v0.9.2 — 200+ Qubit Simulation (MPS + Sparse)

🚀 Quanta SDK v0.9.2

Highlights

• 200+ qubit simulation via new MPS (tensor network) simulator
• Sparse statevector for 40-50 qubit circuits with O(k) memory
• SimulatorBackend ABC — clean architecture for all 6 backends
• Circuit-aware router — automatic simulator selection
• Security hardening — exec() sandbox, eval() elimination, traceback fix
• L3 decoupling — all 10 algorithms use factory pattern

New Simulator Backends

Simulator	Max Qubits	Memory	Best For
Dense	27	O(2ⁿ)	Full accuracy
Sparse (NEW)	50	O(k)	GHZ, oracle, product
MPS (NEW)	200+	O(n·χ²)	QAOA, VQE, 1D circuits
Clifford	1000+	O(n²)	Stabilizer circuits

Quality

• 820 tests pass (up from 774)
• 0 ruff lint errors
• 91% code coverage
• Cross-validated: all new simulators verified against dense simulator

Security (from v0.9.1)

• exec() sandboxed with _SAFE_BUILTINS + _validate_code()
• eval() eliminated — pure AST walker in QASM import
• All MCP error responses sanitized (no traceback leaks)

Install

pip install --upgrade quanta-sdk

Full changelog: CHANGELOG.md

v0.9.0 — Primitives, @quantum, Async, Benchmarks

What's New in v0.9.0

Estimator/Sampler Primitives (IBM V2 Compatible)

• Estimator().run(circuit, observables=[('ZZ', 1.0)]) — exact ⟨ψ|O|ψ⟩
• Sampler().run(circuit, shots=4096) — measurement sampling with batch
• Both support run_async() for parallel batch execution

@quantum Decorator (PennyLane @qml.qnode Equivalent)

• @quantum(qubits=2, observable=[('ZZ', 1.0)])
• .expectation() — exact expectation values
• .gradient() — parameter-shift rule, verified analytical gradients

Async Execution

• run_async(circuits, shots=N) — top-level async batch runner

Benchmark Suite

• 10 benchmarks: Bell (0.45ms), GHZ-20 (83ms), Gradient (0.51ms)

Property-Based Testing

• 21 Hypothesis tests: gate unitarity, Pauli algebra, determinism

Jupyter Notebooks

• 14 notebooks with Google Colab badges

Quality

• 669 tests, 89% coverage
• 0 ruff lint errors

v0.8.1 — GATE_REGISTRY, ReadoutError, Docstrings

Bug Fixes

• GATE_REGISTRY: All 25 gates now registered (added RX, RY, RZ, P, U, RXX, RZZ)
• Simulator: Fixed _get_gate_matrix() in statevector AND density matrix to handle parametric gates from registry
• ReadoutError: Wired into runner pipeline — apply_to_counts() now applies post-measurement noise
• Docstrings: Filled record(), build(), circuit() Args/Returns/Raises

Quality

• 530 tests, 87% coverage
• Lint clean (Ruff)

v0.8.0 — QML, Multi-Backend, IBM Hardware, 14 MCP Tools

🚀 What's New in v0.8.0

Quantum Machine Learning

• QuantumClassifier — Variational quantum classifier with parameter-shift gradients
• QuantumKernel — Quantum kernel matrix K(x,y) = |⟨φ(x)|φ(y)⟩|²
• 3 Feature Maps — Angle, ZZ, Amplitude encoding

IBM Quantum Hardware

• Real hardware submission — Submit circuits to ibm_torino, ibm_fez via REST API
• Job polling — ibm_job_result tool to fetch results from IBM
• ISA transpilation — Automatic Heron native gate decomposition

14 MCP Tools for AI

• run_circuit, create_bell_state, grover_search, shor_factor
• simulate_noise, draw_circuit, list_gates, explain_result
• monte_carlo_price, qaoa_optimize, cluster_data
• run_on_ibm, ibm_backends, ibm_job_result (NEW)

Multi-Backend

• IBM Quantum (156 qubits, Heron r3)
• IonQ (36 qubits, trapped ion)
• Google Quantum (72 qubits, Sycamore)
• Local simulator (27 qubits)

Bug Fixes

• QMC precision plateau eliminated (golden-section MLE)
• Noise fidelity: Kraus density matrix + Bhattacharyya
• Shor prime input validation
• Sandbox import restrictions removed

