json-schema-validator

This multiplatform library is an implementation of JSON schema that can validate JsonElement from kotlinx.serialization-json library.

Usage

Supported targets

Target
jvm
js
wasmJs
macosX64
macosArm64
iosArm64
iosSimulatorArm64
linuxX64
linuxArm64
mingwX64

Dependencies

Releases

In order to use releases add Maven Central repository to the list of repositories.

Kotlin

repositories {
  mavenCentral()
}

implementation("io.github.optimumcode:json-schema-validator:0.3.0")

Groovy

repositories {
  mavenCentral()
}

implementation 'io.github.optimumcode:json-schema-validator:0.3.0'

Release are published to Sonatype repository. The synchronization with Maven Central takes time. If you want to use the release right after the publication you should add Sonatype Release repository to your build script.

Kotlin

repositories {
  maven(url = "https://s01.oss.sonatype.org/content/repositories/releases/")
}

Groovy

repositories {
  maven { url 'https://s01.oss.sonatype.org/content/repositories/releases/' }
}

Snapshots

If you want to use SNAPSHOT version you should add Sonatype Snapshot repository to your build script.

Kotlin

repositories {
  maven(url = "https://s01.oss.sonatype.org/content/repositories/snapshots")
}

implementation("io.github.optimumcode:json-schema-validator:0.3.1-SNAPSHOT")

Groovy

repositories {
  maven { url 'https://s01.oss.sonatype.org/content/repositories/snapshots' }
}

implementation 'io.github.optimumcode:json-schema-validator:0.3.1-SNAPSHOT'

Example

If you have just one JSON schema or many independent schemes you can create it using factory methods defined on JsonSchema class.

import io.github.optimumcode.json.schema.JsonSchema
import io.github.optimumcode.json.schema.ValidationError
import kotlinx.serialization.json.JsonElement

val key = "\$" // to use $ in multiline string
val schema = JsonSchema.fromDefinition(
  """
  {
    "${key}schema": "http://json-schema.org/draft-07/schema#",
    "definitions": {
      "positiveInteger": {
        "type": "integer",
        "minimum": 0
      }
    },
    "properties": {
      "size": { "${key}ref": "#/definitions/positiveInteger" }
    }
  }
  """.trimIndent(),
)
val errors = mutableListOf<ValidationError>()
val elementToValidate: JsonElement = loadJsonToValidate()

val valid = schema.validate(elementToValidate, errors::add)

You can also use predefined ValidationOutputs to collect the results. Output formats are defined in draft 2020-12. The most performance can be achieved by using either flag or basic collectors. The detailed and verbose provide more structured information but this adds additional cost to the validation process (because they collect hierarchical output).

import io.github.optimumcode.json.schema.JsonSchema
import io.github.optimumcode.json.schema.OutputCollector
import io.github.optimumcode.json.schema.ValidationOutput.Flag
import io.github.optimumcode.json.schema.ValidationOutput.Basic
import io.github.optimumcode.json.schema.ValidationOutput.OutputUnit

val flag: Flag = schema.validate(elementToValidate, OutputCollector.flag())
val basic: Basic = schema.validate(elementToValidate, OutputCollector.basic())
val detailed: OutputUnit = schema.validate(elementToValidate, OutputCollector.detailed())
val verbose: OutputUnit = schema.validate(elementToValidate, OutputCollector.verbose())

If you need to use more than one schema, and they have references to other schemas you should use JsonSchemaLoader class.

import io.github.optimumcode.json.schema.JsonSchemaLoader
import io.github.optimumcode.json.schema.JsonSchema
import io.github.optimumcode.json.schema.ValidationError
import kotlinx.serialization.json.JsonElement

val schema: JsonSchema = JsonSchemaLoader.create()
  .register(
    """
    {
      "${KEY}id": "https://test.com",
      "properties": {
        "name": {
          "type": "string"
        }
      }
    }
    """.trimIndent(),
  ).fromDefinition(
    """
    {
      "properties": {
        "anotherName": {
          "${KEY}ref": "https://test.com#/properties/name"
        }
      }
    }
    """.trimIndent(),
  )

val errors = mutableListOf<ValidationError>()
val elementToValidate: JsonElement = loadJsonToValidate()

val valid = schema.validate(elementToValidate, errors::add)

