rules

Validate objects against predefined rules (node.js).

rules (node.js)

A tiny rules framework that can be used to validate any value, either by you creating a rules/schema object or applying validations to a single value. In both cases a fluent interface is used.

You create an object to declare the rules/invariants you want to apply (something akin to a schema). A fluent interface makes it easy to specify the invariants for each property:

var nameRules = {
    first  : mustBe().populated().string({ minLength: 5, maxLength: 20}), [1]
    second : mustBe().populated().string({ minLength: 5, maxLength: 20}),
}
 
var personRules = {
    name:        nameRules,
    weight:      mustBe().populated().numeric({min : 0, max: 130}),
    dateOfBirthfunction() { 
        this.populated().date({ before: now.subtract("years", 1) }); 
    } [2]
}

As shown you can access this fluent interface using two approaches:

  • [1] mustBe() - Acts as the entry point to the fluent interface.
  • [2] function - 'this' inside the function being the entry point to the fluent interface.

The function based approach is designed primarily for use with CoffeeScript:

# This schema is not showing how to validate a real address, it just shows a few validators 
addressRules = {
  streetOne: mustBe().populated()
  streetTwo: -> @.populated().string( minLength: 10maxLength : 50 ) 
  streetThree: -> @.populated().string( minLength : 10maxLength: 50) 
  town: -> @.populated()
  postCode: -> @.populated().matchFor(/.../)
}

The same validators are available for use validating individual values:

var doSomeStuff = function(nameage) {
    ensure(name).populated().string();
    ensure(age, "age").integer();
 
    ...
}

You trigger validation using:

result = rules.apply(person, personRules)

The returned object has the per-property details of any validation failures, e.g.:

{ 
    name: { 
        first: { 
            message: 'The value must be populated.',
            type: 'not_populated',
            value: '' 
        },
        second: { 
            message: 'The value must be populated.',
            type: 'not_populated',
            value:  } 
        },
    weight: { 
        message: 'The value must be populated.',
        type: 'not_populated',
        value:  
    } 
}

Note in this case both the first name (e.g. person.name.first) and second name (person.name.second) needed to be populated, along with the weight.

The framework comes with several validators, to understand them further you may want to run the examples.

  • populated - Checks the value is not null, , "", or an empty array.
  • array
  • numeric - Optionally you can also pass in object with min and/or max values
  • integer
  • matchFor - You can pass in an object with pattern and optionally flags, alternatively you can pass in the RegExp object to use.
  • date - Optionally you can specify that the date must be before and/or after specified dates. To make this easier you use now.add or now.subtract to specify the dates to use for before/after.
  • string - Optionally you can pass in minLength and/or maxLength.

The project comes with examples in the examples directory.

mocha -R spec spec/testFixture spec/ --recursive
  • Numeric validators - >, <, >=, <=
  • Date validator - Support now()
  • boolean validator
  • enum style validator - valueIn(list), valueNotIn(list).
  • Potentially UMD support
  • Trying to apply multiple of same validator, multiple type validators (integer and string, numeric and boolean), regex with anything other than string
  • Ensure interface e.g. ensure(5).populate().numeric(), allowing direct validation of single values
  • API for throwing
  • Cyclical rules objects warning
  • Numeric validator - failing if passed "15.5"