Quality

• 515 tests, 87% coverage
• Lint clean (Ruff)
• Python 3.10+ compatible

v0.7.0 — Quantum Monte Carlo, QAOA Optimizer, Clustering

What's New in v0.7.0

3 New Quantum Algorithms

• Quantum Monte Carlo (layer3/monte_carlo.py): Amplitude estimation (BHMT) for option pricing — European call/put, lognormal/normal/uniform distributions
• QAOA Optimizer Loop (layer3/optimize.py): scipy COBYLA variational parameter training with multi-start optimization
• Quantum Clustering (layer3/clustering.py): Swap test circuit for quantum distance computation + k-means clustering

MCP Server: 7 → 10 Tools

• monte_carlo_price: European option pricing via quantum amplitude estimation
• qaoa_optimize: Combinatorial optimization (max-cut, max-ones, min-ones)
• cluster_data: Data clustering via quantum swap test distances

Bug Fixes

• Shor numpy.int64 overflow fixed (works >32K now)
• Grover MCP num_qubits → num_bits parameter fix
• Shor factor_recursive() for full prime decomposition

Stats

• 445 tests passed, 0 lint errors
• 10 MCP tools, 3 new algorithms

v0.6.1 — Structural Fixes & Encapsulation

What's Changed in v0.6.1

Architecture & Security

• NoiseModel integration: run(circ, noise=NoiseModel()) — noise is now a first-class citizen
• Grover encapsulation: sim._state → public apply_phase() + state setter
• Shor QFT via DAG: Real H/RZ/SWAP gates through DAG pipeline (modular exp stays classical)
• Surface code: Stabilizer-based syndrome extraction + deterministic BFS logical error check
• MCP server: exec() sandboxed with restricted __builtins__
• QASM import: eval() replaced with safe arithmetic parser

Encapsulation

• All sim._state access eliminated — public API only (state, apply_phase, apply_noise)
• StateVectorSimulator.apply_noise() public method added

Code Quality

• 0 ruff lint errors (quanta/ + tests/)
• 78 test lint fixes (unsorted/unused imports)
• Turkish comments → English across 12+ files
• MCP tests skip gracefully when fastmcp not installed

Bug Fixes

• color_code.py: RNG reproducibility (rng parameter)
• sdg/tdg gate mappings corrected
• BitFlipCode/PhaseFlipCode distance → 3
• list.pop(0) → deque.popleft() in DAG + surface code BFS
• CircuitSpec removed from __all__
• CY gate added to exports

Tests: 445 passed, 12 skipped

Quanta SDK v0.6.0 — QEC Enhancements

What's New

Color Code (qec/color_code.py)

• Triangular lattice, 3-colorable plaquettes (R, G, B)
• Restriction decoder (inspired by Google's chromobius)
• Transversal Clifford gates (H, S, CX) — advantage over surface code
• ColorCode(distance=3) -> [[7,1,3]], d=5 -> [[19,1,5]], d=7 -> [[37,1,7]]

QEC Decoders (qec/decoder.py)

• MWPMDecoder: Minimum Weight Perfect Matching (greedy, O(n^2))
• UnionFindDecoder: Near-linear O(n·α(n)) cluster-based decoder
• Works with both surface code and color code

Pauli Frame Simulator (simulator/pauli_frame.py)

• Aaronson-Gottesman stabilizer tableau
• O(n) per gate, O(n^2) memory
• 50-qubit GHZ circuit in <5 seconds
• Clifford gates: H, S, X, Y, Z, CX, CZ, SWAP
• Error injection for noise testing

Stats

• 216 tests, all pass
• Google quantumlib equivalent: Stim + chromobius + tesseract-decoder

Install

pip install --upgrade quanta-sdk

Quanta SDK v0.5.0 — Hardware Backends

What's New

IBM Quantum Backend

• QASM 2.0 bridge to Qiskit
• Local Aer simulator + real hardware via SamplerV2
• run(bell, backend=IBMBackend('ibm_brisbane'))

IonQ Backend

• Pure REST API — zero external dependencies
• Native JSON gate format with job polling
• run(bell, backend=IonQBackend('qpu.aria-1'))

Runner Enhancement

• run() now accepts backend= parameter
• Default: built-in statevector simulator (unchanged)

Stats

• 171 tests, all pass
• 4 hardware backends: Local, Google, IBM, IonQ

Install

pip install --upgrade quanta-sdk

Quanta SDK v0.4.0

Multi-paradigm quantum computing SDK for Python.

Install: pip install quanta-sdk
Import: import quanta

150+ tests | 11 examples | 9 algorithms | Apache 2.0