Extended Scalars for graphql-java

Overview

This library provides extended scalars for graphql-java

GraphQL Scalars are the primitive leaf values in the GraphQL type system which cannot be queried further via sub-field selections.

The GraphQL Specification defines String, Int, Float, Boolean and ID as well-defined built-in scalars that must be present in a graphql type system. Beyond these, it is up to an implementation to decide what custom scalars are present.

You would use custom scalars when you want to describe more meaningful behavior or ranges of values.

Getting started

How to install

To use this library put the following into your gradle config

implementation 'com.graphql-java:graphql-java-extended-scalars:20.2'

or the following into your Maven config

<dependency>
  <groupId>com.graphql-java</groupId>
  <artifactId>graphql-java-extended-scalars</artifactId>
  <version>20.2</version>
</dependency>

Note:

use 19.0 or above for graphql-java 19.x and above

use 20.0 or above for graphql-java 20.x and above

How to use extended scalars

Direct use

Register the scalar with graphql-java

RuntimeWiring.newRuntimeWiring().scalar(ExtendedScalars.DateTime)

Spring for GraphQL

If you are using Spring for GraphQL, register the scalar with RuntimeWiringConfigurer

@Configuration
public class GraphQlConfig {
    @Bean
    public RuntimeWiringConfigurer runtimeWiringConfigurer() {
        return wiringBuilder -> wiringBuilder.scalar(ExtendedScalars.DateTime);
    }
}

Netflix DGS

If you are using Netflix DGS, please see their configuration documentation.

How to add extended scalars to your schema

The GraphQL Specification recommends the use of the @specifiedBy built-in schema directive to provide a scalar specification URL for specifying the behavior of custom scalar types.

directive @specifiedBy(url: String!) on SCALAR

To use a extended scalar in your schema, define the scalar like shown below for DateTime

scalar DateTime
  @specifiedBy(url: "https://scalars.graphql.org/andimarek/date-time.html")

type Something {
  someDateTime: DateTime
}

Custom Scalars

Alias Scalars

scalar AliasedScalar

For example a link to a social media post could be representing by a String but the name SocialMediaLink is a more semantically meaningful name for that scalar type.

For example, you would build it like this:

AliasedScalar socialMediaLink = ExtendedScalars.newAliasedScalar("SocialMediaLink")
        .aliasedScalar(Scalars.GraphQLString)
        .build()

And use it in a SDL schema like this :

type Customer {
  name: String
  socialMediaLink: SocialMediaLink
}

Date & Time Scalars

scalar DateTime
  @specifiedBy(url: 
    "https://scalars.graphql.org/andimarek/date-time.html"
  )

scalar Date
  @specifiedBy(url: 
    "https://tools.ietf.org/html/rfc3339"
  )

scalar Time
  @specifiedBy(url: 
    "https://tools.ietf.org/html/rfc3339"
  )

scalar LocalTime

An example declaration in SDL might be:

type Customer {
  birthDay: Date
  workStartTime: Time
  bornAt: DateTime
}

type Query {
  customers(bornAfter: DateTime): [Customers]
}

And example query might look like:

query {
  customers(bornAfter: "1996-12-19T16:39:57-08:00") {
    birthDay
    bornAt
  }
}

ID Scalars

scalar UUID
  @specifiedBy(url: 
    "https://tools.ietf.org/html/rfc4122"
  )

Numeric Scalars

scalar PositiveInt

scalar NegativeInt

scalar NonPositiveInt

scalar NonNegativeInt

scalar PositiveFloat

scalar NegativeFloat

scalar NonPositiveFloat

scalar NonNegativeFloat

The numeric scalars are derivations of the standard GraphQL Int and Float scalars that enforce range limits.

An example declaration in SDL might be:

type Customer {
  name: String
  currentHeight: PositiveInt
  weightLossGoal: NonPositiveInt
  averageWeightLoss: NegativeFloat
}

type Query {
  customers(height: PositiveInt): [Customers]
}

And example query might look like:

query {
  customers(height: 182) {
    name
    height
    weightLossGoal
  }
}

Java Primitives

scalar GraphQLLong

scalar GraphQLShort

scalar GraphQLByte

scalar GraphQLBigDecimal

scalar GraphQLBigInteger

scalar GraphQLChar

Locale Scalar

scalar Locale
  @specifiedBy(url: 
    "https://tools.ietf.org/html/bcp47"
  )

type Customer {
  name: String
  locale: Locale
}

type Query {
  customers(inLocale: Locale): [Customers]
}

An example query to look for customers in the Romanian locale might look like:

query {
  customers(inLocale: "ro-RO") {
    name
    locale
  }
}

Country Code Scalar

scalar CountryCode 
  @specifiedBy(url: 
    "https://en.wikipedia.org/wiki/ISO_3166-1_alpha-2"
  )

An example declaration in SDL might be:

scalar CountryCode

type Customer {
  name: String
  countryCode: CountryCode
}

And example query might look like:

query {
  customers(code: "US") {
    name
    countryCode
  }
}

Currency Scalar

scalar Currency
  @specifiedBy(url: 
    "https://en.wikipedia.org/wiki/ISO_4217"
  )

An example declaration in SDL might be:

scalar Currency

type Account {
  id: String
  currency: Currency
  accountNumber: String
}

And example query might look like:

query {
  accounts(currency: "USD") {
    id
    currency
    accountNumber
  }
}

URL Scalars

scalar URL
  @specifiedBy(url: 
    "https://www.w3.org/Addressing/URL/url-spec.txt"
  )

Object / JSON Scalars

scalar Object

scalar JSON

One of the design goals of GraphQL, is that the type system describes the shape of the data returned.

The Object / JSON scalars work against this some what because they can return compound values outside the type system. As such they should be used sparingly. In general your should aim to describe the data via the GraphQL type system where you can and only resort to the Object / JSON scalars in very rare circumstances.

An example might be an extensible GraphQL system where systems can input custom metadata objects that cant be known at schema type design time.

An example declaration in SDL might be:

type Customer {
  name: String
  associatedMetaData: JSON
}

type Query {
  customers(filterSyntax: JSON): [Customers]
}

And example query might look like:

query {
  customers(
    filterSyntax: {
      startSpan: "First"
      matchCriteria: { countryCode: "AU", isoCodes: ["27B-34R", "95A-E23"] }
    }
  ) {
    name
    associatedMetaData
  }
}

Note : The JSON scalar is a simple alias type to the Object scalar because often the returned data is a blob of JSON. They are all just objects at runtime in graphql-java terms and what network serialization protocol is up to you. Choose whichever name you think adds more semantic readers to your schema consumers.

Regex Scalars

The RegexScalar has a builder where you provide one or more regex patterns that control the acceptable values for a new scalar.

You name the scalar and it provides an implementation.

For example, imagine a phoneNumber scalar like this :

    RegexScalar phoneNumberScalar = ExtendedScalars.newRegexScalar("phoneNumber")
            .addPattern(Pattern.compile("\\([0-9]*\\)[0-9]*"))
            .build()