SlideShare a Scribd company logo

About Data::ObjectDriver

•Download as KEY, PDF•
•
Yoshiki Kurihara
Yoshiki Kurihara

Data::ObjectDriver is a simple and transparent data interface library with caching capabilities. It provides an object-oriented interface to database tables and supports features like master-slave replication, partitioning, and caching using Memcached. Classes define the schema and connection details. Methods provide CRUD functionality and relationships. Custom drivers can be created to support different database architectures or caching strategies.

Related slideshows

Puppet and the HashiCorp Suite
Puppet and the HashiCorp SuitePuppet and the HashiCorp Suite
Puppet and the HashiCorp Suite
 
Remedie: Building a desktop app with HTTP::Engine, SQLite and jQuery
Remedie: Building a desktop app with HTTP::Engine, SQLite and jQueryRemedie: Building a desktop app with HTTP::Engine, SQLite and jQuery
Remedie: Building a desktop app with HTTP::Engine, SQLite and jQuery
 
The worst Ruby codes I’ve seen in my life - RubyKaigi 2015
The worst Ruby codes I’ve seen in my life - RubyKaigi 2015The worst Ruby codes I’ve seen in my life - RubyKaigi 2015
The worst Ruby codes I’ve seen in my life - RubyKaigi 2015
 
1 of 25
Data::ObjectDriver id:clouder Yokohama.pm #5
About D::OD • Author: Benjamin Trott • Recently Version: 0.06 • Simple, transparent data interface, with caching. • Based on MT::ObjectDriver in MT. Now MT included D::OD.
D::OD features • Built-in supportRAM and Apache inPartitioning. Support Memcached, Caching and caching. • Have to support master-slaver_handle()/rw_handle(). Can change process for read/write using structure in mind. • Implementation is becauseso less model feature. But implement by myself, thin, east-to-use. of Has ‘has_a’ but not has ‘has_many’.
Class structures • Driver definition about how to connection to Class for class db and cache server, and partitioning rules. • Object classwhat you call. The model class Class for definition about how to treat data on tables.
Class structures • Other classes - D::OD::ResultSet In the middle of an implementation? Do not use in MT. - D::OD::Profiler Simple profiler. - D::OD::GearmanDBI I do not know how to use;)
Simple usase • Make object class for table Make sub-class of D::OD::BaseObject, and set table information using install_properties(). ‘driver’ is D::OD::Driver::DBI.
Simple usage package Artist; use strict; use base qw( Data::ObjectDriver::BaseObject ); __PACKAGE__->install_properties( datasource => 'artist', columns => [ qw( id name orig_name band_id ) ], primary_key => 'id', driver => Data::ObjectDriver::Driver::DBI->new( %DB_INFO ), ); 1;
CRUD and etc • Create my artist = Artist->new( name => ' ', fullname => ' II ' ); $artist->save; # or Artist->bulk_insert( [col1, col2], [ [d1, d2], [d1, d2] ]);
CRUD and etc • Read my $artist = Artist->lookup(1); print $artist->name; # or $artist_iter = Artist->search( { name => ' ' } ); @artists = Artist->search( { name => ' ' ); # or $artists_ref = Artist->lookup_multi( [ 1, 2, 3 ] );
CRUD and etc • Update $artist->name( ' ' ); $artist->save;
