Data::Verifier - Profile based data verification with Moose type constraints.
version 0.47
use Data::Verifier;
my $dv = Data::Verifier->new(
filters => [ qw(trim) ],
profile => {
name => {
required => 1,
type => 'Str',
filters => [ qw(collapse) ]
},
age => {
type => 'Int'
},
sign => {
required => 1,
type => 'Str'
}
}
);
# Pass in a hash of data
my $results = $dv->verify({
name => 'Cory', age => 'foobar'
});
$results->success; # no
$results->is_invalid('name'); # no
$results->is_invalid('age'); # yes
$results->is_missing('name'); # no
$results->is_missing('sign'); # yes
$results->get_original_value('name'); # Unchanged, original value
$results->get_value('name'); # Filtered, valid value
$results->get_value('age'); # undefined, as it's invalid
Data::Verifier allows you verify data (such as web forms, which was the original idea) by leveraging the power of Moose's type constraint system.
Data::Verifier firstly intends to leverage Moose's type constraint system, which is significantly more powerful than anything I could create for the purposes of this module. Secondly it aims to keep a fairly simple interface by leveraging the aforementioned type system to keep options to a minimum.
It should be noted that if you choose to make a param a Str then validation
will fail if multiple values are provided. To allow multiple values you
must use an ArrayRef[Str].
If you use an ArrayRef-based parameterized type (e.g. ArrayRef[Str]) then Data::Verifier has the following behavior:
Each parameter supplied for the field is checked. If all the members pass
then the field is considered valid. If any of the members fail, then the
entire field is invalid. If any of the members pass then those members will
be included in the values attribute. An example:
use Moose::Util::TypeConstraints;
use Data::Verifier;
subtype 'Over10'
=> as 'Num'
=> where { $_ > 10 };
my $verifier = Data::Verifier->new(
profile => {
foos => {
type => 'ArrayRef[NumberOver10]',
}
}
);
my $res = $verifier->verify(foos => [ 1, 2, 30, 40 ]);
$res->success; # This is false, as 1 and 2 did not pass
$res->get_value('foos'); # [ 30, 40 ] because 30 and 40 passed!
$res->original_value('foos); # [ 1, 2, 30, 40 ] because it's all of them!
It should also be noted that post_checks that are specified in the profile
do not get applied to the individual members, only to the entire, completed
field that they are constituents of.
Note: Filters and such DO get applied to individual fields, so something like:
my $verifier = Data::Verifier->new(
filters => qw(trim),
profile => {
foos => {
type => 'ArrayRef[Str]',
filters => 'collapse'
}
}
);
In the above example, both trim and collapse bill be applied to each
member of foos.
Data::Verifier stops checking a field (not all, just the failed one) if it fails any of it's constraints. Consult the Execution Order below to ascertain the order. For example, if a field exceeds it's max length then it will not be checked against it's type constraint.
Data::Verifier uses MooseX::Storage to allow serialization of
Data::Verifier::Results objects. You can use this to store results for
validation across redirects. Note, however, that the value
attribute is not serialized. Since you can coerce a value into anything
it is not reasonable to expect to be able to serialize it. Have a look at
the original_value or post_filter_value in Data::Verifier::Results
if you want to know more.
Data::Verifier can verify data encapsulated in objects too. Everything works the way that it does for hash references. Each key in the profile is used as the name of a method to call on the object. In order to maintain consistency with the hash reference case, missing methods pass an 'undef' value into the verification process.
It may be important to understand the order in which the various steps of verification are performed:
- Global Filters
Any global filters in the profile are executed.
- Per-Field Filters
Any per-field filters are executed.
- Empty String Check
If the value of the field is an empty string then it is changed to an undef.
- Required Check
The parameter must now be defined if it is set as required.
- Length Check
Minimum then maximum length is checked.
- Type Check (w/Coercion)
At this point the type will be checked after an optional coercion.
- Depedency Checks
If this field has dependents then those will not be processed.
- Post Check
If the field has a post check it will now be executed.
An optional arrayref of filter names through which all values will be passed.
The profile is a hashref. Each value you'd like to verify is a key. The values specify all the options to use with the field. The available options are:
- coerce
If true then the value will be given an opportunity to coerce via Moose's type system. If this is set, coercion will be ignored.
- coercion
Set this attribute to the coercion defined for this type. If coerce is
set this attribute will be ignored. See the coercion method above.
- dependent
Allows a set of fields to be specifid as dependents of this one. The argument for this key is a full-fledged profile as you would give to the profile key:
my $verifier = Data::Verifier->new( profile => { password => { dependent => { password2 => { required => 1, } } } } );
In the above example password is not required. If it is provided then
password2 must also be provided. If any depedents of a field are missing or
invalid then that field is invalid. In our example if password is provided
and password2 is missing then password will be invalid.
- filters
An optional list of filters through which this specific value will be run. See the documentation for Data::Verifier::Filters to learn more. This value my be either a scalar (string or coderef) or an arrayref of strings or coderefs.
- max_length
An optional length which the value may not exceed.
- min_length
An optional length which the value may not be less.
- member_post_check
A post check that is only to be applied to the members of an ArrayRef based type. Because it is verified in something of a vacuum, the results object it receives will have no other values to look at. Therefore member_post_check is only useful if you want to do some sort of weird post-check thing that I can't imagine would be a good idea.
- post_check
The post_check key takes a subref and, after all verification has finished,
executes the subref with the results of the verification as it's only argument.
The subref's return value determines if the field to which the post_check
belongs is invalid. A typical example would be when the value of one field
must be equal to the other, like an email confirmation:
my $verifier = Data::Verifier->new( profile => { email => { required => 1, dependent => { email2 => { required => 1, } }, post_check => sub { my $r = shift; return $r->get_value('email') eq $r->get_value('email2'); } }, } );
my $results = $verifier->verify({ email => 'foo@example.com', email2 => 'foo2@example.com' });
$results->success; # false $results->is_valid('email'); # false $results->is_valid('email2'); # true, as it has no post_check
In the above example, success will return false, because the value of
email does not match the value of email2. is_valid will return false
for email but true for email2, since nothing specifically invalidated it.
In this example you should rely on the email field, as email2 carries no
significance but to confirm email.
Note about post_check and exceptions: If have a more complex post_check
that could fail in multiple ways, you can die in your post_check coderef
and the exception will be stored in the fields reason attribute.
Note about post_check and ArrayRef based types: The post check is not executed for ArrayRef based types. See the note earlier in this documentation about ArrayRefs.
- required
Determines if this field is required for verification.
- type
The name of the Moose type constraint to use with verifying this field's value. Note, this will also accept an instance of Moose::Meta::TypeConstraint, although it may not serialize properly as a result.
Define a coercion to use for verification. This will not define a global Moose type coercion, but is instead just a single coercion to apply to a specific entity.
my $verifier = Data::Verifier->new(
profile => {
a_string => {
type => 'Str',
coercion => Data::Verifier::coercion(
from => 'Int',
via => sub { (qw[ one two three ])[ ($_ - 1) ] }
),
},
}
);
Mike Eldridge
George Hartzell
Stevan Little
Jason May
Dennis Schön
J. Shirley
Cory G Watson gphat@cpan.org
This software is copyright (c) 2011 by Cold Hard Code, LLC.
This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.