Osv NPM | npm.io

osv - Object Schema Validation

This library can be used for validating JavaScript objects or other data types like numbers, strings, booleans or arrays. It is very strongly inspired by @hapijs/joi and mongoose schema validation.

import OSV from 'osv';

const schema = OSV.schema({
  users: OSV.array({
    required: true,
    item: OSV.schema({
      username: OSV.string({
        required: true,
        empty: false,
      }),
      role: OSV.string({
        required: true,
        oneOf: ['owner', 'admin', 'subscriber'],
      }),
    }),
  }),
});

const obj = {
  users: [
    {
      username: 'tomato12',
      role: 'owner',
    },
    {
      username: 'apple56',
      role: 'admin',
    },
  ],
}

// asynchronous

const resultAsync = schema.validate(obj); // Promise<Data>

resultAsync
  .then(value => {
    console.log(value); // the validated value
  })
  .catch(e => {
    console.error(e); // ValidationError
  });

// synchronous

const resultSync = schema.validateSync(obj); // { value?: Data, error?: ValidationError, exec: () => Promise<Data> }

resultSync.exec()
  .then(value => {
    console.log(value); // the validated value

    /*
      {
        users: [
          {
            username: 'tomato12',
            role: 'owner',
          },
          {
            username: 'apple56',
            role: 'admin',
          },
        ],
      }
    */
  })
  .catch(e => {
    console.error(e); // ValidationError
  });

Installation

$ npm install --save osv

$ yarn add osv

API

This library is able to validate arrays, booleans, numbers, strings and of course objects.

By default, all specified keys of an object are optional. If a key is missing in an object but it has an object with additional keys as its value, a plain object will be added (see examples).

ObjectSchema

// types
type Validator<Data> = BaseSchemaType | ObjectSchema<Data>;

interface DefinitionObject<Data> {
  [path: string]: DefinitionObject<Data> | Validator<Data>;
}

type Definition<Data> = DefinitionObject<Data> | Validator<Data>;

Simply define your object schema and create an ObjectSchema instance.

import OSV, { OSVTypeRef } from 'osv';

interface User {
  id: string;
  username: string;
  role: 'owner' | 'admin' | 'subscriber';
  followers: {
    id: string;
  }[];
}

const definition: OSVTypeRef.Schema.Definition<User> = {
  username: OSV.string({
    required: true,
    empty: false,
  }),
  role: OSV.string({
    required: true,
    oneOf: ['owner', 'admin', 'subscriber'],
  }),
  followers: OSV.array({
    item: OSV.schema({
      id: OSV.string({ required: true, empty: false }),
    }),
  }),
};

const schema = OSV.schema<User>(definition);

After that you can validate objects. The validate method returns a Promise which resolves to the validated data if successful. Otherwise it rejects with a ValidationError.

The validateSync method returns an object with the following schema:

// Data: the data type of the validated value
type ValidationResult<Data> = {
  value?: Data;
  error?: ValidationError;
  exec: () => Promise<Data>;
}

const user1 = {
  username: 'tomato',
  role: 'owner',
  followers: [{
    id: 'apple',
  }],
};

// async

schema.validate(user1)
  .then((user) => {
    // use validated user
    const username = user.username;
  });

// sync

const result = schema.validate(user1);

console.log(result.value, result.error);

const user2 = {
  username: 'tomato',
  role: 'author',
  followers: [{
    id: 'apple',
  }],
};

schema.validate(user1)
  .catch((e) => {
    // ValidationError
    // e.message: 'There was an error while validating: string/not-allowed'
    // e.code: 'string/not-allowed'
    // e.path: 'role'
    // value: 'author'
  });

You can also omit specific paths during checking. These are just transfered onto the result value the way they are on the initial value. A whitelist and a blacklist are both optional.

interface ObjectSchemaValidationOptions {
  check?: {
    whitelist?: string[]; // all paths that should be checked,
    blacklist?: string[]; // all paths that shouldn't be checked
  };
}

schema.validate({
  username: 'tomato',
  role: 'owner',
  followers: [{
    id: undefined,
  }],
}, {
  check: {
    blacklist: ['followers.id'],
  }
}) // succeeds
  .then((user) => {
    // use validated user
    const username = user.username;
    const firstFollowerId = user.followers[0].id; // -> undefined
  });

BaseSchemaType - `ObjectSchema.Types.Base`

This is the schema type which all others extend from.