StripAttributes
StripAttributes is an ActiveModel extension that automatically strips leading
and trailing whitespace from string-valued attributes before validation. If
the resulting string is blank, it converts the value to nil by default.
It works by adding a before_validation hook to the record. By default, all
string-valued attributes are stripped of whitespace, but :only and :except
options can be used to limit which attributes are stripped. Both options accept
a single attribute (only: :field) or arrays of attributes (except: [:field1, :field2, :field3]).
It's also possible to skip stripping the attributes altogether per model using the :if and :unless options.
Contents
- Installation
- Quick Start
- Options Reference
- Whitespace and Blank Values
- Examples
- Usage Patterns
- Testing Your Models
- Support
- Contributing
- Versioning and Changelog
Installation
Include the gem in your Gemfile:
gem "strip_attributes"Quick Start
For a Rails model with a name attribute:
class User < ApplicationRecord
strip_attributes only: :name
end
user = User.new(name: " Ada Lovelace \t")
user.name # => " Ada Lovelace \t"
user.valid?
user.name # => "Ada Lovelace"Attributes are normalized before validation, rather than on assignment.
Calling valid? updates the selected attributes in memory even without saving
the record, and even if validation fails. A normal save also runs validation
and triggers normalization. Operations that skip validations, such as
save(validate: false) and update_columns, do not trigger this hook.
Options Reference
Pass options to strip_attributes in your model:
| Option | Accepted values | Default | Effect |
|---|---|---|---|
only |
Attribute name as a symbol or string, or an array of names | All attributes | Process only the named attributes. |
except |
Attribute name as a symbol or string, or an array of names | No exclusions | Skip the named attributes. |
allow_empty |
true or false
|
Disabled | Keep stripped empty strings as "" instead of converting them to nil. Existing nil values remain nil. |
collapse_spaces |
true or false
|
Disabled | Replace runs of horizontal whitespace with one space, preserving internal newlines. |
replace_newlines |
true or false
|
Disabled | Replace runs of carriage returns (\r) and line feeds (\n) with one space. Surrounding spaces remain unless collapse_spaces is also enabled. |
regex |
Regular expression | None | Remove all matches before normal whitespace trimming. |
if |
Predicate method name as a symbol or a proc | No condition | Run the validation callback only when the condition is truthy. |
unless |
Predicate method name as a symbol or a proc | No condition | Skip the validation callback when the condition is truthy. |
With compatible string encodings, collapse_spaces handles tabs, Unicode
horizontal whitespace, and the selected invisible characters described below.
For incompatible encodings, it only collapses repeated ordinary spaces.
Use either only or except in a declaration. If both are supplied, except
takes precedence and only is ignored.
Each strip_attributes declaration registers a separate validation callback.
Multiple declarations can apply different options to different attributes;
overlapping selections can process the same attribute more than once.
Whitespace and Blank Values
Only String values are processed. Numbers, booleans, arrays, hashes, and nil
are left unchanged. Strings nested inside arrays or hashes are not processed.
By default, leading and trailing whitespace is removed, while internal
whitespace is preserved. Empty strings and strings containing only whitespace
become nil. Use allow_empty: true to keep stripped empty strings as "".
The following examples use Ruby string notation (\t is a tab and \n is a
newline):
| Input | Options | Result |
|---|---|---|
" \t " |
Default | nil |
" \t " |
allow_empty: true |
"" |
"a b" |
Default | "a b" |
"a \n b" |
collapse_spaces: true |
"a \n b" |
"a \n b" |
replace_newlines: true |
"a b" |
"a \n b" |
replace_newlines: true, collapse_spaces: true |
"a b" |
With compatible string encodings, trimming also handles Unicode whitespace
and selected invisible characters, including nonbreaking spaces (U+00A0)
and zero-width spaces (U+200B). A string containing only these characters
also becomes nil by default. For incompatible encodings, trimming falls
back to Ruby's String#strip behavior.
Examples
Default Behavior
class DrunkPokerPlayer < ActiveRecord::Base
strip_attributes
endUsing except
# all attributes will be stripped except :boxers
class SoberPokerPlayer < ActiveRecord::Base
strip_attributes except: :boxers
endUsing only
# only :shoe, :sock, and :glove attributes will be stripped
class ConservativePokerPlayer < ActiveRecord::Base
strip_attributes only: [:shoe, :sock, :glove]
endUsing if
The if and unless options accept a predicate method name or a proc. The
condition is evaluated each time validation runs.
This example normalizes email unless the record's preserve_whitespace flag
is set to true:
class User < ActiveRecord::Base
attr_accessor :preserve_whitespace
strip_attributes only: :email, if: :normalize_email?
def normalize_email?
!preserve_whitespace
end
endUsing unless
The same behavior can be expressed with unless and a proc:
class User < ActiveRecord::Base
attr_accessor :preserve_whitespace
strip_attributes only: :email, unless: ->(record) { record.preserve_whitespace }
endUsing allow_empty
# Empty attributes will not be converted to nil
class BrokePokerPlayer < ActiveRecord::Base
strip_attributes allow_empty: true
endUsing collapse_spaces
# Sequential spaces in attributes will be collapsed to one space
class EloquentPokerPlayer < ActiveRecord::Base
strip_attributes collapse_spaces: true
endUsing replace_newlines
# Newlines in attributes will be replaced with a space
class EloquentPokerPlayer < ActiveRecord::Base
strip_attributes replace_newlines: true
endUsing regex
The regex option removes matching characters throughout the string before
normal leading and trailing whitespace trimming. It supplements the default
trimming behavior rather than replacing it.
class User < ActiveRecord::Base
# Remove non-alphabetic, non-whitespace characters
strip_attributes only: [:first_name, :last_name], regex: /[^[:alpha:]\s]/
# Remove non-digit characters
strip_attributes only: :phone, regex: /[^0-9]/
# Keep only alphanumeric characters, underscores, and hyphens
strip_attributes only: :nickname, regex: /[^[:alnum:]_-]/
# Remove trailing horizontal whitespace from each line
strip_attributes only: :code, regex: /[[:blank:]]+$/
endIn the multiline example, normal trimming also removes leading whitespace at the beginning of the entire string and trailing whitespace at its end.
