QDSV Bridge

Source/package version: 0.6.5. See the PyPI badge for publication status.

From Business Meaning To Quantum Artifacts

QDSV Bridge transforms bounded business rules, prepared data and functional objectives into semantically validated and optimized logical quantum artifacts. It is QDSV's interoperability layer for today's circuit-based quantum ecosystem: the semantic problem remains primary, and logical circuits are derived when the destination requires a circuit artifact.

QDSV Bridge is domain-agnostic within its certified semantic operation set. It does not choose a circuit from a fixed catalog of industry templates. It compiles bounded semantic programs assembled from composable numeric and logical operations, including general predicates and ScoreModel v2. Users do not need to design gates, quantum registers, ancillas, reversible cleanup or a framework-specific quantum model. The user describes what the problem means; Bridge derives and verifies its quantum realization.

The circuit is a delivery format, not the source of truth.

The semantic problem specification remains canonical while Bridge derives, validates and documents:

• the supported reversible construction, registers, controls and cleanup;
• an immutable canonical logical circuit;
• an optional exact logical optimization linked to its parent by digest;
• semantic-equivalence and no-regression validation;
• inspectable OpenQASM/Qiskit artifacts, resources and construction evidence;
• a traceable handoff for downstream simulator or hardware execution.

The user provides bounded candidates, prepared numeric inputs, functional rules, thresholds, priorities and objectives. Bridge derives the supported quantum construction without requiring the user to design it. It never fabricates an unsupported construction or embeds precomputed answers when a request cannot be materialized.

• Documentation
• PyPI
• Source

Business-First Quickstart

Start with the candidates and the rule they must satisfy. This example asks Bridge to mark suppliers whose quality is at least 700 and whose compliance flag is 1:

from qdsv_bridge import (
    QDSVBridgeClient,
    build_predicate_spec,
    select_recommended_artifact,
)

suppliers = [
    {"candidate_index": 0, "supplier_id": 101, "quality": 820, "compliance": 1},
    {"candidate_index": 1, "supplier_id": 102, "quality": 680, "compliance": 1},
    {"candidate_index": 2, "supplier_id": 103, "quality": 760, "compliance": 0},
]

business_rule = {
    "op": "and",
    "args": [
        {
            "op": "gte",
            "left": {"op": "field", "name": "quality"},
            "right": {"op": "const", "value": 700},
        },
        {
            "op": "eq",
            "left": {"op": "field", "name": "compliance"},
            "right": {"op": "const", "value": 1},
        },
    ],
}

spec = build_predicate_spec(rows=suppliers, predicate=business_rule)
result = QDSVBridgeClient().generate(spec)
recommended = select_recommended_artifact(result)

print(result["status"])
print(result["recommended_artifact_role"])
print(recommended["format"])
print(result["construction_verification"])

The user does not provide gates, registers, ancillas, reversible cleanup or expected answers. candidate_index is the stable circuit-domain identity; supplier_id remains the organization's business reference. build_predicate_spec() preserves both, normalizes the declared rule without evaluating it, and applies portable defaults for the logical artifact. Advanced users can override those defaults through its typed parameters.

The successful response can contain:

• an immutable canonical logical artifact;
• an accepted optimized child artifact when exact validation and no-regression checks pass;
• the recommended artifact role and inspectable OpenQASM/Qiskit content;
• construction verification, resource metrics, digests and traceability;
• a handoff contract for downstream simulation or hardware execution.

The complete runnable version is examples/compound_business_predicate.py.

How Bridge Differs From Quantum Synthesis Platforms

Most quantum-development platforms reduce the work required to construct a circuit but still expect the user to define or review a quantum program. Bridge starts one layer earlier: with the bounded problem and its functional meaning.

The distinction is not merely fewer lines of code. It is a different user responsibility:

Classiq provides a high-level environment for designing and synthesizing quantum programs. For bounded semantic programs expressible through Bridge's certified composable operation graph, QDSV Bridge removes quantum-program design from the end-user workflow.

Classiq provides a broader integrated quantum-development environment. Bridge provides an earlier entry point for organizations that want to begin with the problem rather than with a quantum program. The comparison above describes the public workflows and architectural responsibility boundary; it is not a claim of universal circuit-performance superiority. See the Classiq documentation for its current public workflow.

