@jeremy-bankes/schema
TypeScript icon, indicating that this package has built-in type declarations

5.0.16 • Public • Published

Schema


TECHNICAL
noun a representation of a plan or theory in the form of an outline or model.

This library allows you to specify the schema for your data, and get compile time typings (help from your IDE) and runtime validation (throw errors if your data isn't in the right format).

With this library, everything is built from "Schema Primitives" (henceforth referred to as "Primitives"), out of the box, all of the JavaScript primitive types are implemented, but this set of Primitives can be extended for your use case. The JavaScript Date is also a supported Primitive. All of your schema objects will be built from this extensible set of Primitives.

Primitives
undefined
boolean
number
bigint
string
symbol
any
Date

Validation with a Primitive

import { Schema } from "@jeremy-bankes/schema";

const dataOfUnknownType: any = 123;
const dataOfNumberType = Schema.validate("number", dataOfUnknownType);

With Primitives, you can now start building more complex data structures. There are multiple way to represent different scenarios. This is done with the following constructs.

Constructs Structure TLDR
Meta { type: <Primitive or another Construct>, required: boolean, default?: <default value>, validate: <validator> } A meta object to describe the nature of a section within a schema.
Array [ <Primitive or another Construct> ] Represents an array within a schema.
Hierarcy { [ <Any keys needed> ]: <Primitive or another Construct> } Represents nested object hierarcy within a schema.
Compound [ <Primitive or another Construct>, "or", <Primitive or another Construct> ] Represents union types. Currently can be used with "and" or "or".
Dynamic { $: <Primitive or another Construct> } Represents a object with dynamic keys. (Produces the TypeScript type { [Key: string]: any })

Creating Schema

These primitives and constructs can come together to allow you to define your own schema.

import { Schema } from "@jeremy-bankes/schema";

/* ↓ Schema Definition */
const PersonSchema = Schema.build({
    name: /* ← Hierarcy ↓ */ {
        first: /* ← Meta ↓ */ {
            type: "string" /* ← Primitive */,
            required: true
        },
        middle: { type: "string", required: false },
        last: { type: "string", required: true }
    },
    favoriteNumber: ["bigint" /* ← Primitive */, "or", "number"] /* ← Compound */,
    age: {
        type: "number",
        required: false,
        validate: (age: number, rawData: any) => {
            const birthDate = Schema.validate({ type: "Date", required: false }, rawData.birthDate);
            if (birthDate === undefined) {
                return undefined;
            } else {
                const now = new Date();
                const expectedAge = (now.getTime() - birthDate.getTime()) / 1000 / 60 / 60 / 24 / 365.25;
                Schema.assert(Math.floor(expectedAge) === age);
                return age;
            }
        }
    },
    birthDate: {
        type: "Date",
        required: false
    }
});

const WeaponSchema = Schema.build({
    damage: "number" /* ← Primitive */,
    owners: ["string"] /* ← Array */,
    forged: /* ← Meta ↓ */ {
        type: "Date",
        required: false
    },
    affixtures: /* ← Dynamic ↓ */ {
        $: "string"
    }
});

Using Schema

With the schema definitions created, they can be used to validate data where the types are unknown. This data might come from fetch requests, information read from disk, or even JavaScript libraries that don't export their own type definitions, just to name a few examples.

import fs from "fs";

const personData = JSON.parse(fs.readFileSync("person.json", "utf8"));
const person = Schema.validate(PersonSchema, personData);

person now has the following compile-time type annotation based on PersonSchema.

const person: {
    name: {
        first: string;
        last: string;
    } & {
        middle?: string | undefined;
    };
    favoriteNumber: number | bigint;
}

At runtime, the object parsed from the person.json file will be validated to match this generated typing. If it does not match, a Schema.ValidationError will be thrown.

To-Do

  • Documentation:
    • Extending Primitive set.
    • Defining a default value / function in Meta
    • Defining a validator in Meta

Readme

Keywords

Package Sidebar

Install

npm i @jeremy-bankes/schema

Weekly Downloads

13

Version

5.0.16

License

GPL-3.0-or-later

Unpacked Size

61.5 kB

Total Files

5

Last publish

Collaborators

  • jeremy-bankes