Data-Record-Serialize/lib/Data/Record/Serialize.pm
package Data::Record::Serialize;
# ABSTRACT: Flexible serialization of a record
use 5.010000;
use strict;
use warnings;
use warnings::register qw( Encode::dbi::queue );
use Data::Record::Serialize::Error -all;
our $VERSION = '2.02';
use Package::Variant
importing => [
'Moo',
'String::RewritePrefix' => [
rewrite => {
-as => 'rewrite_encode',
prefixes => {
q{} => 'Data::Record::Serialize::Encode::',
q{+} => q{},
},
},
],
'String::RewritePrefix' => [
rewrite => {
-as => 'rewrite_sink',
prefixes => {
q{} => 'Data::Record::Serialize::Sink::',
q{+} => q{},
},
},
],
],
subs => [qw( with has rewrite_encode rewrite_sink )];
use namespace::clean;
sub make_variant {
my ( undef, $target, %attr ) = @_;
error( 'attribute::value', 'must specify <encode> attribute' )
unless defined $attr{encode};
with 'Data::Record::Serialize::Role::Base';
my $encoder = rewrite_encode( $attr{encode} );
with $encoder;
if ( $target->does( 'Data::Record::Serialize::Role::Sink' ) ) {
error( 'attribute::value',
"encoder ($encoder) is already a sink; don't specify a sink attribute\n" )
if defined $attr{sink};
}
else {
with rewrite_sink( $attr{sink} ? $attr{sink} : 'stream' );
}
with 'Data::Record::Serialize::Role::Default';
}
sub new {
my $class = shift;
my $attr = 'HASH' eq ref $_[0] ? shift : {@_};
my %class_attr = (
encode => $attr->{encode},
sink => $attr->{sink},
);
$class = Package::Variant->build_variant_of( __PACKAGE__, %class_attr );
return $class->new( $attr );
}
1;
#
# This file is part of Data-Record-Serialize
#
# This software is Copyright (c) 2017 by Smithsonian Astrophysical Observatory.
#
# This is free software, licensed under:
#
# The GNU General Public License, Version 3, June 2007
#
__END__
=pod
=for :stopwords Diab Jerius Smithsonian Astrophysical Observatory deterministically d'etre
raison truthy numified
=head1 NAME
Data::Record::Serialize - Flexible serialization of a record
=head1 VERSION
version 2.02
=head1 SYNOPSIS
use Data::Record::Serialize;
# simple output to json
$s = Data::Record::Serialize->new( encode => 'json', \%attr );
$s->send( \%record );
# cleanup record before sending
$s = Data::Record::Serialize->new( encode => 'json',
fields => [ qw( obsid chip_id phi theta ) ],
format => 1,
format_types => { N => '%0.4f' },
format_fields => { obsid => '%05d' },
rename_fields => { chip_id => 'CHIP' },
types => { obsid => 'I', chip_id => 'S',
phi => 'N', theta => 'N' },
);
$s->send( \%record );
# send to an SQLite database
$s = Data::Record::Serialize->new(
encode => 'dbi',
dsn => [ 'SQLite', [ dbname => $dbname ] ],
table => 'stuff',
format => 1,
fields => [ qw( obsid chip_id phi theta ) ],
format_types => { N => '%0.4f' },
format_fields => { obsid => '%05d' },
rename_fields => { chip_id => 'CHIP' },
types => { obsid => 'I', chip_id => 'S',
phi => 'N', theta => 'N' },
);
$s->send( \%record );
=head1 DESCRIPTION
B<Data::Record::Serialize> encodes data records and sends them
somewhere. This module is primarily useful for output of sets of
uniformly structured data records. It provides a uniform, thin,
interface to various serializers and output sinks. Its I<raison
d'etre> is its ability to manipulate the records prior to encoding
and output.
=over
=item *
A record is a collection of fields, i.e. keys and I<scalar>
values.
=item *
All records are assumed to have the same structure.
=item *
Fields may have simple types.
=item *
Fields may be renamed upon output.
=item *
A subset of the fields may be selected for output.
=item *
Field values may be transformed prior to output.
