[javascript] Interface type check with Typescript

This question is the direct analogon to Class type check with TypeScript

I need to find out at runtime if a variable of type any implements an interface. Here's my code:

interface A{
    member:string;
}

var a:any={member:"foobar"};

if(a instanceof A) alert(a.member);

If you enter this code in the typescript playground, the last line will be marked as an error, "The name A does not exist in the current scope". But that isn't true, the name does exist in the current scope. I can even change the variable declaration to var a:A={member:"foobar"}; without complaints from the editor. After browsing the web and finding the other question on SO I changed the interface to a class but then I can't use object literals to create instances.

I wondered how the type A could vanish like that but a look at the generated javascript explains the problem:

var a = {
    member: "foobar"
};
if(a instanceof A) {
    alert(a.member);
}

There is no representation of A as an interface, therefore no runtime type checks are possible.

I understand that javascript as a dynamic language has no concept of interfaces. Is there any way to type check for interfaces?

The typescript playground's autocompletion reveals that typescript even offers a method implements. How can I use it ?

I found an example from @progress/kendo-data-query in file filter-descriptor.interface.d.ts

Checker

declare const isCompositeFilterDescriptor: (source: FilterDescriptor | CompositeFilterDescriptor) => source is CompositeFilterDescriptor;

Example usage

const filters: Array<FilterDescriptor | CompositeFilterDescriptor> = filter.filters;

filters.forEach((element: FilterDescriptor | CompositeFilterDescriptor) => {
    if (isCompositeFilterDescriptor(element)) {
        // element type is CompositeFilterDescriptor
    } else {
        // element type is FilterDescriptor
    }
});

Type guards in Typescript:

TS has type guards for this purpose. They define it in the following manner:

Some expression that performs a runtime check that guarantees the type in some scope.

This basically means that the TS compiler can narrow down the type to a more specific type when it has sufficient information. For example:

function foo (arg: number | string) {
    if (typeof arg === 'number') {
        // fine, type number has toFixed method
        arg.toFixed()
    } else {
        // Property 'toFixed' does not exist on type 'string'. Did you mean 'fixed'?
        arg.toFixed()
        // TSC can infer that the type is string because 
        // the possibility of type number is eliminated at the if statement
    }
}

To come back to your question, we can also apply this concept of type guards to objects in order to determine their type. To define a type guard for objects, we need to define a function whose return type is a type predicate. For example:

interface Dog {
    bark: () => void;
}

// The function isDog is a user defined type guard
// the return type: 'pet is Dog' is a type predicate, 
// it determines whether the object is a Dog
function isDog(pet: object): pet is Dog {
  return (pet as Dog).bark !== undefined;
}

const dog: any = {bark: () => {console.log('woof')}};

if (isDog(dog)) {
    // TS now knows that objects within this if statement are always type Dog
    // This is because the type guard isDog narrowed down the type to Dog
    dog.bark();
}

It's now possible, I just released an enhanced version of the TypeScript compiler that provides full reflection capabilities. You can instantiate classes from their metadata objects, retrieve metadata from class constructors and inspect interface/classes at runtime. You can check it out here

Usage example:

In one of your typescript files, create an interface and a class that implements it like the following:

interface MyInterface {
    doSomething(what: string): number;
}

class MyClass implements MyInterface {
    counter = 0;

    doSomething(what: string): number {
        console.log('Doing ' + what);
        return this.counter++;
    }
}

now let's print some the list of implemented interfaces.

for (let classInterface of MyClass.getClass().implements) {
    console.log('Implemented interface: ' + classInterface.name)
}

compile with reflec-ts and launch it:

$ node main.js
Implemented interface: MyInterface
Member name: counter - member kind: number
Member name: doSomething - member kind: function

See reflection.d.ts for Interface meta-type details.

UPDATE: You can find a full working example here

I would like to point out that TypeScript does not provide a direct mechanism for dynamically testing whether an object implements a particular interface.