Supported JSON schema drafts:

• Draft 7

  Supported keywords

  • Core

  Keyword	Status
  $id	Supported. $id in sub-schemas are collected as well and can be used in $ref
  $schema	Supported. Validates if schema is one of the supported schemas. The last supported is used if empty
  $ref	Supported
  definitions	Supported. Definitions are loaded and can be referenced

  • Assertions

  Category	Assertion	Status
  General	type	Supported all type defined in the specification
  	enum	Supported
  	const	Supported
  Numbers	multipleOf	Supported
  	maximum	Supported
  	exclusiveMaximum	Supported
  	minimum	Supported
  	exclusiveMinimum	Supported
  Strings	maxLength	Supported
  	minLength	Supported
  	pattern	Supported (kotlin.text.Regex is used)
  Arrays	items	Supported
  	additionalItems	Supported
  	maxItems	Supported
  	uniqueItems	Supported
  	contains	Supported
  Objects	maxProperties	Supported
  	minProperties	Supported
  	required	Supported
  	properties	Supported
  	patternProperties	Supported (kotlin.text.Regex is used)
  	additionalProperties	Supported
  	dependencies	Supported
  	propertyNames	Supported
  Conditions	if/then/else	Supported
  Boolean logic	allOf	Supported
  	anyOf	Supported (all validation will be executed even if the element matches the first one)
  	oneOf	Supported
  	not	Supported

• Draft 2019-09

  Supported keywords

  • Core

  Keyword	Status
  $id	Supported. $id in sub-schemas are collected as well and can be used in $ref
  $schema	Supported. Validates if schema is one of the supported schemas. The last supported is used if empty
  $ref	Supported
  $recursiveRef	Supported
  $defs/definitions	Supported. Definitions are loaded and can be referenced
  $vocabulary	Supported. You can disable and enable vocabularies through custom meta-schemes

  • Assertions

  Category	Assertion	Status
  General	type	Supported all type defined in the specification
  	enum	Supported
  	const	Supported
  Numbers	multipleOf	Supported
  	maximum	Supported
  	exclusiveMaximum	Supported
  	minimum	Supported
  	exclusiveMinimum	Supported
  Strings	maxLength	Supported
  	minLength	Supported
  	pattern	Supported (kotlin.text.Regex is used)
  Arrays	items	Supported
  	additionalItems	Supported
  	unevaluatedItems	Supported
  	maxItems	Supported
  	uniqueItems	Supported
  	contains	Supported
  	minContains	Supported (does not affect the work of contains assertion anyhow even if minContains=0)
  	maxContains	Supported
  Objects	maxProperties	Supported
  	minProperties	Supported
  	required	Supported
  	properties	Supported
  	patternProperties	Supported (kotlin.text.Regex is used)
  	additionalProperties	Supported
  	unevaluatedProperties	Supported
  	dependentRequired	Supported
  	dependentSchemas	Supported
  	propertyNames	Supported
  Conditions	if/then/else	Supported
  Boolean logic	allOf	Supported
  	anyOf	Supported (all validation will be executed even if the element matches the first one)
  	oneOf	Supported
  	not	Supported

• Draft 2020-12

  Supported keywords

  • Core

  Keyword	Status
  $id	Supported. $id in sub-schemas are collected as well and can be used in $ref
  $schema	Supported. Validates if schema is one of the supported schemas. The last supported is used if empty
  $ref	Supported
  $dynamicRef/$dynamicAnchor	Supported
  $defs/definitions	Supported. Definitions are loaded and can be referenced
  $vocabulary	Supported. You can disable and enable vocabularies through custom meta-schemes

  • Assertions

  Category	Assertion	Status
  General	type	Supported all type defined in the specification
  	enum	Supported
  	const	Supported
  Numbers	multipleOf	Supported
  	maximum	Supported
  	exclusiveMaximum	Supported
  	minimum	Supported
  	exclusiveMinimum	Supported
  Strings	maxLength	Supported
  	minLength	Supported
  	pattern	Supported (kotlin.text.Regex is used)
  Arrays	prefixItems	Supported
  	items	Supported
  	unevaluatedItems	Supported
  	maxItems	Supported
  	uniqueItems	Supported
  	contains	Supported
  	minContains	Supported
  	maxContains	Supported
  Objects	maxProperties	Supported
  	minProperties	Supported
  	required	Supported
  	properties	Supported
  	patternProperties	Supported (kotlin.text.Regex is used)
  	additionalProperties	Supported
  	unevaluatedProperties	Supported
  	dependentRequired	Supported
  	dependentSchemas	Supported
  	propertyNames	Supported
  Conditions	if/then/else	Supported
  Boolean logic	allOf	Supported
  	anyOf	Supported (all validation will be executed even if the element matches the first one)
  	oneOf	Supported
  	not	Supported

Format assertion

The library supports format assertion. All formats from JSON schema draft 2020-12 are supported.

