Did dht test vectors

[andresuribe87]

Overview

Adds test vectors for did:dht for the create and resolve functions. The motivation is to ensure interoperability across implementations of the standard. This PR is related to TBD54566975/web5-spec#11, and will form the basis to fix the issue.

Description

This PR includes the following changes:

• Adds typed exceptions to ExceptionDeclarations.kt.
• Adds test vector files for create and resolve.
• Modifies the implementation of DidDht so that tests pass.
• Adds a README for the test-vectors added.

Once this is accepted, the test vectors will be promoted to the sdk-development repository.

Note

The test vectors were based off of https://did-dht.com/#test-vectors, and adjusted to match our test-vector structure.

How Has This Been Tested?

Unit tests

[nitro-neal]

remove

[andresuribe87]

Replaced with a logger.

[codecov]

Codecov Report

Merging #181 (02932be) into main (e822caf) will increase coverage by 0.03%.
The diff coverage is 85.18%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #181      +/-   ##
==========================================
+ Coverage   80.15%   80.18%   +0.03%     
==========================================
  Files          37       37              
  Lines        2116     2150      +34     
  Branches      297      299       +2     
==========================================
+ Hits         1696     1724      +28     
- Misses        283      288       +5     
- Partials      137      138       +1     

Components	Coverage Δ
credentials	81.38% <ø> (ø)
crypto	38.07% <ø> (ø)
dids	90.34% <85.18%> (-0.25%)	⬇️
common	80.55% <ø> (ø)

[frankhinek]

I could be misreading the Kotlin (KT newb 🖐️)...

It seems that the create operation test vector provides an identity key as input:

web5-kt/test-vectors/did_dht/create.json

Lines 7 to 13 in 02932be

	"identityPublicJwk": {
	"kty": "OKP",
	"crv": "Ed25519",
	"x": "YCcHYL2sYNPDlKaALcEmll2HHyT968M4UWbr-9CFGWE",
	"alg": "EdDSA",
	"kid": "0"
	}

But the DidDht implementation doesn't support specifying a pre-existing identity key and mocks the key generation so that the test vector passes:

web5-kt/dids/src/test/kotlin/web5/sdk/dids/methods/dht/DidDhtTest.kt

Lines 368 to 369 in 02932be

	val identityKeyId = keyManager.import(listOf(vector.input.identityPublicJwk!!)).first()
	doReturn(identityKeyId).whenever(keyManager).generatePrivateKey(JWSAlgorithm.EdDSA, Curve.Ed25519)

Had a number of long discussions with Moe and Adam this week on how to approach DID key management across all of our SDKs. It seems there's an opportunity to continue to get feedback from our partners and tweak how create(), load(), etc. work. But that should probably wait for another day and not block this PR.

[andresuribe87]

I could be misreading the Kotlin (KT newb 🖐️)...

It seems that the create operation test vector provides an identity key as input:

web5-kt/test-vectors/did_dht/create.json

Lines 7 to 13 in 02932be

	"identityPublicJwk": {
	"kty": "OKP",
	"crv": "Ed25519",
	"x": "YCcHYL2sYNPDlKaALcEmll2HHyT968M4UWbr-9CFGWE",
	"alg": "EdDSA",
	"kid": "0"
	}

But the DidDht implementation doesn't support specifying a pre-existing identity key and mocks the key generation so that the test vector passes:

web5-kt/dids/src/test/kotlin/web5/sdk/dids/methods/dht/DidDhtTest.kt

Lines 368 to 369 in 02932be

	val identityKeyId = keyManager.import(listOf(vector.input.identityPublicJwk!!)).first()
	doReturn(identityKeyId).whenever(keyManager).generatePrivateKey(JWSAlgorithm.EdDSA, Curve.Ed25519)

Had a number of long discussions with Moe and Adam this week on how to approach DID key management across all of our SDKs. It seems there's an opportunity to continue to get feedback from our partners and tweak how create(), load(), etc. work. But that should probably wait for another day and not block this PR.

Your reading is correct. The dht identity key cannot be passed in.

There were also other long discussions about the same topic during the ION implementation. The decision at that point in time is that the sdk should allow both paradigms: manually specifying the keys and also automatically creating them. That's the design that motivated

web5-kt/dids/src/main/kotlin/web5/sdk/dids/methods/ion/DidIon.kt

Line 794 in e822caf

public interface VerificationMethodSpec

I agree that such a discussion shouldn't block this.

[phoebe-lew]

@frankhinek might be good for us to schedule time to discuss key management across the SDK?

[nitro-neal]

need to remove Test

[nitro-neal]

should just be Web5TestVectorsDidDht