Instead, TypeScript code can use the JavaScript technique of checking whether an appropriate set of members are present on the object. For example:

var obj : any = new Foo();

if (obj.someInterfaceMethod) {
    ...
}

typescript 2.0 introduce tagged union

Typescript 2.0 features

interface Square {
    kind: "square";
    size: number;
}

interface Rectangle {
    kind: "rectangle";
    width: number;
    height: number;
}

interface Circle {
    kind: "circle";
    radius: number;
}

type Shape = Square | Rectangle | Circle;

function area(s: Shape) {
    // In the following switch statement, the type of s is narrowed in each case clause
    // according to the value of the discriminant property, thus allowing the other properties
    // of that variant to be accessed without a type assertion.
    switch (s.kind) {
        case "square": return s.size * s.size;
        case "rectangle": return s.width * s.height;
        case "circle": return Math.PI * s.radius * s.radius;
    }
}

Here's the solution I came up with using classes and lodash: (it works!)

// TypeChecks.ts
import _ from 'lodash';

export class BakedChecker {
    private map: Map<string, string>;

    public constructor(keys: string[], types: string[]) {
        this.map = new Map<string, string>(keys.map((k, i) => {
            return [k, types[i]];
        }));
        if (this.map.has('__optional'))
            this.map.delete('__optional');
    }

    getBakedKeys() : string[] {
        return Array.from(this.map.keys());
    }

    getBakedType(key: string) : string {
        return this.map.has(key) ? this.map.get(key) : "notfound";
    }
}

export interface ICheckerTemplate {
    __optional?: any;
    [propName: string]: any;
}

export function bakeChecker(template : ICheckerTemplate) : BakedChecker {
    let keys = _.keysIn(template);
    if ('__optional' in template) {
        keys = keys.concat(_.keysIn(template.__optional).map(k => '?' + k));
    }
    return new BakedChecker(keys, keys.map(k => {
        const path = k.startsWith('?') ? '__optional.' + k.substr(1) : k;
        const val = _.get(template, path);
        if (typeof val === 'object') return val;
        return typeof val;
    }));
}

export default function checkType<T>(obj: any, template: BakedChecker) : obj is T {
    const o_keys = _.keysIn(obj);
    const t_keys = _.difference(template.getBakedKeys(), ['__optional']);
    return t_keys.every(tk => {
        if (tk.startsWith('?')) {
            const ak = tk.substr(1);
            if (o_keys.includes(ak)) {
                const tt = template.getBakedType(tk);
                if (typeof tt === 'string')
                    return typeof _.get(obj, ak) === tt;
                else {
                    return checkType<any>(_.get(obj, ak), tt);
                }
            }
            return true;
        }
        else {
            if (o_keys.includes(tk)) {
                const tt = template.getBakedType(tk);
                if (typeof tt === 'string')
                    return typeof _.get(obj, tk) === tt;
                else {
                    return checkType<any>(_.get(obj, tk), tt);
                }
            }
            return false;
        }
    });
}

custom classes:

// MyClasses.ts

import checkType, { bakeChecker } from './TypeChecks';

class Foo {
    a?: string;
    b: boolean;
    c: number;

    public static _checker = bakeChecker({
        __optional: {
            a: ""
        },
        b: false,
        c: 0
    });
}

class Bar {
    my_string?: string;
    another_string: string;
    foo?: Foo;

    public static _checker = bakeChecker({
        __optional: {
            my_string: "",
            foo: Foo._checker
        },
        another_string: ""
    });
}

to check the type at runtime:

if (checkType<Bar>(foreign_object, Bar._checker)) { ... }

Working with string literals is difficult because if you want to refactor you method or interface names then it could be possible that your IDE don't refactor these string literals. I provide you mine solution which works if there is at least one method in the interface

export class SomeObject implements interfaceA {
  public methodFromA() {}
}

export interface interfaceA {
  methodFromA();
}

Check if object is of type interface:

const obj = new SomeObject();
const objAsAny = obj as any;
const objAsInterfaceA = objAsAny as interfaceA;
const isObjOfTypeInterfaceA = objAsInterfaceA.methodFromA != null;
console.log(isObjOfTypeInterfaceA)

Note: We will get true even if we remove 'implements interfaceA' because the method still exists in the SomeObject class

Here's another option: the module ts-interface-builder provides a build-time tool that converts a TypeScript interface into a runtime descriptor, and ts-interface-checker can check if an object satisfies it.

