Modelgun 👾👾👾 🔫 - Data model and util auto-generation tool for TypeScript.

Overview

This is a tool to generate model classes and related utils with validations. What you should do is just define model structures and validation rules. Currently following files are generated by Modelgun.

• model: A class to represent a model which validation methods. For creating models with validations and holding model data and converting to plain JavaScript Objects.
• parser: Parser to convert any object to models. For checking API json payload etc.
• faker (alpha): Fake value generation tool for tests.

What is Modelgun🔫 originally?

Modelguns are Japanese replica or toy guns, which are usually made of zinc alloys or plastic materials. https://en.wikipedia.org/wiki/Modelguns

Motivation

We write data validation code so many times to build a system. But the work is very boring, time consuming and can be done by robots. What human should do is just design the structures of models and defines the validation rules. This tool make you free from the boring works and your life will be easier 😃.

How to

1. Install

npm

npm install --save-dev modelgun

yarn

yarn add --dev modelgun

2. Define models

Modelgun uses toml as the format of definition files. Use [MODEL_NAME].model.toml format for file names. [MODEL_NAME].model.ts, [MODEL_NAME].parser.ts and [MODEL_NAME].faker.ts (alpha) are generated for each model definition. See Model definition file spec section for more details.

3. Run

To run Modelgun generator, use following commands. Modelgun processes definition files in a specified directory and generates files in the same directory. (* Currently recursive checking of deep directories is not supported.)

In scripts in package.json, use

modelgun gen path/to/dir

To call script directory, use

./node_modules/modelgun/bin/cli.js gen path/to/dir

Examples

Example model definition files and generated files are stored here.

Supported languages

Currently only TypeScript is supported. If number of users increase, will consider to write JavaScript version.

Model definition file spec

Top level options

Model setting parameters.

description

Description about the model.

description = "This is the data structure to "

header

Header is placed at the top of auto-generated model file. You can write anything like constants, import statements and functions. This was introduced for custom validations. But you may find more cases. Please let me know if you find some.

header = "
// Importing
import anyFunc from \"./path/to/file\"

function customValidator (value: string) {
  ...
}
"

Generator

• parser: Whether to generate a parser file.
• faker: Whether to generate a faker file.

[generator]
parser = false
faker = true

Properties

props Defines model properties. Property name follows after props.

type

Type of property. Please use supported types.

[props.id]
type = "uuid"

[props.name]
type = "string"

[props.age]
type = "int"

Supported types

• boolean
• float
• int
• string
• uuid
• url
• email

null

Modelgun supports nullable for non-array properties. Specify null option like below.

[props.name]
type = "string"
null = true

Referring another model

Use "ref:[MODEL_NAME]"

TODO: Let users to use file path to a model.

validation

Depending on types of properties, modelgun auto generates validation code in model class. For some types, modelgun uses validator.js. Please add the npm package when you use types which requires it (e.g. uuid, email). Additionally, you can specify detailed validation.

Number

• min
• max
• candidates

String

• regex
• maxLength
• minLength
• candidates

Custom validator for any types

You can use your custom validator in files in your project. For those case, use header to import your functions or write directly in the generated model file. Throw InvalidPropertyError defined in the auto-generated model file for error cases.

header = "
import nameValidator from \"./path/to/file\"

function ageValidator (value: number) {
  if (value === 1000) {
    InvalidPropertyError(\"1000 years old is unbelievable.\")
  }
}
"

[props.name]
type = "string"
customValidator = "nameValidator"

[props.age]
type = "number"
customValidator = "ageValidator"