UNPKG

@palmares/schemas

Version:

This defines a default schema definition for validation of data, it abstract popular schema validation libraries like zod, yup, valibot and others"

523 lines 21.8 kB
import { Schema } from './schema'; import type { DefinitionsOfSchemaType, ExtractTypeFromObjectOfSchemas } from './types'; import type { SchemaAdapter } from '../adapter'; import type { FieldAdapter } from '../adapter/fields'; import type { StandardSchemaV1 } from '@standard-schema/spec'; export declare class ObjectSchema<TType extends { input: any; validate: any; internal: any; representation: any; output: any; } = { input: Record<any, any>; output: Record<any, any>; validate: Record<any, any>; internal: Record<any, any>; representation: Record<any, any>; }, TDefinitions extends DefinitionsOfSchemaType = DefinitionsOfSchemaType<SchemaAdapter & Palmares.PSchemaAdapter>, TData extends Record<any, any> = Record<any, any>> extends Schema<TType, TDefinitions> implements StandardSchemaV1<TType['input'], TType['output']> { protected fieldType: string; protected __data: Record<any, Schema>; protected __cachedDataAsEntries: [string, Schema][]; protected __type: { message: string; check: (value: TType['input']) => boolean; }; constructor(data: TData); protected __retrieveDataAsEntriesAndCache(): [string, Schema][]; protected __transformToAdapter(options: Parameters<Schema['__transformToAdapter']>[0]): Promise<ReturnType<FieldAdapter['translate']>>; /** * Transform the data to the representation without validating it. This is useful when you want to return a data * from a query directly to the user. The core idea of this is that you can join the data from the database "by hand". * In other words, you can do the joins by yourself directly on code. For more complex cases, this can be really * helpful. * * @param value - The value to be transformed. * * @returns The transformed value. */ data(value: TType['output']): Promise<TType['representation']>; /** * This let's you refine the schema with custom validations. This is useful when you want to validate something * that is not supported by default by the schema adapter. * * @example * ```typescript * import * as p from '@palmares/schemas'; * * const numberSchema = p.number().refine((value) => { * if (value < 0) return { code: 'invalid_number', message: 'The number should be greater than 0' }; * }); * * const { errors, parsed } = await numberSchema.parse(-1); * * // [{ isValid: false, code: 'invalid_number', message: 'The number should be greater than 0', path: [] }] * console.log(errors); * ``` * * @param refinementCallback - The callback that will be called to validate the value. * @param options - Options for the refinement. * @param options.isAsync - Whether the callback is async or not. Defaults to true. */ refine<TRefinementCallback extends (args: { value: TType['input']; context: TDefinitions['context']; }) => Promise<void | undefined | { code: string; message: string; }>>(refinementCallback: TRefinementCallback): ObjectSchema<{ input: TType['input']; validate: TType['validate']; internal: TType['internal']; output: TType['output']; representation: TType['representation']; }, TDefinitions, TData>; /** * Allows the value to be either undefined or null. Different from the `optional` method on other schemas, You can * pass `outputOnly` as `true` to this method. This will allow you to pass `null` or `undefined` as a value on the * {@link Schema.data} method, but it will not allow the value to be `null` or `undefined`. This is useful for * typing purposes * * @example * ```typescript * import * as p from '@palmares/schemas'; * * const numberSchema = p.number().optional(); * * const { errors, parsed } = await numberSchema.parse(undefined); * * console.log(parsed); // undefined * * const { errors, parsed } = await numberSchema.parse(null); * * console.log(parsed); // null * * const { errors, parsed } = await numberSchema.parse(1); * * console.log(parsed); // 1 * * * const companySchema = p.object({ id: p.number(), name: p.string() }); * const userSchema = p.object({ * id: p.number(), * name: p.string(), * company: companySchema.optional({ outputOnly: true }) * }); * * // Will not allow the company to be null or undefined on a typing level. * const { errors, parsed } = await userSchema.data({ id: 1, name: 'John Doe' }); * * // Will allow the company to be null or undefined on a typing level * const value = await userSchema.data({ id: 1, name: 'John Doe' }); * ``` * * @returns - The schema we are working with. */ optional<TOutputOnly extends boolean = false>(options?: { message?: string; allow?: false; outputOnly?: TOutputOnly; }): ObjectSchema<TOutputOnly extends true ? { input: TType["input"]; validate: TType["validate"]; internal: TType["internal"]; output: