GraphQL code generator plugin to generate form validation schema from your GraphQL schema.
Start by installing this plugin and write simple plugin config;
$ npm i graphql-codegen-typescript-validation-schema
generates:
path/to/graphql.ts:
plugins:
- typescript
- typescript-validation-schema # specify to use this plugin
config:
# You can put the config for typescript plugin here
# see: https://www.graphql-code-generator.com/plugins/typescript
strictScalars: true
# Overrides built-in ID scalar to both input and output types as string.
# see: https://the-guild.dev/graphql/codegen/plugins/typescript/typescript#scalars
scalars:
ID: string
# You can also write the config for this plugin together
schema: yup # or zod
It is recommended to write scalars
config for built-in type ID
, as in the yaml example shown above. For more information: #375
You can check example directory if you want to see more complex config example or how is generated some files.
The Q&A for each schema is written in the README in the respective example directory.
schema
type: ValidationSchema
default: 'yup'
Specify generete validation schema you want.
You can specify yup
or zod
or myzod
.
generates:
path/to/graphql.ts:
plugins:
- typescript
- typescript-validation-schema
config:
schema: yup
importFrom
type: string
When provided, import types from the generated typescript types file path. if not given, omit import statement.
generates:
path/to/graphql.ts:
plugins:
- typescript
path/to/validation.ts:
plugins:
- typescript-validation-schema
config:
importFrom: ./graphql # path for generated ts code
Then the generator generates code with import statement like below.
import { GeneratedInput } from './graphql'
/* generates validation schema here */
schemaNamespacedImportName
type: string
If defined, will use named imports from the specified module (defined in importFrom
) rather than individual imports for each type.
generates:
path/to/types.ts:
plugins:
- typescript
path/to/schemas.ts:
plugins:
- graphql-codegen-validation-schema
config:
schema: yup
importFrom: ./path/to/types
schemaNamespacedImportName: types
Then the generator generates code with import statement like below.
import * as types from './graphql'
/* generates validation schema here */
useTypeImports
type: boolean
default: false
Will use import type {}
rather than import {}
when importing generated TypeScript types.
This gives compatibility with TypeScript's "importsNotUsedAsValues": "error" option.
Should used in conjunction with importFrom
option.
typesPrefix
type: string
default: (empty)
Prefixes all import types from generated typescript type.
generates:
path/to/graphql.ts:
plugins:
- typescript
path/to/validation.ts:
plugins:
- typescript-validation-schema
config:
typesPrefix: I
importFrom: ./graphql # path for generated ts code
Then the generator generates code with import statement like below.
import { IGeneratedInput } from './graphql'
/* generates validation schema here */
typesSuffix
type: string
default: (empty)
Suffixes all import types from generated typescript type.
generates:
path/to/graphql.ts:
plugins:
- typescript
path/to/validation.ts:
plugins:
- typescript-validation-schema
config:
typesSuffix: I
importFrom: ./graphql # path for generated ts code
Then the generator generates code with import statement like below.
import { GeneratedInputI } from './graphql'
/* generates validation schema here */
enumsAsTypes
type: boolean
default: false
Generates enum as TypeScript type
instead of enum
.
notAllowEmptyString
type: boolean
default: false
Generates validation string schema as do not allow empty characters by default.
scalarSchemas
type: ScalarSchemas
Extends or overrides validation schema for the built-in scalars and custom GraphQL scalars.
config:
schema: yup
scalarSchemas:
Date: yup.date()
Email: yup.string().email()
config:
schema: zod
scalarSchemas:
Date: z.date()
Email: z.string().email()
defaultScalarTypeSchema
type: string
Fallback scalar type for undefined scalar types in the schema not found in scalarSchemas
.
config:
schema: yup
defaultScalarSchema: yup.unknown()
config:
schema: zod
defaultScalarSchema: z.unknown()
withObjectType
type: boolean
default: false
Generates validation schema with GraphQL type objects. But excludes Query
, Mutation
, Subscription
objects.
It is currently added for the purpose of using simple objects. See also #20, #107.
This option currently does not support fragment generation. If you are interested, send me PR would be greatly appreciated!
validationSchemaExportType
type: ValidationSchemaExportType
default: 'function'
Specify validation schema export type.
useEnumTypeAsDefaultValue
type: boolean
default: false
Uses the full path of the enum type as the default value instead of the stringified value.
namingConvention
type: NamingConventionMap
default: { enumValues: "change-case-all#pascalCase" }
Uses the full path of the enum type as the default value instead of the stringified value.
Related: https://the-guild.dev/graphql/codegen/docs/config-reference/naming-convention#namingconvention
directives
type: DirectiveConfig
Generates validation schema with more API based on directive schema. For example, yaml config and GraphQL schema is here.
input ExampleInput {
email: String! @required(msg: "Hello, World!") @constraint(minLength: 50, format: "email")
message: String! @constraint(startsWith: "Hello")
}
generates:
path/to/graphql.ts:
plugins:
- typescript
- typescript-validation-schema
config:
schema: yup
directives:
# Write directives like
#
# directive:
# arg1: schemaApi
# arg2: ["schemaApi2", "Hello $1"]
#
# See more examples in `./tests/directive.spec.ts`
# https://github.com/Code-Hex/graphql-codegen-typescript-validation-schema/blob/main/tests/directive.spec.ts
required:
msg: required
constraint:
minLength: min
# Replace $1 with specified `startsWith` argument value of the constraint directive
startsWith: [matches, /^$1/]
format:
# This example means `validation-schema: directive-arg`
# directive-arg is supported String and Enum.
email: email
Then generates yup validation schema like below.
export function ExampleInputSchema(): yup.SchemaOf<ExampleInput> {
return yup.object({
email: yup.string().defined().required('Hello, World!').min(50).email(),
message: yup.string().defined().matches(/^Hello/)
})
}
generates:
path/to/graphql.ts:
plugins:
- typescript
- typescript-validation-schema
config:
schema: zod
directives:
# Write directives like
#
# directive:
# arg1: schemaApi
# arg2: ["schemaApi2", "Hello $1"]
#
# See more examples in `./tests/directive.spec.ts`
# https://github.com/Code-Hex/graphql-codegen-typescript-validation-schema/blob/main/tests/directive.spec.ts
constraint:
minLength: min
# Replace $1 with specified `startsWith` argument value of the constraint directive
startsWith: [regex, /^$1/, message]
format:
# This example means `validation-schema: directive-arg`
# directive-arg is supported String and Enum.
email: email
Then generates zod validation schema like below.
export function ExampleInputSchema(): z.ZodSchema<ExampleInput> {
return z.object({
email: z.string().min(50).email(),
message: z.string().regex(/^Hello/, 'message')
})
}
Please see example directory.
Their is currently a compatibility issue with the client-preset. A workaround for this is to split the generation into two (one for client-preset and one for typescript-validation-schema).
generates:
path/to/graphql.ts:
plugins:
- typescript-validation-schema
/path/to/graphql/:
preset: 'client',
plugins:
...