For OP's example,

interface A {
  member: string;
}

You'd first run ts-interface-builder which produces a new concise file with a descriptor, say, foo-ti.ts, which you can use like this:

import fooDesc from './foo-ti.ts';
import {createCheckers} from "ts-interface-checker";
const {A} = createCheckers(fooDesc);

A.check({member: "hello"});           // OK
A.check({member: 17});                // Fails with ".member is not a string" 

You can create a one-liner type-guard function:

function isA(value: any): value is A { return A.test(value); }

How about User-Defined Type Guards? https://www.typescriptlang.org/docs/handbook/advanced-types.html

interface Bird {
    fly();
    layEggs();
}

interface Fish {
    swim();
    layEggs();
}

function isFish(pet: Fish | Bird): pet is Fish { //magic happens here
    return (<Fish>pet).swim !== undefined;
}

// Both calls to 'swim' and 'fly' are now okay.

if (isFish(pet)) {
    pet.swim();
}
else {
    pet.fly();
}

same as above where user-defined guards were used but this time with an arrow function predicate

interface A {
  member:string;
}

const check = (p: any): p is A => p.hasOwnProperty('member');

var foo: any = { member: "foobar" };
if (check(foo))
    alert(foo.member);

export interface ConfSteps {
    group: string;
    key: string;
    steps: string[];
}

private verify(): void {
    const obj = `{
      "group": "group",
      "key": "key",
      "steps": [],
      "stepsPlus": []
    } `;
    if (this.implementsObject<ConfSteps>(obj, ['group', 'key', 'steps'])) {
      console.log(`Implements ConfSteps: ${obj}`);
    }
  }

private objProperties: Array<string> = [];

private implementsObject<T>(obj: any, keys: (keyof T)[]): boolean {
    JSON.parse(JSON.stringify(obj), (key, value) => {
      this.objProperties.push(key);
    });
    for (const key of keys) {
      if (!this.objProperties.includes(key.toString())) {
        return false;
      }
    }
    this.objProperties = null;
    return true;
  }

Based on Fenton's answer, here's my implementation of a function to verify if a given object has the keys an interface has, both fully or partially.

Depending on your use case, you may also need to check the types of each of the interface's properties. The code below doesn't do that.

function implementsTKeys<T>(obj: any, keys: (keyof T)[]): obj is T {
    if (!obj || !Array.isArray(keys)) {
        return false;
    }

    const implementKeys = keys.reduce((impl, key) => impl && key in obj, true);

    return implementKeys;
}

Example of usage:

interface A {
    propOfA: string;
    methodOfA: Function;
}

let objectA: any = { propOfA: '' };

// Check if objectA partially implements A
let implementsA = implementsTKeys<A>(objectA, ['propOfA']);

console.log(implementsA); // true

objectA.methodOfA = () => true;

// Check if objectA fully implements A
implementsA = implementsTKeys<A>(objectA, ['propOfA', 'methodOfA']);

console.log(implementsA); // true

objectA = {};

// Check again if objectA fully implements A
implementsA = implementsTKeys<A>(objectA, ['propOfA', 'methodOfA']);

console.log(implementsA); // false, as objectA now is an empty object

Because the type is unknown at run-time, I wrote code as follows to compare the unknown object, not against a type, but against an object of known type:

1. Create a sample object of the right type
2. Specify which of its elements are optional
3. Do a deep compare of your unknown object against this sample object

