feat: 0.4.0 migration script

[TimoGlastra]

Adds migration script for 0.4.0 covering most of the changes between 0.3.1 and 0.4.0. There's some issues to resolve with how we are going to handle legacy identifiers vs qualified identifiers in AnonCreds (see #1391). did:sov dids are all transformed to did:indy dids already as we store the qualified indy did. Same for AnonCreds schema and cred def records, so it's mainly about anoncreds credential records that need some work.

Need to extend the migration docs a bit more, to say you need to add the anoncreds module to update the anoncreds related records, and that any dids created using agent.ledger. should be imported manually into the agent as we don't have the DidRecord for them.

This doesn't yet solve #998, but I left some comments there on my current thoughts

[codecov-commenter]

Codecov Report

Merging #1392 (fa9974d) into main (996c08f) will increase coverage by 0.06%.
The diff coverage is 90.32%.

@@            Coverage Diff             @@
##             main    #1392      +/-   ##
==========================================
+ Coverage   84.90%   84.96%   +0.06%     
==========================================
  Files         788      795       +7     
  Lines       19445    19606     +161     
  Branches     3162     3172      +10     
==========================================
+ Hits        16509    16658     +149     
- Misses       2929     2941      +12     
  Partials        7        7              

Impacted Files	Coverage Δ
.../repository/AnonCredsCredentialDefinitionRecord.ts	100.00% <ø> (ø)
.../anoncreds/src/repository/AnonCredsSchemaRecord.ts	100.00% <ø> (ø)
packages/core/src/error/index.ts	100.00% <ø> (ø)
packages/indy-sdk/src/wallet/IndySdkWallet.ts	46.69% <0.00%> (-0.39%)	⬇️
...kages/indy-vdr/src/dids/IndyVdrIndyDidRegistrar.ts	94.50% <ø> (ø)
packages/askar/src/wallet/AskarWallet.ts	47.61% <33.33%> (-0.14%)	⬇️
...re/src/wallet/error/WalletExportPathExistsError.ts	50.00% <50.00%> (ø)
...ages/core/src/storage/migration/UpdateAssistant.ts	86.32% <60.00%> (-3.20%)	⬇️
...e/src/storage/migration/updates/0.3.1-0.4/cache.ts	87.50% <87.50%> (ø)
...ages/anoncreds/src/updates/0.3.1-0.4/linkSecret.ts	89.47% <89.47%> (ø)
... and 12 more

... and 1 file with indirect coverage changes

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[berendsliedrecht]

Awesome! Looks extensively tested, very nice!

Do you think we should create some tests that do all the migrations in different order? Maybe they change some state in a weird way which breaks it (and seemingly unrevertable).

[berendsliedrecht]

which specific fields do you mean?

[TimoGlastra]

When you call create credential request in indy-sdk / anoncreds rs) it will return the following metadata:

{
          "master_secret_blinding_data": {
            "v_prime": "6088566065720309491695644944398283228337587174153857313170975821102428665682789111613194763354086540665993822078019981371868225077833338619179176775427438467982451441607103798898879602785159234518625137830139620180247716943526165654371269235270542103763086097868993123576876140373079243750364373248313759006451117374448224809216784667062369066076812328680472952148248732117690061334364498707450807760707599232005951883007442927332478453073050250159545354197772368724822531644722135760544102661829321297308144745035201971564171469931191452967102169235498946760810509797149446495254099095221645804379785022515460071863075055785600423275733199",
            "vr_prime": null
          },
          "nonce": "131502096406868204437821",
          "master_secret_name": "walletId28c602347-3f6e-429f-93cd-d5aa7856ef3f"
        }

If we move to using link secret in AnonCreds RS, we should update all metadata and transform the master_secret keys to link_secret keys in the Indy SDK holder service in AFJ.

Did you also update these fields in your link secret pr in anoncreds rs?

[TimoGlastra]

Do you think we should create some tests that do all the migrations in different order? Maybe they change some state in a weird way which breaks it (and seemingly unrevertable).

Do you mean the indy sdk to askar update script and the normal update assistant? If so, yes that'd probably be a good idea!