Usage Patterns
Other ORMs implementing ActiveModel
It also works on other ActiveModel classes, such as Mongoid documents:
class User
include Mongoid::Document
strip_attributes only: :email
endUsing it with ActiveAttr
class Person
include ActiveAttr::Model
include ActiveModel::Validations::Callbacks
attribute :name
attribute :email
strip_attributes
endUsing it directly
# where record is an ActiveModel instance
StripAttributes.strip(record, collapse_spaces: true)
# works directly on Strings too
StripAttributes.strip(" foo \t") #=> "foo"
StripAttributes.strip(" foo bar", collapse_spaces: true) #=> "foo bar"Testing Your Models
StripAttributes provides an RSpec/Shoulda-compatible matcher for easier
testing of attribute normalization during validation. The matcher assigns test
values, calls valid?, and checks the resulting attribute values. You can use it with
RSpec, Shoulda,
Minitest-MatchersVaccine
(preferred), or
Minitest-Matchers.
Setup spec_helper.rb or test_helper.rb
To initialize RSpec, add this to your spec_helper.rb:
require "strip_attributes/matchers"
RSpec.configure do |config|
config.include StripAttributes::Matchers
endTo initialize Shoulda (with test-unit), add this to your test_helper.rb:
require "strip_attributes/matchers"
class Test::Unit::TestCase
extend StripAttributes::Matchers
endOR if in a Rails environment, you might prefer this:
require "strip_attributes/matchers"
class ActiveSupport::TestCase
extend StripAttributes::Matchers
endMinitest-MatchersVaccine
Add the adapter gem to your Gemfile and run bundle install:
gem "minitest-matchers_vaccine", group: :testThen add this to your test_helper.rb:
require "minitest/autorun"
require "minitest/matchers_vaccine"
require "strip_attributes/matchers"
class Minitest::Spec
include StripAttributes::Matchers
endOR if in a Rails environment, you might prefer this:
require "minitest/matchers_vaccine"
require "strip_attributes/matchers"
class ActiveSupport::TestCase
include StripAttributes::Matchers
endMinitest-Matchers
Add the adapter gem to your Gemfile and run bundle install:
gem "minitest-matchers", group: :testThen add this to your test_helper.rb:
require "minitest/autorun"
require "minitest/matchers"
require "strip_attributes/matchers"
class Minitest::Spec
include StripAttributes::Matchers
endWriting Tests
The matcher uses "string" as its default test value. Use .using("AAPL") to
provide a value that remains unchanged by other setters or callbacks, such as
one that uppercases a stock ticker. The matcher adds whitespace around that
value and expects it to be removed during validation.
A positive multi-attribute matcher requires every listed attribute to match. Negating it only establishes that at least one attribute does not match. Use separate negative assertions to check that each attribute preserves whitespace, as shown below.
RSpec:
describe User do
it { is_expected.to strip_attribute(:name).collapse_spaces }
it { is_expected.to strip_attribute(:name).replace_newlines }
it { is_expected.to strip_attribute :email }
it { is_expected.to strip_attributes(:name, :email) }
it { is_expected.to strip_attributes(:ticker).using("AAPL") }
it { is_expected.not_to strip_attribute :password }
it { is_expected.not_to strip_attribute :encrypted_password }
endShoulda (with test-unit):
class UserTest < ActiveSupport::TestCase
should strip_attribute(:name).collapse_spaces
should strip_attribute(:name).replace_newlines
should strip_attribute :email
should strip_attributes(:name, :email)
should strip_attributes(:ticker).using("AAPL")
should_not strip_attribute :password
should_not strip_attribute :encrypted_password
endMinitest-MatchersVaccine:
describe User do
subject { User.new }
it "should strip attributes" do
must strip_attribute(:name).collapse_spaces
must strip_attribute(:name).replace_newlines
must strip_attribute :email
must strip_attributes(:name, :email)
must strip_attributes(:ticker).using("AAPL")
wont strip_attribute :password
wont strip_attribute :encrypted_password
end
endMinitest-Matchers:
describe User do
subject { User.new }
must { strip_attribute(:name).collapse_spaces }
must { strip_attribute(:name).replace_newlines }
must { strip_attribute :email }
must { strip_attributes(:name, :email) }
must { strip_attributes(:ticker).using("AAPL") }
wont { strip_attribute :password }
wont { strip_attribute :encrypted_password }
endSupport
Submit suggestions or feature requests as a GitHub Issue or Pull Request (preferred). If you send a pull request, remember to update the corresponding unit tests. In fact, I prefer new features to be submitted in the form of new unit tests.
Credits
The idea was originally triggered by the information at the (now defunct) Rails Wiki but was modified from the original to include more idiomatic ruby and rails support.
Versioning and Changelog
StripAttributes follows Semantic Versioning 2.0. See CHANGELOG.md for release notes and changes between versions.