feat: wit for arbit qubits by vprusso · Pull Request #615 · unitaryfoundation/metriq-gym

vprusso · 2025-11-01T21:11:16Z

Closes: #609

Additionally, this MR fixes the following warnings:

=========================================== warnings summary ===========================================
tests/unit/benchmarks/test_mirror_circuits.py::TestMirrorCircuitGeneration::test_generate_mirror_circuit
tests/unit/benchmarks/test_mirror_circuits.py::TestMirrorCircuitGeneration::test_generate_mirror_circuit
tests/unit/benchmarks/test_mirror_circuits.py::TestMirrorCircuitGeneration::test_generate_mirror_circuit
tests/unit/benchmarks/test_mirror_circuits.py::TestMirrorCircuitGeneration::test_generate_mirror_circuit
tests/unit/benchmarks/test_mirror_circuits.py::TestMirrorCircuitGeneration::test_generate_mirror_circuit
tests/unit/benchmarks/test_mirror_circuits.py::TestMirrorCircuitGeneration::test_generate_mirror_circuit
tests/unit/benchmarks/test_mirror_circuits.py::TestMirrorCircuitGeneration::test_generate_mirror_circuit
  /home/vrusso/Projects/metriq-gym/metriq_gym/benchmarks/mirror_circuits.py:335: DeprecationWarning: Treating CircuitInstruction as an iterable is deprecated legacy behavior since Qiskit 1.2, and will be removed in Qiskit 3.0. Instead, use the `operation`, `qubits` and `clbits` named attributes.
    for instr, qargs, _ in pauli_layer.data:

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html

Copilot

Pull Request Overview

This PR generalizes the Wormhole Inspired Teleportation (WIT) benchmark to support arbitrary qubit counts and configuration parameters, addressing issue #609. The implementation replaces hard-coded 6- and 7-qubit circuits with a configurable approach while maintaining backward compatibility through legacy support.

Key changes:

Introduced WormholeTeleportationConfig dataclass and WormholeTeleporterFactory for parameterized circuit generation
Extended JSON schema to include physics parameters (coupling strength, rotation angles, time steps, etc.)
Added comprehensive test coverage for the new generalized circuit builder and parameter validation

Reviewed Changes

Copilot reviewed 12 out of 12 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
metriq_gym/benchmarks/wit.py	Core refactoring: replaced fixed circuits with configurable builder; added config validation and parameter parsing
metriq_gym/schemas/wit.schema.json	Extended schema with new physics parameters; made num_qubits optional for backward compatibility
tests/unit/benchmarks/test_wit.py	Added tests verifying generalized circuits match legacy behavior and validate config constraints
metriq_gym/schemas/examples/wit.example.json	Updated example to demonstrate new parameterization with 7-qubit swap-method circuit
tests/conftest.py	Updated fixture params to include all required WIT configuration fields
tests/unit/quantinuum/test_run_fetch_result_quantinuum.py	Updated test params to match new schema requirements
metriq_gym/suites/uf_frugal_3.json	Migrated suite config to new parameter format
docs/source/quickstart.rst	Added note explaining new parameterization approach
docs/source/end_to_end_tutorial.ipynb	Updated notebook output reflecting new param structure
docs/source/cli_workflows.rst	Updated suite example with expanded WIT parameters
README.md	Added brief explanation of new WIT configuration flexibility
metriq_gym/benchmarks/mirror_circuits.py	Minor formatting improvements (unrelated to WIT changes)

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2025-11-03T12:29:26Z

metriq_gym/benchmarks/wit.py

+        z_rotation_angles_raw = cast(Sequence[SupportsFloat], raw_fields["z_rotation_angles"])
+        z_rotation_angles = tuple(float(v) for v in z_rotation_angles_raw)
+        time_steps = int(cast(SupportsInt, raw_fields["time_steps"]))
+        insert_method_str = str(raw_fields["insert_message_method"]).lower()


The .lower() call may fail if insert_message_method is not a string. Consider validating the type before calling .lower() or handling potential AttributeError.

cosenal · 2025-11-03T15:02:40Z

Did you run it for n_qubits = 6 and 7, and cross-check the results against the ones we obtained with the old non-general version?

vprusso · 2025-11-03T16:13:51Z

Did you run it for n_qubits = 6 and 7, and cross-check the results against the ones we obtained with the old non-general version?

