The Perl Toolchain Summit needs more sponsors. If your company depends on Perl, please support this very important event.
 / 
Validate-Tiny-0.04
River stage one • 7 direct dependents • 7 total dependents
13 ++
/ Validate::Tiny

NAME

Validate::Tiny - Minimalistic data validation

VERSION

Version 0.04

SYNOPSIS

Filter and validate user input from forms, etc.

    use Validate::Tiny qw/validate :util/;

    my $rules = {

        # List of fields to look for
        fields => [qw/name email pass pass2 gender/],

        # Filters to run on all fields
        filters => [

            # Remove spaces from all
            qr/.+/ => filter(qw/trim strip/),

            # Lowercase email
            email => filter('lc'),

            # Remove non-alphanumeric symbols from
            # both passwords
            qr/pass?/ => sub {
                $_[0] =~ s/\W/./g;
                $_[0];
            },
        ],

        # Checks to perform on all fields
        checks => [

            # All of these are required
            [qw/name email pass pass2/] => is_required(),

            # pass2 must be equal to pass
            pass2 => is_equal('pass'),

            # custom sub validates an email address
            email => sub {
                my ( $value, $params ) = @_;
                Email::Valid->address($value) ? undef : 'Invalid email';
            },

            # custom sub to validate gender
            gender => sub {
                my ( $value, $params ) = @_;
                return $value eq 'M'
                  || $value eq 'F' ? undef : 'Invalid gender';
            }

        ]
    };

    # Validate the input agains the rules
    my $result = validate( $input, $rules );

    if ( $result->{success} ) {
        my $values_hash = $result->{data};
        ...
    }
    else {
        my $errors_hash = $result->{error};
        ...
    }

    ...

DESCRIPTION

This module provides a simple, light and minimalistic way of validating user input. Except perl core modules and some test modules it has no other dependencies, which is why it does not implement any complicated checks and filters such as email and credit card matching. The basic idea of this module is to provide the validation functionality, and leave it up to the user to write their own data filters and checks. If you need a complete data validation solution that comes with many ready features, I recommend you to take a look at Data::FormValidator. If your validation logic is not too complicated or your form is relatively short, this module is a decent candidate for your project.

LOGIC

The basic principle of data/form validation is that any user input must be sanitized and checked for errors before used in the logic of the program. Validate::Tiny breaks this process in three steps:

  1. Specify the fields you want to work with via "fields". All others will be disregarded.

  2. Filter the fields' values using "filters". A filter can be as simple as changing to lower case or removing excess white space, or very complicated such as parsing and removing HTML tags.

  3. Perform a series of "checks" on the filtered values, to make sure they match the requirements. Again, the checks can be very simple as in checking if the value was defined, or very complicated as in checking if the value is a valid credit card number.

The validation returns a hash ref which contains success => 1|0, data and error hashes. If success is 1, data will contain the filtered values, otherwise error will contain the error messages for each field.

EXPORT

This module does not automatically export anything. You can optionally export validate and the following basic utility functions via the :util tag: "filter()", "is_required()", "is_equal()", "is_long_between()", "is_long_at_least()", "is_long_at_most()".

SUBROUTINES

validate()

    use Validate::Tiny qw/validate/;

    my $result = validate( \%input, \%rules );

Validates user input against a set of rules. The input is expected to be a reference to a hash.

%rules

    my %rules = (
        fields  => \@field_names,
        filters => \@filters_array,
        checks  => \@checks_array
    );

rules is a hash containing references to the following three arrays: "fields", "filters" and "checks".

fields

An array containing the names of the fields that must be filtered, checked and returned. All others will be disregarded.

    my @field_names = qw/username email password password2/;

filters

An array containing name matches and filter subs. The array must have an even number of elements. Each odd element is a field name match and each even element is a reference to a filter subroutine or a chain of filter subroutines. A filter subroutine takes one parameter - the value to be filtered, and returns the modified value.

    my @filters_array = (
        email => sub { return lc $_[0] },    # Lowercase the email
        password =>
          sub { $_[0] =~ s/\s//g; $_[0] }    # Remove spaces from password
    );

The field name is matched with the perl smart match operator, so you could have a regular expression or a reference to an array to match several fields:

    my @filters_array = (
        qr/.+/ => sub { lc $_[0] },    # Lowercase ALL

        [qw/password password2/] => sub {    # Remove spaces from both
            $_[0] =~ s/\s//g;                # password and password2
            $_[0];
        }
    );

Instead of a single filter subroutine, you can pass an array of subroutines to provide a chain of filters:

    my @filters_array = ( 
        qr/.+/ = [ sub { lc $_[0] }, sub { ucfirst $_[0] } ] 
    );

The above example will first lowercase the value then uppercase its first letter.

Some simple text filters are provided by the "filter()" subroutine.

    use Validate::Tiny qw/validate :util/;
    
    my @filters_array = (
        name => filter(qw/strip trim lc/)
    );

checks

An array ref containing name matches and check subs. The array must have an even number of elements. Each odd element is a field name match and each even element is a reference to a check subroutine or a chain of check subroutines.

