Circuits¶

The divi.circuits module provides circuit abstractions for quantum program generation, execution, and error mitigation.

Warning

Developer-Facing Classes: The core circuit class MetaCircuit is intended for advanced users and developers. Most users should interact with circuits through higher-level APIs in the divi.qprog module.

Core¶

Functions¶

`build_overlap_meta`(cost_circuit)	Build the compute-uncompute overlap circuit from a cost ansatz.
`build_template`(qasm_body, symbol_names)	Split a QASM string at symbol boundaries into a `QASMTemplate`.
`dag_to_qasm_body`(dag[, precision])	Emit a body-only parametric OpenQASM 2.0 string from a DAG.
`measurement_qasms_from_groups`(...[, measure_all])	Emit body-only measurement QASM per commuting observable group.
`qnode_to_meta`(qnode, *[, arg_shapes, precision])	Convert a `QNode` directly to a `MetaCircuit`.
`qscript_to_meta`(qscript[, precision, ...])	Shared helper: convert a PennyLane `QuantumScript` to a `MetaCircuit`.
`render_template`(template, formatted_values)	Render a `QASMTemplate` with concrete parameter values.

Classes¶

`MetaCircuit`(circuit_bodies[, parameters, ...])	Logical circuit IR.
`QASMTemplate`(fragments, slot_indices)	Pre-split QASM body for fast parameter substitution.
`TemplateEntry`(template_qasm, ...)	One parametric circuit ready for backend-side substitution.

Variables¶

`DEFAULT_PRECISION`	int([x]) -> integer int(x, base=10) -> integer
`AxisLabel`	Built-in immutable sequence.
`QASMTag`	Built-in immutable sequence.

Error Mitigation Protocols¶

Divi provides quantum error mitigation (QEM) capabilities to improve the accuracy of quantum computations in the presence of noise. All protocols inherit from QEMProtocol.

Classes¶

`QEMProtocol`()	Abstract base class for Quantum Error Mitigation protocols.
`_NoMitigation`()	A dummy default mitigation protocol — pass the circuit through.

Zero Noise Extrapolation (ZNE)¶

Zero-noise extrapolation protocols and folding helpers.

Functions¶

`global_fold`(dag, scale)	Apply `GlobalFoldPass` and return `(folded_dag, effective_scale)`.
`local_fold`(dag, scale, *[, selection, ...])	Apply `LocalFoldPass` and return `(folded_dag, effective_scale)`.

Classes¶

`ZNE`(scale_factors[, folding_fn, extrapolator])	Zero Noise Extrapolation.
`ZNEExtrapolator`(args, *kwargs)	Structural type for zero-noise extrapolation.
`LinearExtrapolator`()	Fit a line `y = a + b·s` and return `a` (the intercept at s=0).
`RichardsonExtrapolator`()	Richardson (Lagrange) extrapolation through all `N` points to s=0.
`GlobalFoldPass`(scale_factor)	Global unitary folding with fractional scale-factor support.
`LocalFoldPass`(scale_factor[, selection, ...])	Per-gate folding with fractional scale-factor support.

Variables¶

FoldingFn

Type for the folding callable — given a DAGCircuit and a requested scale_factor, return (folded_dag, effective_scale).

Quantum Enhanced Pauli Propagation (QuEPP)¶

Quantum Enhanced Pauli Propagation (QuEPP) error mitigation protocol.

Implements the hybrid classical-quantum error mitigation scheme from Majumder et al. (arXiv:2603.14485). QuEPP decomposes a quantum circuit into alternating Clifford layers and non-Clifford Pauli rotations via Clifford Perturbation Theory (CPT). Low-order Pauli paths are simulated classically; the residual is corrected using noisy quantum execution with an empirical rescaling factor.

The decomposition uses the paper’s Heisenberg-picture back-propagation with weights cos(θ)/sin(θ) and R_P(π/2) Clifford replacements. The observable is tracked as a Pauli string; gates that commute with the current observable contribute weight 1 (no branching), which can dramatically reduce path count.

Classes¶

QuEPP([truncation_order, ...])

Quantum Enhanced Pauli Propagation (QuEPP) error mitigation.