Coupling Analysis¶
The coupling-analysis surface exposes the part of the RQM stack that looks beyond single-qubit canonicalization and asks whether cross-qubit structure is present, preserved, or measurably supported.
Base URL: https://rqm-api.onrender.com
Current endpoints¶
POST /v1/analysis/couplingPOST /v1/analysis/coupling/compare
Why this surface exists¶
rqm-core owns coupling / entanglement analysis primitives as part of the mathematical spine. That matters because optimization trust is not only about gate count reduction; it is also about whether meaningful structure was preserved.
This surface helps answer questions such as:
- does this circuit contain entangling structure?
- is the analysis measured or qualitative?
- did an optimized candidate preserve the relevant structure?
Qualitative vs measured¶
The docs should keep these modes separate.
Qualitative mode¶
Qualitative analysis identifies entangling structure and compatibility signals without claiming a measured entanglement metric outside supported scope.
Measured mode¶
Measured analysis uses the currently supported measured scope. It should be documented honestly as limited, not as a generic unrestricted entanglement engine.
Honest scope limitations¶
The current platform should not be described as providing unrestricted measured entanglement analysis for arbitrary circuits.
Instead:
- measured coverage exists where the underlying analysis primitives support it
- unsupported cases fall back to qualitative signals
- limitations are part of the contract, not hidden implementation detail
Why compare/preservation matters¶
POST /v1/analysis/coupling/compare is the natural endpoint for before/after trust checks around optimization. It connects directly to RQM's compiler story:
- optimization candidates are verified before commit
- coupling/preservation analysis offers an additional trust surface for users and workflows
- Studio and reports can use this to explain what changed without blurring it with internal IR