(fix) VM resolution strategy correction

[gmulhearn]

The current VM resolution strategy will get potential VMs by getting the set VMs of that proof_purpose (e.g. "assertionMethod", "authentication") from the pydid formatted DIDDocument. It does this with a getattr (e.g. getattr(doc, "assertionMethod", []).

The problem is that the class attributes of the DIDDocument class are snake case (assertion_method), so attempts to use the proof purpose assertionMethod result in an empty result when you try to get the assertionMethod attribute.

I've just changed this to .get the attribute from the original raw doc dict instead, which has the attributes/keys in the camelCase format.

The new test is a regression test - it fails before the fix

[gmulhearn]

Discovered this while testing W3C VCDM issuance (over DIDComm) with a cheqd DID.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
100.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[swcurran]

Thanks! Can you please take a look at the Ruff/lint failure?

I’ve asked @ff137 and @esune to review this.

[ff137]

Makes sense 👍

---

I think that the camelCase vs snake_case headaches should be avoided by aliasing the Pydantic fields, but that requires changes in pydid
@dbluhm

e.g.

class DIDDocumentRoot(Resource):
    assertion_method: Optional[List[Union[DIDUrl, VerificationMethod]]] = None

could be:

class DIDDocumentRoot(Resource):
    assertion_method: Optional[List[Union[DIDUrl, VerificationMethod]]] = Field(None, alias="assertionMethod")

---

Looks like Pydantic has a way to generate aliases automatically: https://docs.pydantic.dev/2.1/usage/model_config/#alias-generator

I'll post as an issue on pydid (edit: Indicio-tech/pydid#186)

[esune]

👍🏻

[esune]

@ff137 are we good with merging this now, or do we want to wait and see if the upstream library fixes the issue first with aliasing field names?

[ff137]

So it seems that camelCase aliases are already implemented for the pydid resources. However, the getattr method only sees the original field names, not the aliased names.

The way to get aliased fields would be to dump the model to a dictionary: .model_dump_json(by_alias=True).
But, we already have the original dictionary in this context, so doing doc_raw.get(proof_purpose, []) should be fine. The assumption is just that this resolved doc will always have the expected camelCase format.

The pydid DIDDocument can deserialize either snake_case or camelCase fields, so that would be the benefit in first converting to a DIDDocument, so that it's standardised. But, if the standard is to use camelCase already, then it's much of a muchness.