Never Print Magazines


    0.1.0-beta • Public • Published


    An integer isn't a data type, it's a compiler hint.

    Models are sophisticated entities, not bundles of generic primitives.

    If you clone the project you can see the following in a browser by running `npm run docs' and opening the debug console in your browser.

    Define a model to match reality:

    let EF = EntityForge
    let UserForge = EF.obj({
        uuid: EF.string().minLength(20).maxLength(20).ascii(),
        email: EF.string().minLength(5).maxLength(255),

    Unless we specify otherwise, we can't create or modify the created model type in a way that would cause it to become invalid:

    let NullableUserModel = UserForge.asNewable()
    let example = new NullableUserModel()
    console.log("UUID: ", example.uuid) // null
    console.log("Email: ", // null
    console.log("MemberSince: ", example.memberSince) // 2016
    example.uuid = "-JhLeOlGIEjaIOFHR0xd" = ""
    try {
        example.uuid = example.uuid.substring(0, 10)
    } catch (e) {
        console.log("I'm sorry dave....") // Nope, not allowed.
        console.log("Validation errors provide a cause", e.cause) restrictions
        console.log("To be clear, the messaging system needs some work. The cause message is: ", e.cause.minLength.message) // @restriction.minLength
    console.log("UUID wasn't modified by the attempt to set it to illegal value:", example.uuid) // "-JhLeOlGIEjaIOFHR0xd" --- value not modified if invalid.

    Take care to recognize that there is a difference between setting minLength to zero and not allowing null. A null string is still valid even if minLength is set to zero.

    Generating data

    One advantage of specifying our data type in detail is that we can use that specification to do cool things. Like generate semi-random instances of our models:

    let randomUser = UserForge.gen()
    console.log(randomUser) // This instance will be as valid (or invalid) as your Forge definition constraints allow.

    There is still work to be done on the data generation side. Of the missing functionality, the most important is the handling of cases that are hard to code for explicitly, such as string matching on a regex.

    Built In Forges

    The initial batch of forges cover only the basic primitive types, with Arrays/Lists 'on the way'.

    let BiggerUserForge = EF.obj({
        uuid: EF.string().minLength(20).maxLength(20).ascii(),
        email: EF.string().minLength(5).maxLength(255),
        karmaScore: EF.number().min(0).max(1).initTo(0.5),
        groups: EF.enumeration().values(["admin", "guest", "subscriber", "paid-member"]),
        contact: EF.obj({
            surname: EF.string().minLength(1).maxLength(255).ascii(),
            forename: EF.string().minLength(1).maxLength(255).ascii(),
            addressLine1: EF.string().minLength(1).maxLength(255).ascii(),
            addressLine2: EF.string().minLength(1).maxLength(255).ascii(),
            postcode: EF.string().minLength(3).maxLength(25).ascii(),
    let biggerUser = BiggerUserForge.gen()
    console.log("A randomly generated 'BiggerUser:", biggerUser)


    Getting started

    npm install typings --global
    npm install typescript --global
    npm install
    npm run
    # In another console tab
    npm run serve.test


    npm i @tangential/entityforge

    DownloadsWeekly Downloads






    Last publish


    • tangential