✨ Add native RCCX (relative-phase Toffoli) gate support

[simon1hofmann]

Description

Adds native RCCX (relative-phase Toffoli) support across MQT Core.

Gate convention (Qiskit / OpenQASM): rccx(q0, q1, q2) — q0 and q1 are the inner controls, q2 is the target. Controlled variants follow the usual pattern: crccx(c, q0, q1, q2) and mcrccx({c…}, q0, q1, q2).

RCCX is self-inverse and differs from CCX by relative phases on the computational basis.

IR & Python

• New OpType::RCCX with rccx / crccx / mcrccx on QuantumComputation
• Python bindings and regenerated stubs
• StandardOperation::invert() treats RCCX as self-inverse
• OpenQASM 3 import/export and I/O round-trip tests

MLIR (QC / QCO / QIR)

• RCCXOp in QC, QCO, and QIR dialects with closed-form 8×8 unitary
• Program builders: rccx, crccx, mcrccx
• Translation and conversion coverage (QC ↔ QCO ↔ QIR ↔ Jeff)
• QASM3 and QuantumComputation translation tests
• Compiler pipeline and round-trip tests
• Measurement fixtures added to RCCX program tests (aligned with DCX and similar gates)

Decision diagrams (DD)

• getRccxDD builds the gate via canonical elementary decomposition
• Outer controls merged for crccx / mcrccx; single-qubit steps on the target only
• RCCX listed as self-inverse in DD inverse handling
• Package and functionality tests

ZX functionality construction

• addRccx — bare RCCX via addZSpider / addCnot
• addCrccx — single outer control via addCphase, addMcphase, addMcx
• addMcrccx — multi-control via addMcrz / addMcrx / addMcz / addMcphase / addMcx (delegates to addCrccx for one control)
• parseOp dispatches by control count, consistent with other gates

Canonical decomposition (matches decompose-multi-controlled-gates / emitRCCXSequence):
H(t) → T(t) → CX(c1,t) → Tdg(t) → CX(c0,t) → T(t) → CX(c1,t) → Tdg(t) → H(t)

Qiskit plugin & QDMI

• Qiskit ↔ MQT conversion for rccx / crccx / mcrccx
• Backend and mock-backend tests
• RCCX listed in QDMI DDSIM device gate set

AI Assistance

Used Composer 2.5 via Cursor for parts of this change. I reviewed the full
diff and take responsibility for everything in this PR.

Checklist

• The pull request only contains commits that are focused and relevant to this change.
• I have added appropriate tests that cover the new/changed functionality.
• I have updated the documentation to reflect these changes.
• I have added entries to the changelog for any noteworthy additions, changes, fixes, or removals.
• I have added migration instructions to the upgrade guide (if needed).
• The changes follow the project's style guidelines and introduce no new warnings.
• The changes are fully tested and pass the CI checks.
• I have reviewed my own code changes.

If PR contains AI-assisted content:

• I have disclosed the use of AI tools in the PR description as per our AI Usage Guidelines.
• AI-assisted commits include an Assisted-by: [Model Name] via [Tool Name] footer.
• I confirm that I have personally reviewed and understood all AI-generated content, and accept full responsibility for it.

[simon1hofmann]

What is still missing and up to debate:

• DD package simulation (getRccxDD)
• ZX functionality construction
• QDMI DDSIM backend listing
• StandardOperation::invert() for RCCX

[coderabbitai]

Actionable comments posted: 2

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@python/mqt/core/plugins/qiskit/qiskit_to_mqt.py`:
- Around line 321-324: Update the rccx handling in the conversion function to
pass the already computed qubits list to StandardOperation instead of
recomputing the qubit_map comprehension. Preserve the existing OpType.rccx
operation and return behavior.

In `@test/ir/test_io.cpp`:
- Around line 80-86: Add the direct <string_view> standard header include to
test_io.cpp for the std::string_view parameters in expectRccxExport, removing
reliance on transitive includes.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: ASSERTIVE

Plan: Pro

Run ID: 6b41173f-e440-43b8-88a9-0da83fe59c08

📥 Commits

Reviewing files that changed from the base of the PR and between f2b6a38 and c44cfad.

📒 Files selected for processing (25)

• docs/qdmi/qdmi_backend.md
• include/mqt-core/ir/QuantumComputation.hpp
• include/mqt-core/ir/operations/OpType.hpp
• include/mqt-core/ir/operations/StandardOperation.hpp
• include/mqt-core/qasm3/StdGates.hpp
• include/mqt-core/zx/FunctionalityConstruction.hpp
• mlir/include/mlir/Dialect/QCO/IR/QCOOps.td
• mlir/include/mlir/Dialect/QCO/QCOUtils.h
• mlir/lib/Conversion/JeffToQCO/JeffToQCO.cpp
• mlir/lib/Conversion/QCOToJeff/QCOToJeff.cpp
• mlir/lib/Dialect/QCO/IR/Operations/StandardGates/RCCXOp.cpp
• mlir/lib/Dialect/QIR/Builder/QIRProgramBuilder.cpp
• mlir/unittests/Compiler/test_compiler_pipeline.cpp
• mlir/unittests/Conversion/JeffRoundTrip/test_jeff_round_trip.cpp
• mlir/unittests/Dialect/QCO/IR/test_qco_ir.cpp
• mlir/unittests/programs/qco_programs.cpp
• mlir/unittests/programs/qco_programs.h
• python/mqt/core/ir/__init__.pyi
• python/mqt/core/ir/operations.pyi
• python/mqt/core/plugins/qiskit/mqt_to_qiskit.py
• python/mqt/core/plugins/qiskit/qiskit_to_mqt.py
• src/ir/operations/StandardOperation.cpp
• src/qdmi/devices/dd/Device.cpp
• test/ir/test_io.cpp
• test/python/plugins/test_qiskit.py

[simon1hofmann]

@coderabbitai review

[coderabbitai]

✅ Action performed

Review finished.

Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.

[simon1hofmann]

Did not anticipate that many changes to add another gate, @denialhaag would it be possible to get a review?

[burgholzer]

Did not anticipate that many changes to add another gate

Hahah. Yeah, it's quite wild how far reaching adding a single native gate is 😅

[denialhaag]

Thanks a lot for going through all of this, @simon1hofmann! 🙂

This already looks good to me. I just have some minor nitpicks, which you can find below. Don't be surprised by the number of comments; 16 of them are the same comment.

[simon1hofmann]

@coderabbitai review

[coderabbitai]

✅ Action performed

Review finished.

Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.

[denialhaag]

Thanks for addressing my comments! This LGTM now! 🙂

We should also add a changelog entry for the non-MLIR changes. I'm also wondering if we want to backport this. If we do so, I would not update MQTOpt and MQTRef, though. 🤔

I'll leave the final say to @burgholzer! 😌

[simon1hofmann]

Thanks for addressing my comments! This LGTM now! 🙂

We should also add a changelog entry for the non-MLIR changes. I'm also wondering if we want to backport this. If we do so, I would not update MQTOpt and MQTRef, though. 🤔

I'll leave the final say to @burgholzer! 😌

Thanks a lot for the review @denialhaag 🙏
I pushed two more commits to further simplify and align the code (handling R3CX as a controlled RCCX was not correct).

[denialhaag]

I pushed two more commits to further simplify and align the code (handling R3CX as a controlled RCCX was not correct).

Ahh, good catch! The commits look good to me! 👍