Improve api documentation and error handling #2690
Conversation
Signed-off-by: jamshale <jamiehalebc@gmail.com>
Signed-off-by: jamshale <jamiehalebc@gmail.com>
Signed-off-by: jamshale <jamiehalebc@gmail.com>
| @@ -116,13 +131,17 @@ class Meta: | |||
| issuer_id = fields.Str( | |||
| description="Issuer Identifier of the credential definition or schema", | |||
| data_key="issuerId", | |||
| example=INDY_OR_KEY_DID_EXAMPLE, | |||
There was a problem hiding this comment.
I'm a little hesitant about the only examples showing as old Indy values. Is it possible to provide more than one example so we could do something like did:example:123?
Also, I believe most of these fields should be in the metadata kwarg now after updates to marshmallow, as a side note. That change probably belongs in a separate PR.
There was a problem hiding this comment.
I had thought about that. I was thinking when we add more methods we could figure something out for the other examples. For the metadata stuff I saw both methods used and wasn't sure the correct way.
I think the casing caused some confusion when this was getting implemented. The camel casing still applies for the anoncred objects so I'm pretty sure it's good. I basically just removed any discrepancies with search params and outer objects being mixed cases anymore. |
|
Add examples and descriptions to the swagger docs. Mostly used the indy examples.
There was some confusion with request params object attributes with camel vs snake casing. The anoncreds library does camel case like the spec decribes. Everywhere else has snake case. I changed everything so that only the inner anoncreds schema, cred_def and rev_reg_def objects are in camel case. I think this makes things a bit more clear but it is still a bit confusing when you are used to the indy api.
I changed the POST rev_reg_def endpoint to have a wrapper like schema and cred_def with an options object where you can put the endorser connection id.
Changed the GET all creds defs return object to be in a wrapper like schemas.
Fixed a couple mistakes and improved the error handling. I think a majority of errors give information about the cause now. Can still get 500 errors with unexpected param values.