revisited later after integration with the graph-decomposition pass.
• `control_values are not supported yet. This can be added later.

Related GitHub Issues:
[sc-122044]

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 96.96%. Comparing base (c30fe9a) to head (64d9ea8).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #3003   +/-   ##
=======================================
  Coverage   96.96%   96.96%           
=======================================
  Files         167      167           
  Lines       19346    19346           
  Branches     1801     1801           
=======================================
  Hits        18758    18758           
  Misses        433      433           
  Partials      155      155           

☔ 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.