ClientSideValidations
made easy for your Rails 6.1, 7.x, and 8.0 applications!
- Follow the best practices for client side validations developed by Luke Wroblewski
- Automatically extract and apply validation rules defined on the server to the client.
- In the cases where a server-side validation rule would not work on the client (i.e. conditional callbacks like :if, :unless) then do not attempt client side validations. Fall back to the server side validation.
- The client side validation error rendering should be indistinguishable from the server side validation error rendering.
- Wide browser compliancy.
- Work with any ActiveModel::Validations based model
- Validate nested fields
- Support custom validations
- Client side validation callbacks
- Plugin system to support additional FormBuilders, ORMs, etc...
Add the following line to your Gemfile:
gem 'client_side_validations'
Then run bundle install
Please run spring stop
if you are using Spring
Next you need to run the generator:
rails g client_side_validations:install
This will install the initializer:
config/initializers/client_side_validations.rb
Instructions depend on your technology stack.
Please note that CSV depends on jQuery >= 1.12.4 (jQuery slim is fine).
Make sure that you are requiring jQuery.
Add the following package:
yarn add @client-side-validations/client-side-validations
Then add the following line to your app/javascript/packs/application.js
pack:
// If you are using `import` syntax
import '@client-side-validations/client-side-validations/src'
// If you are using `require` syntax
require('@client-side-validations/client-side-validations')
If you are using Turbo, use the
import
syntax and make sure that @client-side-validations/client-side-validations/src
is imported after @hotwired/turbo-rails
, so ClientSideValidations can properly detect
window.Turbo
and attach its event handlers.
If you are using Turbolinks 5.2,
use the require
syntax and make sure that @client-side-validations/client-side-validations
is required after Turbolinks.start()
, so ClientSideValidations can properly
detect window.Turbolinks
and attach its event handlers.
Since ClientSideValidations can also be used via webpacker, it does not require
by default jquery-rails
gem.
Make sure that jquery-rails
is part of your bundled gems and application.js
,
otherwise add:
gem 'jquery-rails', '~> 4.3'
to your Gemfile
, run bundle
, and add
//= require jquery
to your app/assets/javascripts/application.js
file.
Then, add the following to your app/assets/javascripts/application.js
file
after //= require jquery
.
//= require rails.validations
If you are using Turbolinks,
make sure that rails.validations
is required after turbolinks
, so
ClientSideValidations can properly attach its event handlers.
If you need to copy the asset files from the gem into your project, run:
rails g client_side_validations:copy_assets
Note: If you run copy_assets
, you will need to run it again each time you update this project.
The initializer includes a commented out ActionView::Base.field_error_proc
.
Uncomment this to render your error messages inline with the input fields.
I recommend you to not use a solution similar to error_messages_for
. Client
Side Validations does not support this type of error rendering. If you want to
maintain consistency between the client side rendered validation error messages
and the server side rendered validation error messages please use what is in
config/initializers/client_side_validations.rb
There is additional support for other ActiveModel
based ORMs and other
Rails FormBuilders
. Please see the Plugin wiki page
(feel free to add your own)
In your FormBuilder
you only need to enable validations:
<%= form_for @user, validate: true do |f| %>
...
That should be enough to get you going.
Starting from version 14.0, ClientSideValidations also supports form_with
.
The syntax is the same as form_for
:
<%= form_with model: @user, validate: true do |f| %>
...
Note: ClientSideValidations requires id
attributes on form fields to
work, so it will force form_with
to generate ids.
By default, ClientSideValidations will perform the validations in the same order specified in your models. In other words, if you want to validate the format of an email field before its presence, you can use the following:
class User < ApplicationRecord
validates :email, format: { with: /\A[^@\s]+@[^@\s]+\z/ }, presence: true
end
By default conditional validators are not evaluated and passed to the client. We do this because the state model when the form is rendered is not necessarily the state of the model when the validations fire server-side. However, if you wish to override this behavior you can do so in the form. Given the following model:
class Person < ActiveRecord::Base
validates :name, presence: true, length: { maximum: 10 }, if: :can_validate?
def can_validate?
true
end
end
You can force in the form:
<%= f.text_field :name, validate: true %>
Passing validate: true
will force all the validators for that attribute. If there are conditionals
they are evaluated with the state of the model when rendering the form. You can also force
individual validators:
<%= f.text_field :name, validate: { presence: true } %>
In the above case only the presence
validator will be passed to the client.
This is also the case with other supported conditional validations (such as Procs, Arrays or Strings).