Yes. The QASM for the circuits obtained for the 6 and 7-qubit circuits are identical to the QASM used with the parameterized circuits for 6 and 7-qubits. Running using the parameterized version on the IBM Kingston device, we have that:

6-qubits:

Using the schema parameterization:

{
    "benchmark_name": "WIT",
    "shots": 8192,
    "n_qubits_per_side": 3,
    "message_size": 1,
    "insert_message_method": "reset",
    "interaction_coupling_strength": 1.5707963267948966,
    "x_rotation_transverse_angle": 0.7853981633974483,
    "zz_rotation_angle": 0.7853981633974483,
    "z_rotation_angles": [0.0283397, 0.00519953, 0.0316079],
    "time_steps": 3,
    "num_qubits": 6
}

Polling job...
{'app_version': '0.4.2.dev19+gfd416d3b7',
 'job_type': 'WIT',
 'platform': {'device': 'ibm_kingston',
              'device_metadata': {'num_qubits': 156,
                                  'simulator': False,
                                  'version': '1.0.0'},
              'provider': 'ibm'},
 'suite_id': None,
 'timestamp': '2025-11-03T11:02:19.016473'}

Results:
  expectation_value: 0.8526611328125 ± 0.0039160833325412965

7-qubits:

Using the schema parameterization:

{
    "benchmark_name": "WIT",
    "shots": 8192,
    "n_qubits_per_side": 3,
    "message_size": 1,
    "insert_message_method": "swap",
    "interaction_coupling_strength": 1.5707963267948966,
    "x_rotation_transverse_angle": 0.7853981633974483,
    "zz_rotation_angle": 0.7853981633974483,
    "z_rotation_angles": [0.0283397, 0.00519953, 0.0316079],
    "time_steps": 3,
    "num_qubits": 7
}

Polling job...
{'app_version': '0.4.2.dev19+gfd416d3b7',
 'job_type': 'WIT',
 'platform': {'device': 'ibm_kingston',
              'device_metadata': {'num_qubits': 156,
                                  'simulator': False,
                                  'version': '1.0.0'},
              'provider': 'ibm'},
 'suite_id': None,
 'timestamp': '2025-11-03T11:01:19.417911'}

Results:
  expectation_value: 0.8983154296875 ± 0.0033392350226119256

cosenal

Some preliminary comments

cosenal · 2025-11-06T14:25:20Z

docs/source/cli_workflows.rst

+          "n_qubits_per_side": 3,
+          "message_size": 1,
+          "insert_message_method": "swap",
+          "interaction_coupling_strength": 1.5707963267948966,
+          "x_rotation_transverse_angle": 0.7853981633974483,
+          "zz_rotation_angle": 0.7853981633974483,
+          "z_rotation_angles": [
+            0.0283397,
+            0.00519953,
+            0.0316079
+          ],
+          "time_steps": 3,


Do we really want to give this much freedom in the parameters? Can't we simplify and have defaults for all parameters except the num of qubits?

