ENH: introduce conditional annotation for schema generation

[chenqi0805]

Description

This PR

• introduces IfPresentAlsoRequire annotation to mark mutually dependent properties in JSON schema generation
• modifies json schema generator to respect the annotation.

Note: The generated schema example is as follows:

{
  "$schema" : "https://json-schema.org/draft/2020-12/schema",
  "type" : "object",
  "properties" : {
    "custom_test_attribute" : {
      "type" : "string"
    },
    "test_attribute_with_default_value" : {
      "type" : "string",
      "default" : "default_value"
    },
    "test_attribute_with_getter" : {
      "type" : "string"
    },
    "test_mutually_exclusive_attribute_a" : {
      "type" : "string"
    }
  },
  "dependentRequired" : {
    "test_mutually_exclusive_attribute_a" : [ "test_mutually_exclusive_attribute_b:null" ]
  },
  "description" : "test config"
}

Issues Resolved

Resolves #[Issue number to be closed when this PR is merged]

Check List

• New functionality includes testing.
• New functionality has a documentation issue. Please link to it in this PR.
  • New functionality has javadoc added
• Commits are signed with a real name per the DCO

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check here.

[dlvenable]

We should put this in the package org.opensearch.dataprepper.model.annotations.

[dlvenable]

Schemas and validations are tightly related. So we should aim to have a common language for both schemas and validations. Otherwise the developers will need to track schemas and validations independently and this may introduce misses.

I think a few modifications to the documentation and package will help here.

What do you think about using the name @AlsoRequires()?

[chenqi0805]

Unfortunately I am not able to find an out-of-the-box annotation from Jakarta or JsonProperty to achieve this schema property. Thus the custom annotation. I agree the tradeoff would be to maintain a set of custom annotation for schema only.

So we should aim to have a common language for both schemas and validations.

Could you specify what you mean here?

[dlvenable]

I believe you are correct - Jakarta does not have this annotation.

However, you are adding an annotation that you expect to be useful for schema generation. I suggest that we take a different approach. You should be adding annotations that let plugin developers express constraints in one place. Data Prepper should be able to use these annotations for both schema generation and for validation.

So:

1. We need a new annotation because there is not one available.
2. We should design it to support validation.
3. We should document it as being able to express configuration constraints, not as simply a schema generation tool.

[chenqi0805]

Thanks! I will open a separate issue for that. To incorporate validation requires additional work in pipeline parsing.

[chenqi0805]

#4935

[dlvenable]

What is this :null part for? Should we have a distinct model to represent the additional requirement?

e.g.

@IfPresentAlsoRequire({@Required(name = "test_mutually_exclusive_attribute", allowedValues = {null}))

[dlvenable]

I made a small documentation suggestion.

[dlvenable]

Let's make the context clearer. Here are two options.

1. Make this a nested annotation in AlsoRequires
2. Rename to @AlsoRequired.

I think I like option 1.

[dlvenable]

Thanks @chenqi0805 for this contribution!