TinyQubit

Minimal, deterministic quantum circuit compiler with explainable routing.

Why?

Existing quantum SDKs have problems:

• API Churn — Qiskit breaks everything between versions
• Non-Deterministic — Same circuit, different output every run
• Black-Box — No explanation of routing decisions
• Bloat — 1+ second imports, heavy dependencies
• Vendor Lock-in — Best features tied to one provider

TinyQubit: tiny, no dependencies, deterministic, explainable.

Usage

from tinyqubit import Circuit, to_openqasm2

c = Circuit(2).h(0).cx(0, 1).measure(0).measure(1)
print(to_openqasm2(c))

Gates

16 primitives: X Y Z H S T SDG TDG RX RY RZ CX CZ CP SWAP MEASURE

Dependencies

Dependency	Used for	Plan
numpy	Simulator (statevector math)	Replace with pure Python or make optional

Dev/test:

• pytest — test framework
• hypothesis — property-based tests

Optional (for hardware submission):

• qiskit — to_qiskit() export
• qiskit-ibm-runtime — IBM hardware submission
• amazon-braket-sdk — AWS Braket submission

Goal: Core compiler (circuit → routed → decomposed → QASM) runs with zero dependencies.

Tests & Commands

# Run examples
python examples/circuit.py       # Basic circuit + transpile
python examples/grover.py        # Grover's algorithm
python examples/submit_to_ibm.py # IBM hardware submission (requires qiskit-ibm-runtime)

# Run all tests
python -m pytest tests/ -v

# Run specific test categories
python -m pytest tests/test_determinism.py -v   # Golden/determinism tests
python -m pytest tests/test_properties.py -v    # Property-based tests (hypothesis)
python -m pytest tests/test_metrics.py -v       # Routing/optimization metrics
python -m pytest tests/test_performance.py -v   # Import time, transpile speed

# Run benchmarks (vs Qiskit)
python benchmarks/run_all.py

# Update golden test baselines (after intentional changes)
python scripts/update_golden.py