Here's the (interface-agnostic) code I use for the deep compare:

function assertTypeT<T>(loaded: any, wanted: T, optional?: Set<string>): T {
  // this is called recursively to compare each element
  function assertType(found: any, wanted: any, keyNames?: string): void {
    if (typeof wanted !== typeof found) {
      throw new Error(`assertType expected ${typeof wanted} but found ${typeof found}`);
    }
    switch (typeof wanted) {
      case "boolean":
      case "number":
      case "string":
        return; // primitive value type -- done checking
      case "object":
        break; // more to check
      case "undefined":
      case "symbol":
      case "function":
      default:
        throw new Error(`assertType does not support ${typeof wanted}`);
    }
    if (Array.isArray(wanted)) {
      if (!Array.isArray(found)) {
        throw new Error(`assertType expected an array but found ${found}`);
      }
      if (wanted.length === 1) {
        // assume we want a homogenous array with all elements the same type
        for (const element of found) {
          assertType(element, wanted[0]);
        }
      } else {
        // assume we want a tuple
        if (found.length !== wanted.length) {
          throw new Error(
            `assertType expected tuple length ${wanted.length} found ${found.length}`);
        }
        for (let i = 0; i < wanted.length; ++i) {
          assertType(found[i], wanted[i]);
        }
      }
      return;
    }
    for (const key in wanted) {
      const expectedKey = keyNames ? keyNames + "." + key : key;
      if (typeof found[key] === 'undefined') {
        if (!optional || !optional.has(expectedKey)) {
          throw new Error(`assertType expected key ${expectedKey}`);
        }
      } else {
        assertType(found[key], wanted[key], expectedKey);
      }
    }
  }

  assertType(loaded, wanted);
  return loaded as T;
}

Below is an example of how I use it.

In this example I expect the JSON contains an array of tuples, of which the second element is an instance of an interface called User (which has two optional elements).

TypeScript's type-checking will ensure that my sample object is correct, then the assertTypeT function checks that the unknown (loaded from JSON) object matches the sample object.

export function loadUsers(): Map<number, User> {
  const found = require("./users.json");
  const sample: [number, User] = [
    49942,
    {
      "name": "ChrisW",
      "email": "[email protected]",
      "gravatarHash": "75bfdecf63c3495489123fe9c0b833e1",
      "profile": {
        "location": "Normandy",
        "aboutMe": "I wrote this!\n\nFurther details are to be supplied ..."
      },
      "favourites": []
    }
  ];
  const optional: Set<string> = new Set<string>(["profile.aboutMe", "profile.location"]);
  const loaded: [number, User][] = assertTypeT(found, [sample], optional);
  return new Map<number, User>(loaded);
}

You could invoke a check like this in the implementation of a user-defined type guard.

TypeGuards

interface MyInterfaced {
    x: number
}

function isMyInterfaced(arg: any): arg is MyInterfaced {
    return arg.x !== undefined;
}

if (isMyInterfaced(obj)) {
    (obj as MyInterfaced ).x;
}

In TypeScript 1.6, user-defined type guard will do the job.

interface Foo {
    fooProperty: string;
}

interface Bar {
    barProperty: string;
}

function isFoo(object: any): object is Foo {
    return 'fooProperty' in object;
}

let object: Foo | Bar;

if (isFoo(object)) {
    // `object` has type `Foo`.
    object.fooProperty;
} else {
    // `object` has type `Bar`.
    object.barProperty;
}

And just as Joe Yang mentioned: since TypeScript 2.0, you can even take the advantage of tagged union type.

interface Foo {
    type: 'foo';
    fooProperty: string;
}

interface Bar {
    type: 'bar';
    barProperty: number;
}

let object: Foo | Bar;

// You will see errors if `strictNullChecks` is enabled.
if (object.type === 'foo') {
    // object has type `Foo`.
    object.fooProperty;
} else {
    // object has type `Bar`.
    object.barProperty;
}

And it works with switch too.

You can validate a TypeScript type at runtime using ts-validate-type, like so (does require a Babel plugin though):

const user = validateType<{ name: string }>(data);