CRUD and etc • Delete $artist->remove;
CRUD and etc • has_a() __PACKAGE__->has_a( { class => 'Band', column => 'band_id', cached => 1, } );
CRUD and etc • add_trigger() post_load pre_search pre_insert pre_save post_save post_insert pre_update post_update pre_remove post_remove post_inflate __PACKAGE__->add_trigger( pre_insert => sub { my ( $obj, $orig_obj ) = @_; ... }, );
Caching • Only change driver If there is not cache, connect to db using D::OD::Driver::DBI.
Caching package Artist; use strict; use base qw( Data::ObjectDriver::BaseObject ); __PACKAGE__->install_properties( ... driver => Data::ObjectDriver::Driver::Cache::Memcached->new( cache => Cache::Memcached->new( servers => @servers ), fallback => Data::ObjectDriver::Driver::DBI->new( %DB_INFO ), ), ... ); 1;
Master-Slave structure • Only override r_handle().in read process, r_handle() is method that execute so this method is used to connect to slave database.
Master-Slave structure • Object class package Artist; use strict; use base qw( Data::ObjectDriver::BaseObject ); __PACKAGE__->install_properties( ... driver => Data::ObjectDriver::Driver::Cache::Memcached->new( cache => Cache::Memcached->new( servers => @servers ), fallback => ReplDriver->new( %DB_INFO, slaves => [ slave01, ... ] ), ), ... );
Master-Slave structure • Driver class package ReplDriver; use strict; use base qw( Data::ObjectDriver::Driver::DBI ); __PACKAGE__->mk_accessors( qw( slaves ) ); sub init { my $driver = shift; my %param = @_; $driver->slaves( delete $param{ slaves } ); $driver->SUPER::init( %param ); return $driver; } # cont.
Master-Slave structure • Driver class(cont.) # cont. sub r_handle { my $driver = shift; my $db = shift || 'main'; for my $slave ( shuffle @{ $driver->slaves } ) { # connect to $slave my $dbh = DBI->connect( $slave->{DB_INFO} ); $driver->dbd->init_dbh($dbh); return $dbh; } $driver->rw_handle($db); } 1;
Partitioning package CD; use strict; use base qw( Data::ObjectDriver::BaseObject ); __PACKAGE__->install_properties( datasource => 'cd', columns => [ qw( artist_id id title ) ], primary_key => [ qw( artist_id id ) ], driver => PartitionDriver->driver, ); 1;
Partitioning package PartitionDriver; use strict; sub driver { my $fallback = Data::ObjectDriver::Driver::Partition->new( get_driver => &find_partition, ); Data::ObjectDriver::Driver::Cache::Memcached->new( cache => Cache::Memcached->new( servers => @servers ), fallback => $fallback, ), } # cont.
Partitioning # cont. sub find_partition { my ( $terms, $args ) = @_; my $artist = Artist->lookup( $terms->{ artist_id } ); return ReplDriver->new( %{ $artist->partition_obj->master }, slaves => $artist->partition_obj->slaves, pk_generator => &pk_generator, ); } sub pk_generator { my $obj = shift; $obj->id( generate_id() ); 1; }, 1;
Partitioning my $cd = CD->new( artist_id => 1, title => ' ' ); $cd->save; lookup() is depends on PartitionDriver implementation in partitioning.
At the end, I wish... • Built-in support pager using Data::Page. • Wants count() and more useful methods. • Hard to execute simple SQL. (Just do using D::OD::SQL?) • And hard to execute ‘JOIN’.
Fin.

