dcos/raml-validator-loader

Name: raml-validator-loader

Owner: DC/OS

Description: A webpack plugin that converts RAML rules into pure javascript-only validation routines

Created: 2017-01-09 16:17:08.0

Updated: 2017-05-12 07:28:56.0

Pushed: 2018-01-29 14:06:21.0

Homepage: null

Size: 158

Language: JavaScript

GitHub Committers

User	Most Recent Commit	# Commits

Other Committers

User	Email	Most Recent Commit	# Commits

README

raml-validator-loader

A webpack plugin that converts RAML rules into pure javascript-only validation routines

Usage

rt TypeValidators from './ramls/myTypes.raml';

he raml-validator-loader will convert the RAML document into an object with
 validation function for every type defined in your RAML.

userInput = read_user_input();
validationErrors = TypeValidators.MyType(userInput);

isplay errors
dationErrors.forEach((error) => {
nsole.log('Error message: /' + error.path.join('/'));
nsole.log('  Error path : ' + error.message);

Customizing validator

It is possible to customize the validator instance in order for example to provide custom error messages:

rt TypeValidators from './ramls/myTypes.raml';


ou can clone the default validator instance and customize it

t CustomTypeValidators = TypeValidators.clone({
'errorMessages': {
    // You can provide a string override
    TYPE_NOT_OBJECT: 'Expecting an object here',

    // Or a function override if you want more control over
    // the error message
    STRING_PATTERN: (templateVars, path) => {
        if (path[0] == 'name') {
            return 'Name must only contain letters';
        }

        return `Must match '${templateVars.pattern}'`;
    }
}

Installation

First install the module

install --save-dev raml-validator-loader

And then use it in your webpack.config.js like so:

module: {
    loaders: [
        { test: /\.raml$/, loader: "raml-validator-loader" }
    ]
}

API Reference

When you are importing a .raml document you will end-up loading an instance of a RAMLValidator class and all the RAML types are exposed as instance property functions.

rt TypeValidators from './ramls/myTypes.raml';

RAMLValidator Class

A default instance of RAMLValidator class is returned when importing a RAML file. You don't have access to it's constructor directly.

Function .clone([config])

Creates a new instance of RAMLValidator allowing further customisations to the configuration. Parameters not overriden through the configuration will be kept intact.

t CustomValidator = TypeValidators.clone({
errorMessages: { ... }

Parameters

• config - [Optional] An object with the new configuration parameters to pass down to the new validator instance. Accepts the following properties:
  • errorMessages : An object with custom error messages. The value to the error message can either be a string, or a function with the following signature: (messageVariables, path) that returns an error string. For the full list of error messages you can refer to the end of this documentation.

Returns

• RAMLValidator - Returns a new RAMLValidator instance

Function .<TypeName>(input)

For each type in the RAML document, a validation function will be created with the above signature. You can call this function, passing it the data you want to validate and it will return an array with the errors found in it.

t errors = TypeValidators.SomeRAMLType(userData);
errors.length > 0) {
throw new TypeError('You must provide some valid data');

Parameters

• input - The data to validate. Can be anything

Returns

• [RAMLError] - Returns an array of RAMLError instances, one for every error encountered in the input data.

le.exports = {

*
 For every type in the RAML document, an equivalent validation function
 will be generated by the loader.
/
mlTypeName: function( validatorInput ) {

...

// Each validation function will return an array of RAMLError objects
// with the information for the validation errors occured.
return [ RAMLError() ];

RAMLError Class

This class is instantiated by the type validator function and it contains the information to locate and describe an error in the type.

t errors = TypeValidators.SomeRAMLType(userData);
errors.length > 0) {
const error = errors[0];
console.log('Error path = ', error.path);
console.log('Error message = ', error.message);

Property .path

• Array - Returns the path in the object where the validation error ocurred, as an array of path components. For example: ['objects', 0, 'name']

Property .type

• String - Returns the type of the error as a string. For example "PROP_IS_MISSING"

Property .variables

• Object - Returns an object with the template variable values for the error message. For example {value: 2} for the ITEMS_MAX error type, in order to be able to compose the dynamic result of "Must contain at most 2 items in the array".

Property .message

• String - Returns the human-readable description of the error. For example: Missing property 'name'.

Error Messages

The following error messages are used by the RAML validator. You can override them using the .clone({errorMessages: ...}) function.

Error Type	Default Value
ENUM	Must be one of {{values}}
ITEMS_MAX	Must contain at most {{value}} items in the array
ITEMS_MIN	Must contain at least {{value}} items in the array
ITEMS_UNIQUE	Must contain only unique items
LENGTH_MAX	Must be at most {{value}} characters long
LENGTH_MIN	Must be at least {{value}} characters long
NUMBER_MAX	Must be smaller than or equal to {{value}}
NUMBER_MIN	Must be bigger than or equal to {{value}}
NUMBER_MULTIPLEOF	Must be multiple of {{value}}
NUMBER_TYPE	Must be of type `{{type}}`
PROPS_MAX	Must contain at most {{value}} properties
PROPS_MIN	Must contain at least {{value}} properties
PROP_ADDITIONAL_PROPS	Contains extraneous property `{{name}}`
PROP_IS_MISSING	Must be defined
PROP_MISSING	Must define property `{{name}}`
PROP_MISSING_MATCH	Must contain a property that matches `{{pattern}}`
STRING_PATTERN	Must match the pattern `{{pattern}}`
TYPE_NOT_ARRAY	Must be an array
TYPE_NOT_BOOLEAN	Must be a boolean value
TYPE_NOT_DATETIME	Must be a date/time string
TYPE_NOT_INTEGER	Must be an integer number
TYPE_NOT_NULL	Must be null
TYPE_NOT_NUMBER	Must be a number
TYPE_NOT_OBJECT	Must be an object
TYPE_NOT_STRING	Must be a string

Function Overrides

If you want more control over the error message, you can use a function instead of string for the error message. The parameters passed to the function are the messageVariables that provide some contextualization to the error message, and the path of the error.

For example:

t CustomTypeValidators = TypeValidators.clone({
'errorMessages': {
    STRING_PATTERN: (templateVars, path) => {
        switch (path.join('.')) {
            case 'name':
                return 'Name must only contain numbers and letters'

            case 'file.name':
                return 'All file names must only contain numbers and letters'

            default:
                return `Must match ${templateVars.pattern}`;
        }
    }
}

Work in progress

The following facets are not yet implemented:

• discriminator in Unions
• discriminatorValue Unions

The following types are not yet implemented:

• date-only
• time-only
• datetime-only
• file

The following concepts are not yet implemented:

• Multiple Inheritance