NOTE: when :if
conditional includes a symbol or a string with
changed?
in it or start with will_save_change_to
, validator will forced automatically.
class Person < ActiveRecord::Base
validates :name, presence: true, if: :name_changed?
end
The presence of name in the example above will be validated without explicit forcing.
This is done because it is always assumed the value will change on the form.
Conditionals defined with :unless
key do not have this optimization.
If you wish to skip validations on a given attribute force it to false
:
<%= f.text_field :name, validate: false %>
If you want to be more selective about the validation that is turned off you can simply do:
<%= f.text_field :name, validate: { presence: false } %>
You can even turn them off per fieldset:
<%= f.fields_for :profile, validate: false do |p| %>
...
Please note that pass
callback will also be performed on fields that skip validations.
A rendered form with validations will always have a data-client-side-validations
attribute.
The objects it contains will have different keys depending upon the FormBuilder
being used. However, html_settings
and validators
will always be present.
This will always contain the type to the class of the FormBuilder
that did the rendering. The type will be used by the JavaScript to determine how to add
and remove
the error messages. If you create a new FormBuilder
, you will need to write your own handlers for adding and removing.
This object contains the validators for each of the inputs rendered on the FormBuilder
. Each input is keyed to the name
attribute and each containing validator could simply contain the error message itself or also specific options on how that validator should be run.
If you need to add more validators but don't want them rendered on the form immediately you can inject those validators with FormBuilder#validate
:
<%= form_for @user, validate: true do |f| %>
<p>
<%= f.label :name %>
<%= f.text_field :name %>
</p>
<%= f.validate :age, :bio %>
...
In the above example age
and bio
will not render as inputs on the form but their validators will be properly added to the validators
object for use later. If you do intend to dynamically render these inputs later the name
attributes on the inputs will have to match with the keys on the validators
object, and the inputs will have to be enabled for client side validation.
You can add all attributes with validators for the given object by passing nothing:
<%= f.validate %>
You can also force validators similarly to the input syntax:
<%= f.validate :email, presence: false %>
Take care when using this method. The embedded validators are overwritten based upon the order they are rendered. So if you do something like:
<%= f.text_field :email, validate: { presence: false } %>
<%= f.validate %>
The presence
validator will not be turned off because the options
were overwritten by the call to FormBuilder#validate
ClientSideValidations
will use ActiveRecord::Base.field_error_proc
to render the error messages. Other FormBuilders
will use their own settings.
If you need to change the markup of how the errors are rendered you can modify that in config/initializers/client_side_validations.rb
Please Note if you modify the markup, you will also need to modify ClientSideValidations.formBuilders['ActionView::Helpers::FormBuilder']
's add
and remove
functions. You can override the behavior by creating a new JavaScript file called rails.validations.actionView.js
that contains the following:
window.ClientSideValidations.formBuilders['ActionView::Helpers::FormBuilder'] = {
add: function($element, settings, message) {
// custom add code here
},
remove: function($element, settings) {
// custom remove code here
}
}
Please view the code in rails.validations.js
to see how the existing add
and remove
functions work and how best to override for your specific use-case.
Client Side Validations supports the use of custom validators. The following is an example for creating a custom validator that validates the format of email addresses.
Let's say you have several models that all have email fields and you are validating the format of that email address on each one. This is a common validation and could probably benefit from a custom validator. We're going to put the validator into config/initializers/email_validator.rb
class EmailValidator < ActiveModel::EachValidator
def validate_each(record, attr_name, value)
unless value =~ /^([^@\s]+)@((?:[-a-z0-9]+\.)+[a-z]{2,})$/i
record.errors.add(attr_name, :email, options.merge(value: value))
end
end
end
# This allows us to assign the validator in the model
module ActiveModel::Validations::HelperMethods
def validates_email(*attr_names)
validates_with EmailValidator, _merge_attributes(attr_names)
end
end
Heads-up!: Put custom initializers in config/initializers
, otherwise named validator helpers will not
be available and migrations will not work.
Next we need to add the error message to the Rails i18n file config/locales/en.yml
# config/locales/en.yml
en:
errors:
messages:
email: "Not an email address"
Finally we need to add a client side validator. This can be done by hooking into the ClientSideValidations.validator
object. Create a new file app/assets/javascripts/rails.validations.customValidators.js
// The validator variable is a JSON Object
// The selector variable is a jQuery Object
window.ClientSideValidations.validators.local['email'] = function($element, options) {
// Your validator code goes in here
if (!/^([^@\s]+)@((?:[-a-z0-9]+\.)+[a-z]{2,})$/i.test($element.val())) {
// When the value fails to pass validation you need to return the error message.
