A JavaScript object merge/validation utility where you can define a different merge and validation strategy for each key. This is helpful when you need to validate complex data structures and then merge them in a way that is more complex than Object.assign()
. This is used in the @eslint/config-array
package but can also be used on its own.
For Node.js and compatible runtimes:
npm install @eslint/object-schema
# or
yarn add @eslint/object-schema
# or
pnpm install @eslint/object-schema
# or
bun install @eslint/object-schema
For Deno:
deno add @eslint/object-schema
Import the ObjectSchema
constructor:
// using ESM
import { ObjectSchema } from "@eslint/object-schema";
// using CommonJS
const { ObjectSchema } = require("@eslint/object-schema");
const schema = new ObjectSchema({
// define a definition for the "downloads" key
downloads: {
required: true,
merge(value1, value2) {
return value1 + value2;
},
validate(value) {
if (typeof value !== "number") {
throw new Error("Expected downloads to be a number.");
}
},
},
// define a strategy for the "versions" key
version: {
required: true,
merge(value1, value2) {
return value1.concat(value2);
},
validate(value) {
if (!Array.isArray(value)) {
throw new Error("Expected versions to be an array.");
}
},
},
});
const record1 = {
downloads: 25,
versions: ["v1.0.0", "v1.1.0", "v1.2.0"],
};
const record2 = {
downloads: 125,
versions: ["v2.0.0", "v2.1.0", "v3.0.0"],
};
// make sure the records are valid
schema.validate(record1);
schema.validate(record2);
// merge together (schema.merge() accepts any number of objects)
const result = schema.merge(record1, record2);
// result looks like this:
const result = {
downloads: 75,
versions: ["v1.0.0", "v1.1.0", "v1.2.0", "v2.0.0", "v2.1.0", "v3.0.0"],
};
Instead of specifying a merge()
method, you can specify one of the following strings to use a default merge strategy:
"assign"
- useObject.assign()
to merge the two values into one object."overwrite"
- the second value always replaces the first."replace"
- the second value replaces the first if the second is notundefined
.
For example:
const schema = new ObjectSchema({
name: {
merge: "replace",
validate() {},
},
});
Instead of specifying a validate()
method, you can specify one of the following strings to use a default validation strategy:
"array"
- value must be an array."boolean"
- value must be a boolean."number"
- value must be a number."object"
- value must be an object."object?"
- value must be an object or null."string"
- value must be a string."string!"
- value must be a non-empty string.
For example:
const schema = new ObjectSchema({
name: {
merge: "replace",
validate: "string",
},
});
If you are defining a key that is, itself, an object, you can simplify the process by using a subschema. Instead of defining merge()
and validate()
, assign a schema
key that contains a schema definition, like this:
const schema = new ObjectSchema({
name: {
schema: {
first: {
merge: "replace",
validate: "string",
},
last: {
merge: "replace",
validate: "string",
},
},
},
});
schema.validate({
name: {
first: "n",
last: "z",
},
});
If the merge strategy for a key returns undefined
, then the key will not appear in the final object. For example:
const schema = new ObjectSchema({
date: {
merge() {
return undefined;
},
validate(value) {
Date.parse(value); // throws an error when invalid
},
},
});
const object1 = { date: "5/5/2005" };
const object2 = { date: "6/6/2006" };
const result = schema.merge(object1, object2);
console.log("date" in result); // false
If you'd like the presence of one key to require the presence of another key, you can use the requires
property to specify an array of other properties that any key requires. For example:
const schema = new ObjectSchema();
const schema = new ObjectSchema({
date: {
merge() {
return undefined;
},
validate(value) {
Date.parse(value); // throws an error when invalid
},
},
time: {
requires: ["date"],
merge(first, second) {
return second;
},
validate(value) {
// ...
},
},
});
// throws error: Key "time" requires keys "date"
schema.validate({
time: "13:45",
});
In this example, even though date
is an optional key, it is required to be present whenever time
is present.
Apache 2.0
The following companies, organizations, and individuals support ESLint's ongoing maintenance and development. Become a Sponsor to get your logo on our READMEs and website.
Technology sponsors allow us to use their products and services for free as part of a contribution to the open source ecosystem and our work.