How to split enum variants and their contents in Rust to generate correct JSON schema

Question

With schemars, it is possible to generate a JSON schema from Rust code.

Consider this JSON example:

{
   "cause" : {
      "type" : "EVT1",
      "payload" : {
         "cause_id" : "ABC-123"
      }
   } 
}

and the Rust code used to generate the schema:

use schemar::JsonSchema;
use serde::Serialize;

#[derive(Serialize, JsonSchema)]
struct Cause {
    _type: EventType,
    payload: EventPayload,
}

#[derive(Serialize, JsonSchema)]
enum EventType {
    EVT1 { param1: String, param2: String },
    EVT2 { param1: String },
}

#[derive(Serialize, JsonSchema)]
enum EventPayload {
    // ?
}

EventType contains the type of event in the name of the variant, and the payload is specified by EventType as well. EventPayload is superfluous.

Is there any possibility to configure serde so that the name of the EventType variant will be used as type, the contents of it as payload? The payload depends on EventType.

What do I expect?

• _type shall be restricted to the variants of EventType
• the payload of cause is directly dependant of _type
• Ideally, the schema generation would use the name of EventType variant as _type, and take the payload from the variants contents.