More Related Content

About Data::ObjectDriver

  • 1. Data::ObjectDriver id:clouder Yokohama.pm #5
  • 2. About D::OD • Author: Benjamin Trott • Recently Version: 0.06 • Simple, transparent data interface, with caching. • Based on MT::ObjectDriver in MT. Now MT included D::OD.
  • 3. D::OD features • Built-in supportRAM and Apache inPartitioning. Support Memcached, Caching and caching. • Have to support master-slaver_handle()/rw_handle(). Can change process for read/write using structure in mind. • Implementation is becauseso less model feature. But implement by myself, thin, east-to-use. of Has ‘has_a’ but not has ‘has_many’.
  • 4. Class structures • Driver definition about how to connection to Class for class db and cache server, and partitioning rules. • Object classwhat you call. The model class Class for definition about how to treat data on tables.
  • 5. Class structures • Other classes - D::OD::ResultSet In the middle of an implementation? Do not use in MT. - D::OD::Profiler Simple profiler. - D::OD::GearmanDBI I do not know how to use;)
  • 6. Simple usase • Make object class for table Make sub-class of D::OD::BaseObject, and set table information using install_properties(). ‘driver’ is D::OD::Driver::DBI.
  • 7. Simple usage package Artist; use strict; use base qw( Data::ObjectDriver::BaseObject ); __PACKAGE__->install_properties( datasource => 'artist', columns => [ qw( id name orig_name band_id ) ], primary_key => 'id', driver => Data::ObjectDriver::Driver::DBI->new( %DB_INFO ), ); 1;
  • 8. CRUD and etc • Create my artist = Artist->new( name => ' ', fullname => ' II ' ); $artist->save; # or Artist->bulk_insert( [col1, col2], [ [d1, d2], [d1, d2] ]);
  • 9. CRUD and etc • Read my $artist = Artist->lookup(1); print $artist->name; # or $artist_iter = Artist->search( { name => ' ' } ); @artists = Artist->search( { name => ' ' ); # or $artists_ref = Artist->lookup_multi( [ 1, 2, 3 ] );
  • 10. CRUD and etc • Update $artist->name( ' ' ); $artist->save;
  • 11. CRUD and etc • Delete $artist->remove;
  • 12. CRUD and etc • has_a() __PACKAGE__->has_a( { class => 'Band', column => 'band_id', cached => 1, } );
  • 13. CRUD and etc • add_trigger() post_load pre_search pre_insert pre_save post_save post_insert pre_update post_update pre_remove post_remove post_inflate __PACKAGE__->add_trigger( pre_insert => sub { my ( $obj, $orig_obj ) = @_; ... }, );
  • 14. Caching • Only change driver If there is not cache, connect to db using D::OD::Driver::DBI.
  • 15. Caching package Artist; use strict; use base qw( Data::ObjectDriver::BaseObject ); __PACKAGE__->install_properties( ... driver => Data::ObjectDriver::Driver::Cache::Memcached->new( cache => Cache::Memcached->new( servers => @servers ), fallback => Data::ObjectDriver::Driver::DBI->new( %DB_INFO ), ), ... ); 1;
  • 16. Master-Slave structure • Only override r_handle().in read process, r_handle() is method that execute so this method is used to connect to slave database.
  • 17. Master-Slave structure • Object class package Artist; use strict; use base qw( Data::ObjectDriver::BaseObject ); __PACKAGE__->install_properties( ... driver => Data::ObjectDriver::Driver::Cache::Memcached->new( cache => Cache::Memcached->new( servers => @servers ), fallback => ReplDriver->new( %DB_INFO, slaves => [ slave01, ... ] ), ), ... );
  • 18. Master-Slave structure • Driver class package ReplDriver; use strict; use base qw( Data::ObjectDriver::Driver::DBI ); __PACKAGE__->mk_accessors( qw( slaves ) ); sub init { my $driver = shift; my %param = @_; $driver->slaves( delete $param{ slaves } ); $driver->SUPER::init( %param ); return $driver; } # cont.
  • 19. Master-Slave structure • Driver class(cont.) # cont. sub r_handle { my $driver = shift; my $db = shift || 'main'; for my $slave ( shuffle @{ $driver->slaves } ) { # connect to $slave my $dbh = DBI->connect( $slave->{DB_INFO} ); $driver->dbd->init_dbh($dbh); return $dbh; } $driver->rw_handle($db); } 1;
  • 20. Partitioning package CD; use strict; use base qw( Data::ObjectDriver::BaseObject ); __PACKAGE__->install_properties( datasource => 'cd', columns => [ qw( artist_id id title ) ], primary_key => [ qw( artist_id id ) ], driver => PartitionDriver->driver, ); 1;
  • 21. Partitioning package PartitionDriver; use strict; sub driver { my $fallback = Data::ObjectDriver::Driver::Partition->new( get_driver => &find_partition, ); Data::ObjectDriver::Driver::Cache::Memcached->new( cache => Cache::Memcached->new( servers => @servers ), fallback => $fallback, ), } # cont.
  • 22. Partitioning # cont. sub find_partition { my ( $terms, $args ) = @_; my $artist = Artist->lookup( $terms->{ artist_id } ); return ReplDriver->new( %{ $artist->partition_obj->master }, slaves => $artist->partition_obj->slaves, pk_generator => &pk_generator, ); } sub pk_generator { my $obj = shift; $obj->id( generate_id() ); 1; }, 1;
  • 23. Partitioning my $cd = CD->new( artist_id => 1, title => ' ' ); $cd->save; lookup() is depends on PartitionDriver implementation in partitioning.
  • 24. At the end, I wish... • Built-in support pager using Data::Page. • Wants count() and more useful methods. • Hard to execute simple SQL. (Just do using D::OD::SQL?) • And hard to execute ‘JOIN’.
  • 25. Fin.

Editor's Notes