Architecture
=============

Module Layout
--------------

.. code-block:: none

   psipose/
   ├── core/                        # Core protocols
   │   ├── ansatz.py                # Ansatz protocol
   │   ├── feature_map.py           # FeatureMap protocol
   │   ├── measurement.py           # Measurement protocol
   │   └── quantum_model.py         # QuantumModel protocol
   ├── feature_maps/                # Data encoding circuits
   │   ├── base.py                  # FeatureMap abstract class
   │   ├── angle.py                 # AngleFeatureMap
   │   └── amplitude.py             # AmplitudeFeatureMap
   ├── ansatze/                     # Parameterized quantum circuits
   │   ├── base.py                  # Ansatz abstract class
   │   ├── hardware_efficient.py    # HardwareEfficientAnsatz
   │   └── strongly_entangling.py   # StronglyEntanglingAnsatz
   ├── measurements/                # Quantum measurements
   │   ├── base.py                  # Measurement abstract class
   │   ├── expectation.py           # PauliZExpectation, etc.
   │   └── probability.py           # ProbabilityMeasurement
   ├── models/                      # Pure quantum models
   │   ├── variational.py           # VariationalModel
   │   └── kernel.py                # QuantumKernel
   ├── estimators/                  # sklearn-compatible estimators
   │   ├── base.py                  # QuantumEstimator
   │   ├── classification.py        # VQCClassifier, QSVC
   │   └── regression.py            # VQCRegressor
   ├── training/                    # Training utilities
   │   ├── loss.py                  # Loss functions
   │   └── optimizer.py             # Optimizer adapter
   └── preprocessing/               # Sklearn transformers
       ├── angle_scaler.py
       └── amplitude_normalizer.py

Key Abstractions
-----------------

**FeatureMap:** Takes classical features and applies them to qubits.

**Ansatz:** A parametrized quantum circuit with trainable weights.

**Measurement:** Quantum measurement operator.

**Estimator:** The sklearn-compatible class combining feature map + ansatz + measurement + optimizer.

**VariationalModel:** Pure quantum model combining feature map + ansatz + measurement.

**QuantumKernel:** Computes kernel matrix from quantum state overlaps.