TType["output"] | undefined | null; representation: TType["representation"]; } : { input: TType["input"] | undefined | null; validate: TType["validate"] | undefined | null; internal: TType["internal"] | undefined | null; output: TType["output"] | undefined | null; representation: TType["representation"] | undefined | null; }, TDefinitions, TData>; /** * Just adds a message when the value is undefined. It's just a syntax sugar for * * ```typescript * p.datetime().optional({ message: 'This value should be defined', allow: false }) * ``` * * @param options - The options of nonOptional function * @param options.message - A custom message if the value is undefined. * * @returns - The schema. */ nonOptional(options?: { message: string; }): ObjectSchema<{ input: TType["input"]; validate: TType["validate"]; internal: TType["internal"]; output: TType["output"]; representation: TType["representation"]; }, TDefinitions>; /** * Allows the value to be null and ONLY null. You can also use this function to set a custom message when the value * is NULL by setting * the { message: 'Your custom message', allow: false } on the options. * * @example * ```typescript * import * as p from '@palmares/schemas'; * * const numberSchema = p.number().nullable(); * * const { errors, parsed } = await numberSchema.parse(null); * * console.log(parsed); // null * * const { errors, parsed } = await numberSchema.parse(undefined); * * console.log(errors); // [{ isValid: false, code: 'invalid_type', message: 'Invalid type', path: [] }] * ``` * * @param options - The options for the nullable function. * @param options.message - The message to be shown when the value is not null. Defaults to 'Cannot be null'. * @param options.allow - Whether the value can be null or not. Defaults to true. * * @returns The schema. */ nullable(options?: { message: string; allow: false; }): ObjectSchema<{ input: TType["input"] | null; validate: TType["validate"] | null; internal: TType["internal"] | null; output: TType["output"] | null; representation: TType["representation"] | null; }, TDefinitions, TData>; /** * Just adds a message when the value is null. It's just a syntax sugar for * * ```typescript * p.datetime().nullable({ message: 'This value cannot be null', allow: false }) * ``` * * @param options - The options of nonNullable function * @param options.message - A custom message if the value is null. * * @returns - The schema. */ nonNullable(options?: { message: string; }): ObjectSchema<{ input: TType["input"]; validate: TType["validate"]; internal: TType["internal"]; output: TType["output"]; representation: TType["representation"]; }, TDefinitions>; /** * This method will remove the value from the representation of the schema. If the value is undefined it will keep * that way otherwise it will set the value to undefined after it's validated. * This is used in conjunction with the {@link data} function, the {@link parse} function or {@link validate} * function. This will remove the value from the representation of the schema. * * By default, the value will be removed just from the representation, in other words, when you call the {@link data} * function. * But if you want to remove the value from the internal representation, you can pass the argument `toInternal` * as true. * Then if you still want to remove the value from the representation, you will need to pass the argument * `toRepresentation` as true as well. * * @example * ```typescript * import * as p from '@palmares/schemas'; * * const userSchema = p.object({ * id: p.number().optional(), * name: p.string(), * password: p.string().omit() * }); * * const user = await userSchema.data({ * id: 1, * name: 'John Doe', * password: '123456' * }); * * console.log(user); // { id: 1, name: 'John Doe' } * ``` * * * @param args - By default, the value will be removed just from the representation, in other words, when you call * the {@link data} function. * But if you want to remove the value from the internal representation, you can pass the argument `toInternal` * as true. * Then if you still want to remove the value from the representation, you will need to pass the argument * `toRepresentation` as true as well. * * @returns The schema. */ omit<TToInternal extends boolean, TToRepresentation extends boolean = boolean extends TToInternal ? true : false>(args?: { toInternal?: TToInternal; toRepresentation?: TToRepresentation; }): ObjectSchema<{ input: TToInternal extends true ? TType["input"] | undefined : TType["input"]; validate: TToInternal extends true ? TType["validate"] | undefined : TType["validate"]; internal: TToInternal extends true ? undefined : TType["internal"]; output: TToRepresentation extends true ? TType["output"] | undefined : TType["output"]; representation: TToRepresentation