How to generate Bento Configuration Validator Schema for Tyk

There is a script named generate_bento_config_schema under apidef/streams/bento/schema folder in Tyk GW repository. This script generates a JSON schema for the input & output resources that Tyk supports. There are two ways to run this script.

Running the script directly

Simply, go run generate_bento_config_schema.go command in the source code folder. It'll generate a file named bento-config-schema.json in the current working folder. You can also set an output path via -output <string> parameter.

This is useful for development purposes.

Running via task runner

Simply run the following command under Tyk GW repository’s root folder.

task generate-bento-config-validator-schema

The task will automatically update apidef/streams/bento/schema/bento-config-schema.json file. This should be done after upgrading Bento in Tyk GW.

How to add a new input & output resources

1- Import the related component for its side effects.

For example if you want to produce a JSON schema that supports redis component, you can import it like the following:

 _ "http://github.com/warpstreamlabs/bento/public/components/redis"`

2- Add the component name to supportedSources slice. You should know that some components exposes different input/output sources.

For example, components/kafka exposes kafka and kafka_franz. You need to dig into the Bento's codebase to understand which input/output is exposed by a component.

See the list of components: https://github.com/warpstreamlabs/bento/tree/main/public/components

Importing all components was not preferred because it results in generating a gigantic JSON schema, and it’s too hard to navigate in that file and debug possible issues.

How to add a new validator rule

A validator rule has been defined by a simple interface:

// customValidationRule is an interface for defining custom validation rules for JSON schemas.
type customValidationRule interface {
	// Name returns the name of the validation rule.
	Name() string

	// Apply applies the validation rule to the provided input data and
	// returns the processed data or an error if validation fails.
	Apply(input []byte) ([]byte, error)
}

Apply method takes the input and returns the processed input by the validator rule implementation.

You can add rules to the pipeline like the following, generateBentoConfigSchema will take a list of rules and apply them respectively.

err = generateBentoConfigSchema(args.output, []customValidationRule{
	&addURIFormatToHTTPClientRule{}
})