refactor: Reorganise the QuantumInfo directory

[jstoobysmith]

No changes to content, just a reorganisation. New directory structure is the following:

QuantumInfo
 ┣ Capacity
 ┃ ┣ Capacity.lean
 ┃ ┗ Capacity_doc.lean
 ┣ Channels
 ┃ ┣ Bundled.lean
 ┃ ┣ CPTP.lean
 ┃ ┣ DegradableOrder.lean
 ┃ ┣ Dual.lean
 ┃ ┣ MatrixMap.lean
 ┃ ┣ Pinching.lean
 ┃ ┗ Unbundled.lean
 ┣ ClassicalInfo
 ┃ ┣ ForMathlib
 ┃ ┃ ┗ Analysis
 ┃ ┃ ┃ ┗ SpecialFunctions
 ┃ ┃ ┃ ┃ ┗ Log
 ┃ ┃ ┃ ┃ ┃ ┗ NegMulLog.lean
 ┃ ┣ Capacity.lean
 ┃ ┣ Channel.lean
 ┃ ┣ Distribution.lean
 ┃ ┣ Entropy.lean
 ┃ ┗ Prob.lean
 ┣ Entropy
 ┃ ┣ Axiomatized
 ┃ ┃ ┣ Defs.lean
 ┃ ┃ ┗ Renyi.lean
 ┃ ┣ DPI.lean
 ┃ ┣ Relative.lean
 ┃ ┣ SSA.lean
 ┃ ┗ VonNeumann.lean
 ┣ ForMathlib
 ┃ ┣  ...
 ┣ Measurements
 ┃ ┗ POVM.lean
 ┣ Operators
 ┃ ┗ Unitary.lean
 ┣ ResourceTheory
 ┃ ┣ FreeState.lean
 ┃ ┣ HypothesisTesting.lean
 ┃ ┣ ResourceTheory.lean
 ┃ ┗ SteinsLemma.lean
 ┣ States
 ┃ ┣ Mixed
 ┃ ┃ ┣ Fidelity.lean
 ┃ ┃ ┣ MState.lean
 ┃ ┃ ┗ TraceDistance.lean
 ┃ ┣ Pure
 ┃ ┃ ┣ BargmannInvariant.lean
 ┃ ┃ ┣ Braket.lean
 ┃ ┃ ┗ Qubit.lean
 ┃ ┣ Ensemble.lean
 ┃ ┗ Entanglement.lean
 ┗ Regularized.lean

[wock9000]

Tested locally — builds clean. Verified that BargmannInvariant.lean works correctly at the new States/Pure/ path with the updated import.

The States/Pure/ vs States/Mixed/ split is clean. For our follow-up PRs (Pancharatnam solid angle, N-level generalization), the placement under States/Pure/ makes sense since geometric phase is a property of pure state evolution.

One note for our own tracking: our open PR #1126 targets QuantumInfo/Finite/GeometricPhase/ which will need rebasing once this merges. Happy to handle that.

[morrison-daniel]

I think the organization makes sense, but I'd like to wait for approval from someone more familiar with QuantumInfo.