=back
=head2 Types
Some output encoders care about the type of a
field. B<Data::Record::Serialize> recognizes these types:
=over
=item *
C<N> - Number (any number)
=item *
C<I> - Integer
=item *
C<S> - String
=item *
C<B> - Boolean
=back
Not all encoders support separate integer or Boolean types. Where not supported,
integers are encoded as numbers and Booleans as integers.
Types may be specified for fields, or may be automatically determined
from the first record which is output. It is not possible to
deterministically determine if a field is Boolean, so such fields
must be explicitly specified. Boolean fields should be "truthy",
e.g., when used in a conditional, they evaluate to true or false.
=head2 Field transformation
Transformations can be applied to fields prior to output, and may be
specified globally for data types as well as for specifically for
fields. The latter take precedence.
Transformations are specified via the L</format_fields> and
L</format_types> parameters. They can either be a C<sprintf>
compatible format string,
format_types => { N => '%0.4f' },
format_fields => { obsid => '%05d' },
or a code reference:
format_types => { B => sub { Lingua::Boolean::Tiny::boolean( $_[0] ) } }
=head2 Encoders
The available encoders and their respective documentation are:
=over
=item *
C<dbi> - L<Data::Record::Serialize::Encode::dbi>
Write to a database via B<DBI>. This is a combined
encoder and sink.
=item *
C<ddump> - L<Data::Record::Serialize::Encode::ddump>
encode via L<Data::Dumper>
=item *
C<json> - L<Data::Record::Serialize::Encode::json>
=item *
C<null> - L<Data::Record::Serialize::Sink::null>
This is a combined encoder and sink.
=item *
C<rdb> - L<Data::Record::Serialize::Encode::rdb>
=item *
C<yaml> - L<Data::Record::Serialize::Encode::yaml>
=back
=head2 Sinks
Sinks are where encoded data are sent.
The available sinks and their documentation are:
=over
=item *
C<stream> - L<Data::Record::Serialize::Sink::stream>
write to a stream
=item *
C<null> - L<Data::Record::Serialize::Sink::null>
send the encoded data to the bit bucket.
=item *
C<array> - L<Data::Record::Serialize::Sink::array>
append the encoded data to an array.
=back
Refer to the documentation for additional constructor options, and
object and class methods and attributes;
=head2 Fields and their types
Which fields are output and how their types are determined depends
upon the L</fields>, L</types>, and L</default_type> attributes.
This seems a bit complicated, but the idea is to provide a DWIM
interface requiring minimal specification.
In the following table:
N => not specified
Y => specified
X => doesn't matter
all => the string 'all'
Automatic type determination is done by examining the first
record sent to the output stream.
Automatic output field determination is done by examining the first
record sent to the output stream. Only those fields will be
output for subsequent records.
fields types default_type Result
------ ----- ------------ ------
N/all N N Output fields are automatically determined.
Types are automatically determined.
N/all N Y Output fields are automatically determined.
Types are set to <default_type>.
Y N N Fields in <fields> are output.
Types are automatically determined.
Y Y N Fields in <fields> are output.
Fields in <types> get the specified type.
Types for other fields are automatically determined.
Y Y Y Fields in <fields> are output.
Fields in <types> get the specified type.
Types for other fields are set to <default_type>.
all Y N Output fields are automatically determined.
Fields in <types> get the specified type.
Types for other fields are automatically determined.
all Y Y Output fields are automatically determined.
Fields in <types> get the specified type.
Types for other fields are set to <default_type>.
N Y X Fields in <types> are output.
Types are specified by <types>.
=head2 Errors
Most errors result in exception objects being thrown, typically in the
L<Data::Record::Serialize::Error> hierarchy.
=head1 CLASS METHODS
=head2 B<new>
$s = Data::Record::Serialize->new( %args );
$s = Data::Record::Serialize->new( \%args );
Construct a new object. In addition to any arguments required or provided
by the specified encoders and sinks, the following arguments are recognized:
=over
=item C<types> => I<hashref>|I<arrayref>
This specifies types (C<N>, C<I>, C<S>, C<B> ) for fields.
If an array, the fields will be output in the specified order,
provided the encoder permits it (see below, however). For example,
# use order if possible
types => [ c => 'N', a => 'N', b => 'N' ]
# order doesn't matter
types => { c => 'N', a => 'N', b => 'N' }
If L</fields> is specified, then its order will override that specified
here.