Likely not. I think we can keep this level of specification strictly to the benchmark logic itself and not necessarily expose it to the user. Done.

cosenal · 2025-11-06T14:27:15Z

metriq_gym/benchmarks/mirror_circuits.py

 class MirrorCircuitsResult(BenchmarkResult):
-    success_probability: BenchmarkScore = Field(..., json_schema_extra={"direction": MetricDirection.HIGHER})
-    polarization: BenchmarkScore = Field(..., json_schema_extra={"direction": MetricDirection.HIGHER})
+    success_probability: BenchmarkScore = Field(


nit: was this file committed before we had a linter and a formatter, or why does it need reformatting now?

cosenal · 2025-11-06T14:29:25Z

metriq_gym/schemas/wit.schema.json

+        "interaction_coupling_strength": {
+            "type": "number",
+            "description": "Interaction coupling strength (g) applied during the two-sided coupling step.",
+            "default": 1.5707963267948966,


I am guessing those values are fractions of pi? It's odd that one can't use symbols in here. One more reason to get rid of those parameters.

Good point. Those have since been removed (see the above comment/response).

cosenal · 2025-11-06T14:33:46Z

metriq_gym/benchmarks/wit.py

+    def __post_init__(self) -> None:
+        if self.n_qubits_per_side < 1:
+            raise ValueError("n_qubits_per_side must be at least 1.")
+        if self.message_size < 1:
+            raise ValueError("message_size must be at least 1.")
+        if self.message_size >= self.n_qubits_per_side:
+            raise ValueError("message_size must be smaller than n_qubits_per_side.")
+        if self.time_steps < 1:
+            raise ValueError("time_steps must be at least 1.")
+        if self.insert_message_method not in VALID_MESSAGE_METHODS:
+            raise ValueError(
+                f"insert_message_method must be one of {', '.join(VALID_MESSAGE_METHODS)}."
+            )


Aren't those constraints validated as part of the json schema validation?

vprusso added 2 commits November 1, 2025 17:10

feat: wit for arbit qubits

765e97f

fix: update notebook output

fd416d3

vprusso marked this pull request as ready for review November 3, 2025 12:28

vprusso requested review from Copilot and cosenal November 3, 2025 12:28

Copilot AI reviewed Nov 3, 2025

View reviewed changes

vprusso mentioned this pull request Nov 3, 2025

Enhancement: Generalize number of qubits for the WIT benchmark #609

Open

cosenal reviewed Nov 6, 2025

View reviewed changes

vprusso added 4 commits November 7, 2025 17:30

chore: merge main

1d28563

chore: merge main

a3fae83

chore: simplify wit benchmark

15334b5

chore: simplify wit benchmark

f194a01

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat: wit for arbit qubits#615

feat: wit for arbit qubits#615
vprusso wants to merge 6 commits intomainfrom
609-generalize-wit-benchmark

vprusso commented Nov 1, 2025 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Nov 3, 2025

Uh oh!

cosenal commented Nov 3, 2025

Uh oh!

vprusso commented Nov 3, 2025 •

edited

Loading

Uh oh!

cosenal left a comment •

edited

Loading

Uh oh!

cosenal Nov 6, 2025

Uh oh!

vprusso Nov 7, 2025

Uh oh!

cosenal Nov 6, 2025

Uh oh!

cosenal Nov 6, 2025

Uh oh!

vprusso Nov 7, 2025

Uh oh!

cosenal Nov 6, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

vprusso commented Nov 1, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Reviewed Changes

Uh oh!

Copilot AI Nov 3, 2025

Choose a reason for hiding this comment

Uh oh!

cosenal commented Nov 3, 2025

Uh oh!

vprusso commented Nov 3, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

cosenal left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

cosenal Nov 6, 2025

Choose a reason for hiding this comment

Uh oh!

vprusso Nov 7, 2025

Choose a reason for hiding this comment

Uh oh!

cosenal Nov 6, 2025

Choose a reason for hiding this comment

Uh oh!

cosenal Nov 6, 2025

Choose a reason for hiding this comment

Uh oh!

vprusso Nov 7, 2025

Choose a reason for hiding this comment

Uh oh!

cosenal Nov 6, 2025

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

vprusso commented Nov 1, 2025 •

edited

Loading

vprusso commented Nov 3, 2025 •

edited

Loading

cosenal left a comment •

edited

Loading