extends true ? undefined : TType["representation"]; }, TDefinitions, TData>; /** * This function is used in conjunction with the {@link validate} function. It's used to save a value to an external * source like a database. You should always return the schema after you save the value, that way we will always have * the correct type of the schema after the save operation. * * You can use the {@link toRepresentation} function to transform and clean the value it returns after the save. * * @example * ```typescript * import * as p from '@palmares/schemas'; * * import { User } from './models'; * * const userSchema = p.object({ * id: p.number().optional(), * name: p.string(), * email: p.string().email(), * }).onSave(async (value) => { * // Create or update the user on the database using palmares models or any other library of your choice. * if (value.id) * await User.default.set(value, { search: { id: value.id } }); * else * await User.default.set(value); * * return value; * }); * * * // Then, on your controller, do something like this: * const { isValid, save, errors } = await userSchema.validate(req.body); * if (isValid) { * const savedValue = await save(); * return Response.json(savedValue, { status: 201 }); * } * * return Response.json({ errors }, { status: 400 }); * ``` * * @param callback - The callback that will be called to save the value on an external source. * * @returns The schema. */ onSave<TSave extends ((value: TType['internal']) => (context: any) => Promise<TType['output']>) | ((value: TType['internal']) => Promise<TType['output']>)>(callback: TSave): ObjectSchema<{ input: TType['input']; validate: TType['validate']; internal: TType['internal']; output: TType['output']; representation: TType['representation']; }, Omit<TDefinitions, 'hasSave' | 'context'> & { hasSave: true; context: ReturnType<TSave> extends (context: any) => any ? Parameters<ReturnType<TSave>>[0] : any; }, TData>; /** * This function is used to add a default value to the schema. If the value is either undefined or null, the default * value will be used. * * @example * ```typescript * import * as p from '@palmares/schemas'; * * const numberSchema = p.number().default(0); * * const { errors, parsed } = await numberSchema.parse(undefined); * * console.log(parsed); // 0 * ``` */ default<TDefaultValue extends TType['input'] | (() => Promise<TType['input']>)>(defaultValueOrFunction: TDefaultValue): ObjectSchema<{ input: TType["input"] | undefined | null; validate: TType["validate"]; internal: TType["internal"]; output: TType["output"] | undefined | null; representation: TType["representation"]; }, TDefinitions, TData>; /** * This function let's you customize the schema your own way. After we translate the schema on the adapter we call * this function to let you customize the custom schema your own way. Our API does not support passthrough? * No problem, you can use this function to customize the zod schema. * * @example * ```typescript * import * as p from '@palmares/schemas'; * * const numberSchema = p.number().extends((schema) => { * return schema.nonnegative(); * }); * * const { errors, parsed } = await numberSchema.parse(-1); * * // [{ isValid: false, code: 'nonnegative', message: 'The number should be nonnegative', path: [] }] * console.log(errors); * ``` * * @param callback - The callback that will be called to customize the schema. * @param toStringCallback - The callback that will be called to transform the schema to a string when you want to * compile the underlying schema to a string so you can save it for future runs. * * @returns The schema. */ extends(callback: (schema: Awaited<ReturnType<NonNullable<TDefinitions['schemaAdapter']['object']>['translate']>>) => Awaited<ReturnType<NonNullable<TDefinitions['schemaAdapter']['field']>['translate']>> | any, toStringCallback?: (schemaAsString: string) => string): this; /** * This function is used to transform the value to the representation of the schema. When using the {@link data} * function. With this function you have full control to add data cleaning for example, transforming the data and * whatever. Another use case is when you want to return deeply nested recursive data. * The schema maps to itself. * * @example * ```typescript * import * as p from '@palmares/schemas'; * * const recursiveSchema = p.object({ * id: p.number().optional(), * name: p.string(), * }).toRepresentation(async (value) => { * return { * id: value.id, * name: value.name, * children: await Promise.all(value.children.map(async (child) => await recursiveSchema.data(child))) * } * }); * * const data = await recursiveSchema.data({ * id: 1, * name: 'John Doe', * }); * ``` * * @example * ``` * import * as p from '@palmares/schemas'; * * const colorToRGBSchema = p.string().toRepresentation(async (value) => { * switch (value) { * case 'red': return { r: 255, g: 0, b: 0 }; * case 'green': return { r: 0, g: 255, b: 0 }; * case 'blue': return { r: 0, g: 0, b: 255 }; * default: return { r: 0, g: 0, b: 0 }; * } * }); * ``` * @param toRepresentationCallback - The callback that will be called to transform the value to the representation. * @param options - Options for the toRepresentation function. * @param options.after - Whether the toRepresentationCallback should be called after the existing * toRepresentationCallback. Defaults to true. * @param options.before - Whether the toRepresentationCallback should be called before the existing * toRepresentationCallback. Defaults to true. * * @returns The schema with a new return type */ toRepresentation<TRepresentation>(toRepresentationCallback: (value: TType['representation']) => Promise<TRepresentation>, options?: { after?: boolean; before?: boolean; }): ObjectSchema<{ input: TType["input"]; validate: TType["validate"]; internal: TType["internal"]; output: TType["output"]; representation: TRepresentation; }, TDefinitions, TData>; /** * This function is used to transform the value to the internal representation of the schema. This is useful when * you want to transform the value to a type that the schema adapter can understand. For example, you might want * to transform a string to a date. This is the function you use. * * @example * ```typescript * import * as p from '@palmares/schemas'; * * const dateSchema = p.string().toInternal((value) => { * return new Date(value); * }); * * const date = await dateSchema.parse('2021-01-01'); * * console.log(date); // Date object * * const rgbToColorSchema = p.object({ * r: p.number().min(0).max(255), * g: p.number().min(0).max(255), * b: p.number().min(0).max(255), * }).toInternal(async (value) => { * if (value.r === 255 && value.g === 0 && value.b === 0) return 'red'; * if (value.r === 0 && value.g === 255 && value.b === 0) return 'green'; * if (value.r === 0 && value.g === 0 && value.b === 255) return 'blue'; * return `rgb(${value.r}, ${value.g}, ${value.b})`; * }); * ``` * * @param toInternalCallback - The callback that will be called to transform the value to the internal representation. * * @returns The schema with a new return type. */ toInternal<TInternal>(toInternalCallback: (value: TType['validate']) => Promise<TInternal>): ObjectSchema<{ input: TType["input"]; validate: TType["validate"]; internal: TInternal; output: TType["output"]; representation: TType["representation"]; }, TDefinitions, TData>; /** * Called before the validation of the schema. Let's say that you want to validate a date that might receive a string, * you can convert that string to a date here BEFORE the validation. This pretty much transforms the value to a type * that the schema adapter can understand. * * @example * ``` * import * as p from '@palmares/schemas'; * import * as z from 'zod'; * * const customRecordToMapSchema = p.schema().appendSchema(z.map()).toValidate(async (value) => { * return new Map(value); // Before validating we transform the value to a map. * }); * * const { errors, parsed } = await customRecordToMapSchema.parse({ key: 'value' }); * ``` * * @param toValidateCallback - The callback that will be called to validate the value. * * @returns The schema with a new return type. */ toValidate<TValidate>(toValidateCallback: (value: TType['input'], context: TDefinitions['context']) => Promise<TValidate> | TValidate): ObjectSchema<{ input: TType["input"]; validate: TValidate; internal: TType["internal"]; output: TType["output"]; representation: TType["representation"]; }, TDefinitions, TData>; /** * Remove extraneous keys from the object, in other words, remove keys that are not in the schema. * * @example * ```typescript * import * as p from '@palmares/schemas'; * * const userSchema = p.object({ * id: p.number().optional(), * name: p.string(), * }).removeExtraneous(); * * const { errors, parsed } = await userSchema.parse({ * id: 1, * name: 'John Doe', * password: '123456' * }); * * console.log(parsed); // { id: 1, name: 'John Doe' } * ``` * * @returns The schema. */ removeExtraneous(): this; static new<TData extends Record<any, Schema<any, TDefinitions>>, TDefinitions extends DefinitionsOfSchemaType = DefinitionsOfSchemaType<SchemaAdapter & Palmares.PSchemaAdapter>>(data: TData): ObjectSchema<{ input: ExtractTypeFromObjectOfSchemas<TData, "input">; validate: ExtractTypeFromObjectOfSchemas<TData, "validate">; internal: ExtractTypeFromObjectOfSchemas<TData, "internal">; representation: ExtractTypeFromObjectOfSchemas<TData, "representation">; output: ExtractTypeFromObjectOfSchemas<TData, "output">; }, TDefinitions, TData>; } export declare const object: typeof ObjectSchema.new; //# sourceMappingURL=object.d.ts.map