To understand how this attribute works in concert with L</fields> and
L</default_type>, please see L</Fields and their types>.
=item C<default_type> => C<S>|C<N>|C<I>|C<B>
The default input type for fields whose types were not specified via
the L</types>.
To understand how this attribute works in concert with L</fields> and
L</types>, please see L</Fields and their types>.
=item C<fields> => I<arrayref>|C<all>
The fields to output. If it is the string C<all>,
all input fields will be output. If it is an arrayref, the
fields will be output in the specified order, provided the encoder
permits it.
To understand how this attribute works in concert with L</types> and
L</default_type>, please see L</Fields and their types>.
=item C<encode> => I<encoder>
I<Required>. The module which will encode the data.
With no prefix, it is assumed to be in the
C<Data::Record::Serialize::Encode> namespace. With a prefix of C<+>
it is a fully qualified module name.
Specific encoders may provide additional, or require specific,
attributes. See L</Encoders> for more information.
=item C<sink> => I<sink>
The module which writes the encoded data.
With no prefix, it is assumed to be in the
C<Data::Record::Serialize::Sink> namespace. With a prefix of C<+>
it is a fully qualified module name.
Specific sinks may provide additional, or require specific
attributes. See L</Sinks> for more information.
The value is C<stream> (corresponding to
L<Data::Record::Serialize::Sink::stream>), unless the encoder is also
a sink.
It is an error to specify a sink if the encoder already acts as one.
=item C<nullify> => I<arrayref>|I<coderef>|Boolean
Fields that should be set to C<undef> if they are
empty. Sinks should encode C<undef> as the C<null> value.
B<nullify> may be passed:
=over
=item * an arrayref of input field names
=item * a coderef
The coderef is called as
@input_field_names = $code->( $serializer_object )
=item * a Boolean
If true, all field names are added to the list. When false, the list
is emptied.
=back
Names are verified against the input fields after the first record is
sent. A C<Data::Record::Serialize::Error::Role::Base::fields> error is thrown
if non-existent fields are specified.
=item C<numify> => I<arrayref>|I<coderef>|Boolean
Specify fields that should be explicitly transformed into a number. It
defaults to C<false>, unless a particular encoder requires it
(e.g. C<json>). To avoid unnecessary conversion overhead, set this to
C<false> if you are sure that your data are appropriately numified, or
if only a few fields need to be transformed and can be done via the
L</format_fields> option.
B<numify> may be passed:
=over
=item * an arrayref of input field names
=item * a coderef
The coderef is called as
@input_field_names = $code->( $serializer_object )
=item * a Boolean
If true, all numeric fields are added to the list. When false, the list
is emptied.
=back
Names are verified against the input fields after the first record is
sent. A C<Data::Record::Serialize::Error::Role::Base::fields> error is thrown
if non-existent fields are specified.
=item C<stringify> => I<arrayref>|I<coderef>|Boolean
Specify fields that should be explicitly transformed into a string. It
defaults to C<false>, unless a particular encoder requires it
(e.g. C<json>). To avoid unnecessary conversion overhead, set this to
C<false> if you are sure that your data are appropriately stringified, or
if only a few fields need to be transformed and can be done via the
L</format_fields> option.
B<stringify> may be passed:
=over
=item * an arrayref of input field names
=item * a coderef
The coderef is called as
@input_field_names = $code->( $serializer_object )
=item * a Boolean
If true, all string fields are added to the list. When false, the list
is emptied.
=back
Names are verified against the input fields after the first record is
sent. A C<Data::Record::Serialize::Error::Role::Base::fields> error is thrown
if non-existent fields are specified.
=item C<format> => I<Boolean>
If true, format the output fields using the formats specified in the
L</format_fields> and/or L</format_types> options. The default is true.
=item C<format_fields> => I<hashref>
Specify transforms specific to fields. The hash keys are input field names;
each value is either a C<sprintf> style format or a coderef. The
transformations will be applied prior to encoding the record, but only
if the L</format> attribute is also set. Formats specified here
override those specified in L</format_types>.
The coderef will be called with the value to format as its first
argument, and should return the formatted value.
