Spected is a low level validation library for validating objects against defined validation rules. Framework specific validation libraries can be built upon spected, leveraging the spected appraoch of separating the speciific input from any validation. Furthermore it can be used to verify the validity deeply nested objects, f.e. server side validation of data or client side validation of JSON objects. Spected can also be used to validate Form inputs etc.
Install Spected via npm or yarn.
npm install --save spected
Spected takes care of running your predicate functions against provided inputs, by separating the validation from the input. For example we would like to define a number of validation rules for two inputs, name and random.
const validationRules = {
name: [
[ isGreaterThan(5),
`Minimum Name length of 6 is required.`
],
],
random: [
[ isGreaterThan(7), 'Minimum Random length of 8 is required.' ],
[ hasCapitalLetter, 'Random should contain at least one uppercase letter.' ],
]
}And imagine this is our input data.
const inputData = { name: 'abcdef', random: 'z'}We would like to have a result that displays any possible errors.
Calling validate spected(validationRules, inputData)
should return
{name: true,
random: [
'Minimum Random length of 8 is required.',
'Random should contain at least one uppercase letter.'
]}You can also pass in an array of items as an input and validate that all are valid. You need to write the appropriate function to handle any specific case.
const userSpec = [
[
items => all(isLengthGreaterThan(5), items),
'Every item must have have at least 6 characters!'
]
]
const validationRules = {
id: [[ notEmpty, notEmptyMsg('id') ]],
users: userSpec,
}
const input = {
id: 4,
users: ['foobar', 'foobarbaz']
}
spected(validationRules, input)There are cases where a validation has to run against an unknown number of items. f.e. submitting a form with dynamic fields. These dynamic fields can be an array or as object keys.
const input = {
id: 4,
users: [
{firstName: 'foobar', lastName: 'action'},
{firstName: 'foo', lastName: 'bar'},
{firstName: 'foobar', lastName: 'Action'},
]
}All users need to run against the same spec.
const capitalLetterMsg = 'Capital Letter needed.'
const userSpec = {
firstName: [[isLengthGreaterThan(5), minimumMsg('firstName', 6)]],
lastName: [[hasCapitalLetter, capitalLetterMsg]],
}As we're only dealing with functions, map over userSpec and run the predicates against every collection item.
const validationRules = {
id: [[ notEmpty, notEmptyMsg('id') ]],
users: map(always(userSpec)),
}
spected(validationRules, input)In case of an object containing an unknown number of properties, the approach is the following.
const input = {
id: 4,
users: {
one: {firstName: 'foobar', lastName: 'action'},
two: {firstName: 'foo', lastName: 'bar'},
three: {firstName: 'foobar', lastName: 'Action'},
}
}Spected can also work with functions instead of an [predFn, errorMsg] tuple array, which means one can specify a function
that expects the input and then maps every rule to the object. Note: This example uses Ramda map, which expects the
function as the first argument and then always returns the UserSpec for every property.
const validationRules = {
id: [[ notEmpty, notEmptyMsg('id') ]],
users: map(() => userSpec)),
}How UserSpec is applied to every Object key is not spected specific, but can be freely implemented as needed.
Spected also accepts a function as an input, i.e. to simulate if a field would contain errors if empty.
const verify = validate(a => a, a => a)
const validationRules = {
name: nameValidationRule,
}
const input = {name: 'foobarbaz'}
const result = verify(validationRules, key => key ? ({...input, [key]: ''}) : input)
deepEqual({name: ['Name should not be empty.']}, result)
import {
compose,
curry,
head,
isEmpty,
length,
not,
prop,
} from 'ramda'
import spected from 'spected'
// predicates
const notEmpty = compose(not, isEmpty)
const hasCapitalLetter = a => /[A-Z]/.test(a)
const isGreaterThan = curry((len, a) => (a > len))
const isLengthGreaterThan = len => compose(isGreaterThan(len), prop('length'))
// error messages
const notEmptyMsg = field => `${field} should not be empty.`
const minimumMsg = (field, len) => `Minimum ${field} length of ${len} is required.`
const capitalLetterMsg = field => `${field} should contain at least one uppercase letter.`
// rules
const nameValidationRule = [[notEmpty, notEmptyMsg('Name')]]
const randomValidationRule = [
[isLengthGreaterThan(2), minimumMsg('Random', 3)],
[hasCapitalLetter, capitalLetterMsg('Random')],
]
const validationRules = {
name: nameValidationRule,
random: randomValidationRule,
}
spected(validationRules, {name: 'foo', random: 'Abcd'})
// {name: true, random: true}A spec can be composed of other specs, enabling to define deeply nested structures to validate against nested input. Let's see this in form of an example.
const locationSpec = {
street: [...],
city: [...],
zip: [...],
country: [...],
}
const userSpec = {
userName: [...],
lastName: [...],
firstName: [...],
location: locationSpec,
settings: {
profile: {
design: {
color: [...]
background: [...],
}
}
}
} Now we can validate against a deeply nested data structure.
import {
compose,
indexOf,
head,
isEmpty,
length,
not,
} from 'ramda'
import spected from 'spected'
const colors = ['green', 'blue', 'red']
const notEmpty = compose(not, isEmpty)
const minLength = a => b => length(b) > a
const hasPresetColors = x => indexOf(x, colors) !== -1
// Messages
const notEmptyMsg = field => `${field} should not be empty.`
const minimumMsg = (field, len) => `Minimum ${field} length of ${len} is required.`
const spec = {
id: [[notEmpty, notEmptyMsg('id')]],
userName: [[notEmpty, notEmptyMsg('userName')], [minLength(5), minimumMsg('userName', 6)]],
address: {
street: [[notEmpty, notEmptyMsg('street')]],
},
settings: {
profile: {
design: {
color: [[notEmpty, notEmptyMsg('color')], [hasPresetColors, 'Use defined colors']],
background: [[notEmpty, notEmptyMsg('background')], [hasPresetColors, 'Use defined colors']],
},
},
},
}
const input = {
id: 1,
userName: 'Random',
address: {
street: 'Foobar',
},
settings: {
profile: {
design: {
color: 'green',
background: 'blue',
},
},
},
}
spected(spec, input)
/* {
id: true,
userName: true,
address: {
street: true,
},
settings: {
profile: {
design: {
color: true,
background: true,
},
},
},
}
*/In case you want to change the way errors are displayed, you can use the low level validate function, which expects a success and a failure
callback in addition to the rules and input.
import {validate} from 'spected'
const verify = validate(
() => true, // always return true
head // return first error message head = x => x[0]
)
const spec = {
name: [
[isNotEmpty, 'Name should not be empty.']
],
random: [
[isLengthGreaterThan(7), 'Minimum Random length of 8 is required.'],
[hasCapitalLetter, 'Random should contain at least one uppercase letter.'],
]
}
const input = {name: 'foobar', random: 'r'}
verify(spec, input)
// {
// name: true,
// random: 'Minimum Random length of 8 is required.',
// }
Check the API documentation for further information.
For a deeper understanding of the underlying ideas and concepts:
Form Validation As A Higher Order Component Pt.1
Written by A.Sharif
Contributions by
and many more.
Also very special thanks to everyone that has contributed documentation updates and fixes (this list will updated).
Original idea and support by Stefan Oestreicher
MIT