ts-interface-checker

What is ts-interface-checker?

The ts-interface-checker package is used to perform runtime type checking of objects against TypeScript interfaces. It works by using a companion compiler that generates runtime type information, which can then be used to validate objects at runtime, ensuring they conform to the specified TypeScript interfaces.

What are ts-interface-checker's main functionalities?

Type Checking

This feature allows you to check if an object matches a TypeScript interface. The example shows how to import the checkers, load the generated type information, and use the checker to validate an object.

const { createCheckers } = require('ts-interface-checker');
const ti = require('./type-info');

const { UserChecker } = createCheckers(ti);

const user = { name: 'Alice', age: 25 };
UserChecker.check(user);

Custom Error Reporting

This feature allows you to catch and report errors when an object fails to match the interface. The example demonstrates how to use a try-catch block to handle validation errors and output a custom error message.

const { createCheckers } = require('ts-interface-checker');
const ti = require('./type-info');

const { UserChecker } = createCheckers(ti);

try {
  const user = { name: 'Alice', age: '25' };
  UserChecker.check(user);
} catch (error) {
  console.error('Validation failed:', error.message);
}

Partial Checking

This feature enables checking for partial conformance to an interface, useful when only a subset of properties is required. The example shows how to check a partial object against a TypeScript interface.

const { createCheckers } = require('ts-interface-checker');
const ti = require('./type-info');

const { UserChecker } = createCheckers(ti);

const partialUser = { name: 'Bob' };
UserChecker.strictCheck(partialUser, 'Partial<User>');

Other packages similar to ts-interface-checker

io-ts

io-ts is a TypeScript library for runtime type checking. It uses a different approach than ts-interface-checker by defining types using codecs, which can then be used to validate data at runtime. It also supports functional programming patterns and is more focused on type transformations.

class-validator

class-validator is a validation library that allows using decorators to define validation rules on class properties. It differs from ts-interface-checker in that it works with class instances rather than plain objects and interfaces, and it integrates with class-transformer for object mapping.

prop-types

prop-types is a runtime type checking library for React props. It is similar to ts-interface-checker in that it validates objects at runtime, but it is specifically designed for React and uses a different syntax and set of validation functions.

ajv

ajv is a JSON schema validator that can validate data against JSON schemas. Unlike ts-interface-checker, which relies on TypeScript interfaces, ajv uses the JSON Schema standard for validation, making it more suitable for validating JSON data formats and inter-operable across different programming languages.

README

ts-interface-checker

Runtime library to validate data against TypeScript interfaces.

This package is the runtime support for validators created by ts-interface-builder. It allows validating data, such as parsed JSON objects received over the network, or parsed JSON or YAML files, to check if they satisfy a TypeScript interface, and to produce informative error messages if they do not.

Installation

npm install --save-dev ts-interface-builder
npm install --save ts-interface-checker

Usage

Suppose you have a TypeScript file defining an interface:

// foo.ts
interface Square {
  size: number;
  color?: string;
}

The first step is to generate some code for runtime checks:

`npm bin`/ts-interface-builder foo.ts

It produces a file like this:

// foo-ti.js
import * as t from "ts-interface-checker";

export const Square = t.iface([], {
  "size": "number",
  "color": t.opt("string"),
});
...

Now at runtime, to check if a value satisfies the Square interface:

import fooTI from "./foo-ti";
import {createCheckers} from "ts-interface-checker";

const {Square} = createCheckers(fooTI);

Square.check({size: 1});                  // OK
Square.check({size: 1, color: "green"});  // OK
Square.check({color: "green"});           // Fails with "value.size is missing"
Square.check({size: 4, color: 5});        // Fails with "value.color is not a string"

Note that ts-interface-builder is only needed for the build-time step, and ts-interface-checker is needed at runtime. That's why the recommendation is to npm-install the former using --save-dev flag and the latter using --save.

Checking method calls

If you have an interface with methods, you can validate method call arguments and return values:

// greet.ts
interface Greeter {
  greet(name: string): string;
}

After generating the runtime code, you can now check calls like:

import greetTI from "./greet-ti";
import {createCheckers} from "ts-interface-checker";

const {Greeter} = createCheckers(greetTI);

Greeter.methodArgs("greet").check(["Bob"]);     // OK
Greeter.methodArgs("greet").check([17]);        // Fails with "value.name is not a string"
Greeter.methodArgs("greet").check([]);          // Fails with "value.name is missing"

Greeter.methodResult("greet").check("hello");   // OK
Greeter.methodResult("greet").check(null);      // Fails with "value is not a string"

Type suites

If one type refers to a type defined in another file, you need to tell the interface checker about all type names when you call createCheckers(). E.g. given

// color.ts
export type Color = RGB | string;
export type RGB = [number, number, number];

// shape.ts
import {Color} from "./color";
export interface Square {
  size: number;
  color?: Color;
}

the produced files color-ti.ts and shape-ti.ts do not automatically refer to each other, but expect you to relate them in createCheckers() call:

import color from "./color-ti";
import shape from "./shape-ti";
import {createCheckers} from "ts-interface-checker";

const {Square} = createCheckers(shape, color);    // Pass in all required type suites.

Square.check({size: 1, color: [255,255,255]});

Strict checking

You may check that data contains no extra properties. Note that it is not generally recommended as it this prevents backward compatibility: if you add new properties to an interface, then older code with strict checks will not accept them.

Following on the example above:

Square.strictCheck({size: 1, color: [255,255,255], bg: "blue"});    // Fails with value.bg is extraneous
Square.strictCheck({size: 1, color: [255,255,255,0.5]});            // Fails with ...value.color[3] is extraneous

Type guards

Standard Checker objects do the type checking logic, but are unable to make the TypeScript compiler aware that an object of unknown type implements a certain interface.

Basic code:

const unk: unknown = {size: 1, color: "green"};
// Type is unknown, so TypeScript will not let you access the members.
console.log(unk.size); // Error: "Object is of type 'unknown'"

With a Checker available:

import fooTI from "./foo-ti";
import {createCheckers} from "ts-interface-checker";

const {Square} = createCheckers(fooTI);

const unk: unknown = {size: 1, color: "green"};

if (Square.test(unk)) {
  // unk does implement Square, but TypeScript is not aware of it.
  console.log(unk.size); // Error: "Object is of type 'unknown'"
}

To enable type guard functionality on the existing test, and strictTest functions, Checker objects should be cast to CheckerT<> using the appropriate type.

Using CheckerT<>:

import {Square} from "./foo";
import fooTI from "./foo-ti";
import {createCheckers, CheckerT} from "ts-interface-checker";

const {Square} = createCheckers(fooTI) as {Square: CheckerT<Square>};

const unk: unknown = {size: 1, color: "green"};

if (Square.test(unk)) {
  // TypeScript is now aware that unk implements Square, and allows member access.
  console.log(unk.size);
}

Type assertions

CheckerT<> will eventually support type assertions using the check and strictCheck functions, however, this feature is not yet fully working in TypeScript.