According to JSON schema specification the format keyword produces assertion by default for all drafts before draft 2019-09. Starting from draft 2019-09 the format keyword produces only annotation by default.

You can change the default behaviour by specifying the corresponding option in JsonSchemaLoader.

import io.github.optimumcode.json.schema.FormatBehavior.ANNOTATION_AND_ASSERTION
import io.github.optimumcode.json.schema.JsonSchemaLoader
import io.github.optimumcode.json.schema.SchemaOption

val schema = JsonSchemaLoader.create()
       // option to change the default behavior
      .withSchemaOption(SchemaOption.FORMAT_BEHAVIOR_OPTION, ANNOTATION_AND_ASSERTION)
      // or
      .withSchemaOption(SchemaOption.FORMAT_BEHAVIOR_OPTION, ANNOTATION_ONLY)
      .fromDefinition(schemaDefinition)

Alternatively, for drafts starting from the draft 2019-09 you can either use a custom meta-schema with format-assertion vocabulary enabled and use this meta-schema in $schema property or define $vocabulary block with enabled format assertion in your schema.

Implementation details:

• regex - to implement regex format Kotlin Regex class is used. Because of that, result might vary depending on the platform where validation is executed (KT-49557). Please, be aware of it when using this library.
  If you know a KMM library that provides support for ECMA-262 Regex format I would appreciate it if you could find some time to create an issue with information about that library.

There is also an API to implement the user's defined format validation. The FormatValidator interface can be user for that. The custom format validators can be register in JsonSchemaLoader.

Please note, that the format validation API is marked as experimental and will require OptIn declaration in your code.

Custom assertions

You can implement custom assertions and use them. Read more here.

Compliance to JSON schema test suites

This library uses official JSON schema test suites as a part of the CI to make sure the validation meet the expected behavior. Not everything is supported right now but the missing functionality might be added in the future. The tests are located here.

NOTE: Python 3.* is required to run test-suites. It is used to generate list of remote schemas using this script

This library is also integrated into bowtie and runs against the official test suite along with other libraries. You can find the report here.

Benchmarking

There is a benchmark project that compares this library with some other ones:

• OpenAPI schema validator
• Networknt Schema Validator

The benchmark is scheduled to run every night on Monday. You can see the results in the latest workflow execution.

Also, benchmark results are available here in nice charts. Many thanks to maintainers of github-action-benchmark repo.

Developer notes

Build process

The update to Kotlin 1.9.22 came with an issue for JS incremental compilation. In case you see an error about main function that already bind please execute clean task.

When you build project for linux target you might get an error about missing native library. This is because com.doist.x:normalize requires this library to perform string normalization. This is needed to support idn-hostname format. Install this library with the following command:

sudo apt-get install -y libunistring-dev

Devcontainer

Devcontainers is a cool feature. However, by default in Codespaces and DevPod you will use VS Code. This is a good IDE but not for Kotlin, unfortunately. The extension that is available for VS Code to support Kotlin works quite slow (when workspace is just started) and sometimes does not work correctly with multiplatform definitions. Because of that I higly recoment using JetBrains Gateway (but it looks like the GitHub Codespace provider is not currently compatible with latest Gateway version). However, there is a way to connect to Codespace and work with project using JetBrains IDE. Please, read details here.

Future plans

• Add $schema property validation (if not set the latest supported will be used)
• Add proper $id support (for nested schemas and for referencing)
• Add support for newer drafts
  • Draft 2019-09 (Draft 8)
  • 2020-12
• Add support for schemas from external documents
  • Load schemas from local sources
  • Load schemas from remote sources
• Formalize error output as it is defined in the latest drafts

License