package PGObject::Simple;
use 5.006;
use strict;
use warnings;
use Carp;
use PGObject;
=head1 NAME
PGObject::Simple - Minimalist stored procedure mapper based on LedgerSMB's DBObject
=head1 VERSION
Version 1.6
=cut
our $VERSION = '1.6';
=head1 SYNOPSIS
use PGObject::Simple;
my $obj = PGObject::Simple->new(%myhash);
$obj->set_dbh($dbh); # Database connection
To call a stored procedure with enumerated arguments.
my @results = $obj->call_procedure(
funcname => $funcname,
funcschema => $funcname,
args => [$arg1, $arg2, $arg3],
);
You can add something like a running total as well:
my @results = $obj->call_procedure(
funcname => $funcname,
funcschema => $funcname,
args => [$arg1, $arg2, $arg3],
running_funcs => [{agg => 'sum(amount)', alias => 'total'}],
);
To call a stored procedure with named arguments from a hashref. This is
typically done when mapping object properties in to stored procedure arguments.
my @results = $obj->call_dbmethod(
funcname => $funcname,
funcschema => $funcname,
running_funcs => [{agg => 'sum(amount)', alias => 'total'}],
);
To call a stored procedure with named arguments from a hashref with overrides.
my @results = $obj->call_dbmethod(
funcname => 'customer_save',
funcschema => 'public',
running_funcs => [{agg => 'sum(amount)', alias => 'total'}],
args => { id => undef }, # force to create new!
);
=head1 DESCRIPTION
PGObject::Simple a top-half object system for PGObject which is simple and
inspired by (and a subset functionally speaking of) the simple stored procedure
object method system of LedgerSMB 1.3. The framework discovers stored procedure
APIs and dispatches to them and can therefore be a base for application-specific
object models and much more.
PGObject::Simple is designed to be light-weight and yet robust glue between your
object model and the RDBMS's stored procedures. It works by looking up the
stored procedure arguments, stripping them of the conventional prefix 'in_', and
mapping what is left to object property names. Properties can be
overridden by passing in a hashrefs in the args named argument. Named arguments
there will be used in place of object properties.
This system is quite flexible, perhaps too much so, and it relies on the
database encapsulating its own logic behind self-documenting stored procedures
using consistent conventions. No function which is expected to be discovered can
be overloaded, and all arguments must be named for their object properties. For
this reason the use of this module fundamentally changes the contract of the
stored procedure from that of a fixed number of arguments in fixed types
contract to one where the name must be unique and the stored procedures must be
coded to the application's interface. This inverts the way we typically think
about stored procedures and makes them much more application friendly.
=head1 SUBROUTINES/METHODS
=head2 new
This constructs a new object. Basically it copies the incoming hash (one level
deep) and then blesses it. If the hash passed in has a dbh member, the dbh
is set to that. This does not set the function prefix, as this is assumed to
be done implicitly by subclasses.
=cut
sub new {
my ($self) = shift @_;
my %args = @_;
my $ref = {};
$ref->{$_} = $args{$_} for keys %args;
bless ($ref, $self);
$ref->set_dbh($ref->{dbh}) if $ref->{dbh};
return $ref;
}
=head2 set_dbh($dbh)
Sets the database handle (needs DBD::Pg 2.0 or later) to $dbh
=cut
sub set_dbh {
my ($self, $dbh) = @_;
$self->{_DBH} = $dbh;
}
=head2 _set_funcprefix
This sets the default funcprefix for future calls. The funcprefix can still be
overridden by passing in an explicit '' in a call. This is used to "claim" a
certain set of stored procedures in the database for use by an object.
It is semi-private, intended to be called by subclasses directly, perhaps in
constructors, but not from outside the object.
=cut
sub _set_funcprefix {
my ($self, $funcprefix) = @_;
$self->{_func_prefix} = $funcprefix;
}
=head2 _set_funcschema
This sets the default funcschema for future calls. This is overwridden by
per-call arguments, (PGObject::Util::DBMethod provides for such overrides on a
per-method basis).
=cut
sub _set_funcschema {
my ($self, $funcschema) = @_;
$self->{_func_schema} = $funcschema;
}
=head2 _set_registry
This sets the registry for future calls. The idea here is that this allows for
application object model wrappers to set which registry they are using, both for
predictability and ensuring that interoperability is possible.
=cut
sub _set_registry {
my ($self, $registry) = @_;
$self->{_registry} = $registry;
}
=head2 call_dbmethod
Does a straight-forward mapping (as described below) to the stored procedure
arguments. Stored procedure arguments are looked up, a leading 'in_' is
stripped off where it exists, and the remaining string mapped back to an
object property. The $args{args} hashref can be used to override arguments by
name. Unknown properties are handled simply by passing a NULL in, so the
stored procedures should be prepared to handle these.
As with call_procedure below, this returns a single hashref when called in a
scalar context, and a list of hashrefs when called in a list context.