Installation

Install the client:

pip install qdsv-bridge

Install the optional Qiskit inspection dependencies:

pip install "qdsv-bridge[qiskit]"

The Qiskit extra is capped at qiskit>=2,<3 to preserve compatibility with the currently tested Qiskit major version. The Public Preview does not require an API key.

Current Availability And Operational Boundaries

QDSV Bridge is publicly available through PyPI and the Qiskit Ecosystem for bounded real-world validation. Its certified semantic operations and deployment options continue to expand under the Public Preview contract.

• Bridge supports bounded compositions of certified semantic operations, not a fixed catalog of industries or use cases. Public paths include general predicates and ScoreModel v2.
• Practical limits are determined by certified reversible lowering and materialized resource limits, not by a fixed catalog of industries or use cases. Relevant resources include candidate count, numeric precision, expression complexity, qubits, ancillas, circuit depth and artifact size. Bridge does not accept every arbitrary business or quantum program.
• The public service is provided without an SLA and may change or be temporarily unavailable before 1.0.
• Bridge does not execute on a simulator or QPU, select a provider, manage credentials, route to hardware, mitigate noise or interpret experiments.
• Do not send confidential, regulated or secret data to the public service.

Bridge validates the semantic-to-circuit construction path and reports the resources required by the generated artifact. It can derive an exact, target-independent logical optimization and recommend it only when contractual replay passes and protected logical metrics do not regress. The canonical artifact is never replaced silently.

The public SDK supports Python 3.9 and later. Before 1.0, minor releases may introduce contract changes; deprecations and migration notes are recorded in the changelog.

Technical Workflows

Use build_predicate_spec() for explicit public predicates with nested boolean composition and field-to-field comparisons. Use build_score_expression_spec() when a numeric expression must be compared with a declared threshold, and build_score_model_spec() for flat or hierarchical ScoreModel v2 decisions. These helpers normalize public semantic inputs without calculating scores, decisions or expected answers.

weighted_sum accepts explicit values and same-length weights. Unary operations accept arg, value, operand or a one-item args list and are canonicalized to one stable operand form. Every request remains bounded and resource checked; a valid semantic program can still be rejected if its materialized artifact exceeds the active limits.

Use the lower-level specification contract only when you need to freeze artifact format, backend family, evidence or resource limits. In every path, do not include labels, expected decisions or precomputed predicate results in the input rows.

When materialization succeeds, generate() returns the canonical circuit and loading guidance. Bridge attempts the public qiskit_structural_exact_v1 logical optimization by default. select_recommended_artifact() returns the accepted optimized artifact when available and otherwise returns the canonical artifact.

The optimization is exact and target-independent. It does not perform layout, routing, scheduling, calibration-aware selection, noise suppression, mitigation or hardware execution. Those operations remain downstream in Qiskit, Qruba or QDSV Runtime/HSP.

For a minimal multi-criteria ScoreModel example, run examples/score_model_v2.py. Prepared metrics and the cutoff use one declared scale, and priority represents a domain priority, not the position of a criterion. The SDK example does not reproduce private ScoreModel aggregation or compiler rules.

Technical Reference

The detailed contracts remain versioned in the documentation:

• Public SDK contract: delivery modes, outputs, errors, limits, privacy, reports and compatibility.
• Conformance and release identity: normative outcomes, validated runtime identities, suite evidence and reproducibility boundaries.
• Migration to 0.6.5: compatibility, contract and evidence changes from 0.6.4.
• Problem-first and similarity boundaries: prepared metrics, declared operations and construction guarantees.
• Canonical and optimized artifacts: profiles, lineage, validation and recommendation.
• IBM/Qiskit handoff: the boundary between logical artifacts and physical execution.
• Examples and tutorials: first workflow, ScoreModel and inspectable artifacts.

Installed clients can inspect the exact public release identity without contacting the service:

from qdsv_bridge import get_release_manifest

manifest = get_release_manifest()
print(manifest["contracts"])
print(manifest["conformance"])

The manifest identifies the contract and the runtime builds used for release validation. It is evidence about those frozen builds, not a claim that a mutable public endpoint will always serve the same deployment.

Support And Security

• General questions and defects: GitHub Issues
• Sensitive security reports: follow SECURITY.md; do not open a public issue
• Release history: CHANGELOG.md
• Roadmap: ROADMAP.md

License

The client SDK, examples, documentation and tests in this repository are licensed under the MIT License.

QDSV, QIntent and Qruba names and marks belong to their respective owners. The MIT License does not grant trademark rights.