A check subroutine takes two parameters - the value to be checked and a reference to the filtered input hash.

A check subroutine must return undef if the check passes or a string with an error message if the check fails.

Example:

    # Make sure the password is good
    sub is_good_password {
        my ( $value, $params ) = @_;

        if ( !defined $value ) {
            return undef;
        }

        if ( length($value) < 6 ) {
            return "The password is too short";
        }

        if ( length($value) > 40 ) {
            return "The password is too long";
        }

        if ( $value eq $params->{username} ) {
            return "Your password can not be the same as your username";
        }

        # At this point we're happy with the password
        return undef;
    }

    my $rules = {
        fields => [qw/username password/],
        checks => [ 
            password => \&is_good_password 
        ]
    };

It may be a bit counter-intuitive for some people to return undef when the check passes and a string when it fails. If you have a huge problem with this concept, then this module may not be right for you.

Important! Notice that in the beginning of is_good_password we check if $value is defined and return undef if it is not. This is because it is not the job of is_good_password to check if password is required. Its job is to determine if the password is good. Consider the following example:

    my $rules = {
        fields => [qw/username password/],
        checks => [ 
            password => [ is_required(), \&is_good_password ] 
        ]
    };

and this one too:

    my $rules = {
        fields => [qw/username password/],
        checks => [
            qr/.+/   => is_required(),
            password => \&is_good_password
        ]
    };

The above examples show how we make sure that password is available before we check if it is a good password. Of course you can check if password is defined inside is_good_password, but it would be redundant.

Chaining

The above example also shows that chaining check subroutines is available in the same fashion as chaining filter subroutines. The difference between chaining filters and chaining checks is that a chain of filters will always run all filters, and a chain of checks will exit after the first failed check and return its error message. This way the $result->{error} hash always has a single error message per field.

Using closures

When writing reusable check subroutines, sometimes you will want to be able to pass arguments. Returning closures (anonymous subs) is the recommended approach:

    sub is_long_between {
        my ( $min, $max ) = @_;
        return sub {
            my $value = shift;
            return length($value) >= $min && length($value) <= $max
              ? undef
              : "Must be between $min and $max symbols";
        };
    }

    my $rules = {
        fields => qw/password/,
        checks => [ 
            password => is_long_between( 6, 40 ) 
        ]
    };

Return value

validate returns a hash ref with three elements:

    my $result = validate(\%input, \%rules);

    # Now $result looks like this
    $result = {
        success => 1,       # or 0 if checks didn't pass
        data    => \%data,
        error   => \%error
    };

If success is 1 all of the filtered input will be in %data, otherwise the error messages will be stored in %error. If success is 0, %data may or may not contain values, but its use not recommended.

filter()

    filter( $name1, $name2, ... );

Provides a shortcut to some basic text filters. In reality, it returns a list of anonymous subs, so the following:

    my $rules = {
        filters => [
            email => filter('lc', 'ucfirst')
        ]
    };

is equivalent to this:

    my $rules = {
        filters => [ 
            email => [ sub{ lc $_[0] }, sub{ ucfirst $_[0] } ]
        ]
    };

It provides a shortcut for the following filters:

trim

Removes leading and trailing white space.

strip

Shrinks two or more white spaces to one.

lc

Lower case.

uc

Upper case.

ucfirst

Upper case first letter

is_required()

    is_required( $opt_error_msg );

is_required provides a shortcut to an anonymous subroutine that checks if the matched field is defined and it is not an empty string. Optionally, you can provide a custom error message to be returned.

is_equal()

    is_equal( $other_field_name, $opt_error_msg )

is_equal checks if the value of the matched field is the same as the value of another field within the input hash. Example:

    my $rules = {
        checks => [
            password2 => is_equal("password", "Passwords don't match")
        ]
    };

is_long_between()

    my $rules = {
        checks => [
            username => is_is_long_between( 6, 25, 'Bad username' )
        ]
    };

Checks if the length of the value is >= $min and <= $max. Optionally you can provide a custom error message. The default is Invalid value.

is_long_at_least()

    my $rules = {
        checks => [
            zip_code => is_is_long_at_least( 5, 'Bad zip code' )
        ]
    };

Checks if the length of the value is >= $length. Optionally you can provide a custom error message. The default is Must be at least %i symbols.

is_long_at_most()

    my $rules = {
        checks => [
            city_name => is_is_long_at_most( 40, 'City name is too long' )
        ]
    };

Checks if the length of the value is <= $length. Optionally you can provide a custom error message. The default is Must be at the most %i symbols.

SEE ALSO

Data::FormValidator

AUTHOR

minimalist, <minimalist at lavabit.com>

BUGS

Bug reports and patches are welcome. Reports which include a failing Test::More style test are helpful will receive priority.

You may also fork the module on Github: https://github.com/naturalist/Validate--Tiny

LICENSE AND COPYRIGHT

Copyright 2011 minimalist.

This program is free software; you can redistribute it and/or modify it under the terms as perl itself.