=cut
sub call_dbmethod {
my ($self) = shift @_;
my %args = @_;
croak 'No function name provided' unless $args{funcname};
if (eval { $self->isa(__PACKAGE__) } and ref $self){
$args{dbh} = $self->{_DBH} if $self->{_DBH} and !$args{dbh};
$args{funcprefix} = $self->{_func_prefix} if !defined $args{funcprefix};
$args{funcschema} = $self->{_func_schema} if !defined $args{funcschema};
}
$args{funcprefix} ||= '';
my $info = PGObject->function_info(%args);
my $dbargs = [];
for my $arg (@{$info->{args}}){
$arg->{name} =~ s/^in_//;
my $db_arg = $self->{$arg->{name}} if ref $self;
if ($args{args}->{$arg->{name}}){
$db_arg = $args{args}->{$arg->{name}};
}
if (eval {$db_arg->can('to_db')}){
$db_arg = $db_arg->to_db;
}
if ($arg->{type} eq 'bytea'){
$db_arg = { type => 'bytea', value => $db_arg};
}
push @$dbargs, $db_arg;
}
$args{args} = $dbargs;
# The conditional return is necessary since the object may carry a registry
# --CT
return $self->call_procedure(%args) if ref $self;
return __PACKAGE__->call_procedure(%args);
}
=head2 call_procedure
This is a lightweight wrapper around PGObject->call_procedure which merely
passes the currently attached db connection in. We use the previously set
funcprefix and dbh by default but other values can be passed in to override the
default object's values.
This returns a single hashref when called in a scalar context, and a list of
hashrefs when called in a list context. When called in a scalar context it
simply returns the single first row returned.
=cut
sub call_procedure {
my ($self) = shift @_;
my %args = @_;
if (eval { $self->isa(__PACKAGE__) } and ref $self ){
$args{funcprefix} = $self->{_func_prefix} if !defined $args{funcprefix};
$args{funcschema} = $self->{_func_schema} if !defined $args{funcschema};
$args{registry} = $self->{_registry} if !defined $args{registry};
$args{dbh} = $self->{_DBH} if $self->{_DBH} and !$args{dbh};
}
$args{funcprefix} ||= '';
croak 'No DB handle provided' unless $args{dbh};
my @rows = PGObject->call_procedure(%args);
return shift @rows unless wantarray;
return @rows;
}
=head1 WRITING CLASSES WITH PGObject::Simple
Unlike PGObject, which is only loosely tied to the functionality in question
and presumes that relevant information will be passed over a functional
interface, PGObject is a specific framework for object-oriented coding in Perl.
It can therefore be used alone or with other modules to provide quite a bit of
functionality.
A PGObject::Simple object is a blessed hashref with no gettors or setters. This
is thus ideal for cases where you are starting and just need some quick mappings
of stored procedures to hashrefs. You reference properties simply with the
$object->{property} syntax. There is very little encapsulation in objects, and
very little abstraction except when it comes to the actual stored procedure
interfaces. In essence, PGObject::Simple generally assumes that the actual
data structure is essentially a public interface between the database and
whatever else is going on with the application.
The general methods can then wrap call_procedure and call_dbmethod calls,
mapping out to stored procedures in the database.
Stored procedures must be written to relatively exacting specifications.
Arguments must be named, with names prefixed optionally with 'in_' (if the
property name starts with 'in_' properly one must also prefix it).
An example of a simple stored procedure might be:
CREATE OR REPLACE FUNCTION customer_get(in_id int) returns customer
RETURNS setof customer language sql as $$
select * from customer where id = $1;
$$;
This stored procedure could then be called with any of:
$obj->call_dbmethod(
funcname => 'customer_get',
); # retrieve the customer with the $obj->{id} id
$obj->call_dbmethod(
funcname => 'customer_get',
args => {id => 3 },
); # retrieve the customer with the id of 3 regardless of $obj->{id}
$obj->call_procedure(
funcname => 'customer_get',
args => [3],
);
=head1 AUTHOR
Chris Travers, C<< <chris.travers at gmail.com> >>
=head1 BUGS
Please report any bugs or feature requests to C<bug-pgobject-simple at rt.cpan.org>, or through
the web interface at L<http://rt.cpan.org/NoAuth/ReportBug.html?Queue=PGObject-Simple>. I will be notified, and then you'll
automatically be notified of progress on your bug as I make changes.
=head1 SUPPORT
You can find documentation for this module with the perldoc command.
perldoc PGObject::Simple
You can also look for information at:
=over 4
=item * RT: CPAN's request tracker (report bugs here)
L<http://rt.cpan.org/NoAuth/Bugs.html?Dist=PGObject-Simple>
=item * AnnoCPAN: Annotated CPAN documentation
L<http://annocpan.org/dist/PGObject-Simple>
=item * CPAN Ratings
L<http://cpanratings.perl.org/d/PGObject-Simple>
=item * Search CPAN
L<http://search.cpan.org/dist/PGObject-Simple/>
=back
=head1 ACKNOWLEDGEMENTS
=head1 LICENSE AND COPYRIGHT
Copyright 2013 Chris Travers.
Redistribution and use in source and compiled forms with or without
modification, are permitted provided that the following conditions are met:
=over
=item
Redistributions of source code must retain the above
copyright notice, this list of conditions and the following disclaimer as the
first lines of this file unmodified.
=item
Redistributions in compiled form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
source code, documentation, and/or other materials provided with the
distribution.
=back
THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) "AS IS" AND
ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR
ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON
ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
=cut
1; # End of PGObject::Simple