=item C<format_types> => I<hashref>
Specify transforms specific to types. The hash keys are field types
(C<N>, C<I>, C<S>, C<B>); each value is either a C<sprintf> style format or a coderef.
The transformations will be applied prior to encoding the record, but
only if the L</format> attribute is also set. Formats specified here
may be overridden for specific fields using the L</format_fields>
attribute.
The coderef will be called with the value to format as its first
argument, and should return the formatted value.
=item C<rename_fields> => I<hashref>
A hash mapping input to output field names. By default the input
field names are used unaltered.
=back
=head1 METHODS
For additional methods, see the following modules
=over
=item L</Data::Serialize::Record::Role::Base>
=item L</Data::Serialize::Record::Role::Default>
=item L</Data::Serialize::Record::Encode>
=item L</Data::Serialize::Record::Sink>,
=back
=head2 B<send>
$s->send( \%record );
Encode and send the record to the associated sink.
B<WARNING>: the passed hash is modified. If you need the original
contents, pass in a copy.
=head1 INTERNALS
=for Pod::Coverage make_variant
=head1 ATTRIBUTES
Object attributes are provided by the following modules
=over
=item L</Data::Serialize::Record::Role::Base>
=item L</Data::Serialize::Record::Role::Default>
=item L</Data::Serialize::Record::Encode>
=item L</Data::Serialize::Record::Sink>,
=back
=head1 EXAMPLES
=head2 Generate a JSON stream to the standard output stream
$s = Data::Record::Serialize->new( encode => 'json' );
=head2 Only output select fields
$s = Data::Record::Serialize->new(
encode => 'json',
fields => [ qw( obsid chip_id phi theta ) ],
);
=head2 Format numeric fields
$s = Data::Record::Serialize->new(
encode => 'json',
fields => [ qw( obsid chip_id phi theta ) ],
format => 1,
format_types => { N => '%0.4f' },
);
=head2 Override formats for specific fields
$s = Data::Record::Serialize->new(
encode => 'json',
fields => [ qw( obsid chip_id phi theta ) ],
format_types => { N => '%0.4f' },
format_fields => { obsid => '%05d' },
);
=head2 Rename fields
$s = Data::Record::Serialize->new(
encode => 'json',
fields => [ qw( obsid chip_id phi theta ) ],
format_types => { N => '%0.4f' },
format_fields => { obsid => '%05d' },
rename_fields => { chip_id => 'CHIP' },
);
=head2 Specify field types
$s = Data::Record::Serialize->new(
encode => 'json',
fields => [ qw( obsid chip_id phi theta ) ],
format_types => { N => '%0.4f' },
format_fields => { obsid => '%05d' },
rename_fields => { chip_id => 'CHIP' },
types => { obsid => 'N', chip_id => 'S', phi => 'N', theta => 'N' }'
);
=head2 Switch to an SQLite database in C<$dbname>
$s = Data::Record::Serialize->new(
encode => 'dbi',
dsn => [ 'SQLite', [ dbname => $dbname ] ],
table => 'stuff',
fields => [ qw( obsid chip_id phi theta ) ],
format_types => { N => '%0.4f' },
format_fields => { obsid => '%05d' },
rename_fields => { chip_id => 'CHIP' },
types => { obsid => 'N', chip_id => 'S', phi => 'N', theta => 'N' }'
);
=for Pod::Coverage BUILD
=head1 SUPPORT
=head2 Bugs
Please report any bugs or feature requests to bug-data-record-serialize@rt.cpan.org or through the web interface at: L<https://rt.cpan.org/Public/Dist/Display.html?Name=Data-Record-Serialize>
=head2 Source
Source is available at
https://gitlab.com/djerius/data-record-serialize
and may be cloned from
https://gitlab.com/djerius/data-record-serialize.git
=head1 SEE ALSO
Please see those modules/websites for more information related to this module.
=over 4
=item *
L<Data::Serializer>
=back
=head1 AUTHOR
Diab Jerius <djerius@cpan.org>
=head1 COPYRIGHT AND LICENSE
This software is Copyright (c) 2017 by Smithsonian Astrophysical Observatory.
This is free software, licensed under:
The GNU General Public License, Version 3, June 2007
=cut