// It can be derived from validator.message
return options.message;
}
}
That's it! Now you can use the custom validator as you would any other validator in your model
# app/models/person.rb
class Person < ActiveRecord::Base
validates_email :email
end
Client Side Validations will apply the new validator and validate your forms as needed.
There are many reasons why you might want to enable, disable, or even completely reset the bound validation events on the client. ClientSideValidations
offers a simple API for this.
If you have rendered a new form via AJAX into your page you will need to enable that form for validation:
$(new_form).enableClientSideValidations();
You should attach this to an event that is fired when the new HTML renders.
You can use the same function if you introduce new inputs to an existing form:
$(new_input).enableClientSideValidations();
If you wish to turn off validations entirely on a form:
$(form).disableClientSideValidations();
You can reset the current state of the validations, clear all error messages, and reattach clean event handlers:
$(form).resetClientSideValidations();
ClientSideValidations
will run callbacks based upon the state of the element or form. The following callbacks are supported:
ClientSideValidations.callbacks.element.after($element, eventData)
ClientSideValidations.callbacks.element.before($element, eventData)
ClientSideValidations.callbacks.element.fail($element, message, callback, eventData)
ClientSideValidations.callbacks.element.pass($element, callback, eventData)
ClientSideValidations.callbacks.form.after($form, eventData)
ClientSideValidations.callbacks.form.before($form, eventData)
ClientSideValidations.callbacks.form.fail($form, eventData)
ClientSideValidations.callbacks.form.pass($form, eventData)
The names of the callbacks should be pretty straight forward. For example, ClientSideValidations.callbacks.form.fail
will be called if a form failed to validate. And ClientSideValidations.callbacks.element.before
will be called before that particular element's validations are run.
All element callbacks will receive the element in a jQuery object as the first parameter and the eventData object as the second parameter. ClientSideValidations.callbacks.element.fail()
will receive the message of the failed validation as the second parameter, the callback for adding the error fields as the third and the eventData object as the third. ClientSideValidations.elementValidatePass()
will receive the callback for removing the error fields. The error field callbacks must be run in your custom callback in some fashion. (either after a blocking event or as a callback for another event, such as an animation)
All form callbacks will receive the form in a jQuery object as the first parameter and the eventData object as the second parameter.
Here is an example callback for sliding out the error message when the validation fails then sliding it back in when the validation passes:
// You will need to require 'jquery-ui' for this to work
window.ClientSideValidations.callbacks.element.fail = function($element, message, callback) {
callback()
if ($element.data('valid') !== false) {
$element.parent().find('.message').hide().show('slide', { direction: 'left', easing: 'easeOutBounce' }, 500)
}
}
window.ClientSideValidations.callbacks.element.pass = function($element, callback) {
// Take note how we're passing the callback to the hide()
// method so it is run after the animation is complete.
$element.parent().find('.message').hide('slide', { direction: 'left' }, 500, callback)
}
.message {
background-color: red;
border-bottom-right-radius: 5px 5px;
border-top-right-radius: 5px 5px;
padding: 2px 5px;
}
div.field_with_errors div.ui-effects-wrapper {
display: inline-block !important;
}
Finally uncomment the ActionView::Base.field_error_proc
override in config/initializers/client_side_validations.rb
If you want to disable some validators, set the disabled_validators
config variable in config/initializers/client_side_validations.rb
:
# Example: disable the presence validator
ClientSideValidations::Config.disabled_validators = [:presence]
Note that the FormBuilder
will automatically skip building validators that are disabled.
By default, ClientSideValidations will automatically validate the form.
If for some reason you would like to manually validate the form (for example you're working with a multi-step form), you can use the following approach:
$input = $('#myInputField');
$form = $($input[0].form);
validators = $form[0].ClientSideValidations.settings.validators;
// Validate a single field
// It might not work for multiple inputs selected at once by `$input`
$input.isValid(validators);
// Validate the whole form
$form.isValid(validators);
To manually validate a single field, you may also trigger a focusout event:
$('#myInputField').trigger('focusout');
We are very thankful for the many contributors
This gem follows Semantic Versioning
Please do! We are always looking to improve this gem. Please see our Contribution Guidelines on how to properly submit issues and pull requests.
DockYard, LLC © 2012-2024