++ed by:
1 non-PAUSE user
Author image Ali Zia
and 2 contributors


JSONAPI::Document - Turn DBIx results into JSON API documents.


version 2.4


    use JSONAPI::Document;
    use DBIx::Class::Schema;

    my $jsonapi = JSONAPI::Document->new({ api_url => 'http://example.com/api' });
    my $schema = DBIx::Class::Schema->connect(['dbi:SQLite:dbname=:memory:', '', '']);
    my $user = $schema->resultset('User')->find(1);

    # Builds a simple JSON API document, without any relationships
    my $doc = $jsonapi->resource_document($user);

    # Same but with all relationships
    my $doc = $jsonapi->resource_document($user, { includes => 'all_related' });

    # With only the author relationship
    my $doc = $jsonapi->resource_document($user, { includes => ['author'] });

    # Fully blown resource document with all relationships and their attributes
    my $doc = $jsonapi->compound_resource_document($user);

    # Multiple resource documents
    my $docs = $jsonapi->resource_documents($schema->resultset('User'));

    # With sparse fieldsets
    my $doc = $jsonapi->resource_document($user, { fields => [qw/name email/] });

    # Relationships with sparse fieldsets
    my $doc = $jsonapi->resource_document($user, { related_fields => { author => [qw/name expertise/] } });


Moo class that builds data structures according to the JSON API specification.


JSON API documents require that you define the type of a document, which this library does using the source_name of the result row. The type is also pluralised using Linua::EN::Inflexion while keeping relationship names intact (i.e. an 'author' relationship will still be called 'author', with the type 'authors').

This module supplies an opt-in Moo role that can be consumed by objects that layer over a DBIx::Class::Row, JSONAPI::Document::Role::Attributes. Consuming objects should implement a method called attributes which will be used throughout the creation of resource documents for that result type to build the attributes of the document. This is useful when you have a more complicated set of attributes that cannot be fulfilled by simply calling get_inflated_columns (the default behaviour).



Required; An absolute URL pointing to your servers JSON API namespace.


Boolean attribute; setting this will make the column keys for each document into kebab-cased-strings instead of snake_cased. Default is false.


compound_resource_document(DBIx::Class::Row|Object $row, HashRef $options)

A compound document is one that includes the resource object along with the data of all its relationships.

Returns a HashRef with the following structure:

        data => {
            id => 1,
            type => 'authors',
            attributes => {},
            relationships => {},
        included => [
                id => 1,
                type => 'posts',
                attributes => { ... },

The following options can be given:


An array reference specifying inclusion of a subset of relationships. By default all the relationships will be included. Use this if you only want a subset of relationships (e.g. when accepting the includes query parameter in your API requests, where you have to return only what relationships were requested).

This argument should contain strings representing direct relationships to the row, and can also contain hash refs which specify nested inclusion. Example:

    $self->compound_resource_document($post, ['author', { comments => ['author'] }]);

This will include the post as the primary document, its direct relationships 'author' and 'comments', and the 'author' of each related comment.

NOTE: Nested relationships are experimental and come with the following limitations:

  • many_to_many relationships are not supported

  • Only one level of depth is supported (so requesting 'include=comments.likes.author' will throw errors)

  • Are only available through compound_resource_document (not resource_document).

resource_document(DBIx::Class::Row|Object $row, HashRef $options)

Builds a single resource document for the given result row. Will optionally include relationships that contain resource identifiers.

Returns a HashRef with the following structure:

        id => 1,
        type => 'authors',
        attributes => {},
        relationships => {},

View the resource document specification here.

Uses decamelize from Mojo::Util to parse the source_name of the DBIx::Class::Row and set the appropriate type of the document. This is used to ensure that your rows source name gets hyphenated appropriately when converted into its plural form.

The following options can be given:

includes Str|ArrayRef

Optional; Used to specify any relationships of the row to include.

This argument can contain either the value 'all_related', which will return all the direct relationships of the row, or an array ref including a subset of direct relationships.

with_attributes Bool

If includes is used, for each resulting relationship row, the attributes (columns) of that relationship will be included.

By default, each relationship will contain a links object.

If this option is true, links object will be replaced with attributes.

fields ArrayRef

An optional list of attributes to include for the given resource. Implements sparse fieldsets in the specification.

Will pass the array reference to the attributes method (if you're using the attributes role), which should make use of the reference and return only those attributes that were requested.

Behaves the same as the fields option but for relationships, returning only those fields for the related resource that were requested.

Not specifying sparse fieldsets for a resource implies requesting all attributes for that relationship.

resource_documents(DBIx::Class::ResultSet|Object|ArrayRef $resultset, HashRef $options)

Builds the structure for multiple resource documents with a given resultset.

$resultset can be either a DBIx::Class::ResultSet object in which case this method will call all on the resultset, an object that extends DBIx::Class::ResultSet, or you can pass in an ArrayRef from your own all call.

Returns a HashRef with the following structure:

        data => [
                id => 1,
                type => 'authors',
                attributes => {},
                relationships => {},

See resource_document for a list of options.


This code is released under the Perl 5 License.