org.lenskit

Class LenskitRecommenderEngine

java.lang.Object

org.lenskit.LenskitRecommenderEngine

All Implemented Interfaces:

java.io.Serializable, RecommenderEngine

public final class LenskitRecommenderEngine
extends java.lang.Object
implements RecommenderEngine, java.io.Serializable

LensKit implementation of a recommender engine. It uses containers set up by the LenskitConfiguration to set up actual recommenders, and can build multiple recommenders from the same model.

If you just want to quick create a recommender for evaluation or testing, consider using LenskitRecommender.build(LenskitConfiguration). For more control, use LenskitRecommenderEngineBuilder; to load a pre-built recommender engine, use LenskitRecommenderEngineLoader.

See Also:

LenskitConfiguration, LenskitRecommender, LenskitRecommenderEngineBuilder, LenskitRecommenderEngineLoader, Serialized Form

Compatibility

Public

Constructor Summary

Constructors

Constructor and Description
LenskitRecommenderEngine(DAGNode<Component,Dependency> graph, boolean instantiable) Build an engine encapsulating a dependency graph.

Method Summary

Modifier and Type	Method and Description
static LenskitRecommenderEngine	build(LenskitConfiguration config) Build a LensKit recommender engine from a configuration.
static LenskitRecommenderEngine	build(LenskitConfiguration config, DataAccessObject dao) Build a LensKit recommender engine from a configuration.
LenskitRecommender	createRecommender() Deprecated. Use createRecommender(DataAccessObject)
LenskitRecommender	createRecommender(DataAccessObject dao) Create a LensKit recommender.
LenskitRecommender	createRecommender(LenskitConfiguration config) Construct a recommender with some additional configuration.
<T> T	getComponent(java.lang.Class<T> type) Get the component of a particular type, if one is already instantiated.
DAGNode<Component,Dependency>	getGraph() Get the dependency graph of the recommender engine.
boolean	isInstantiable() Query whether this engine is instantiable.
static LenskitRecommenderEngine	load(java.io.File file) Create a new recommender engine by reading a previously serialized engine from the given file.
static LenskitRecommenderEngine	load(java.io.File file, java.lang.ClassLoader loader) Deprecated. Use LenskitRecommenderEngineLoader for sophisticated loading.
static LenskitRecommenderEngine	load(java.io.InputStream input) Create a new recommender engine by reading a previously serialized engine from the given input stream.
static LenskitRecommenderEngineBuilder	newBuilder() Create a new recommender engine builder.
static LenskitRecommenderEngineLoader	newLoader() Create a new recommender engine loader.
void	write(java.io.File file) Write the state of this recommender engine to the given file so that it can be recreated later using another DAOFactory.
void	write(java.io.File file, CompressionMode compressed) Write the state of this recommender engine to the given file so that it can be recreated later using another DAOFactory.
void	write(java.io.OutputStream stream) Write the state of this recommender engine to the given stream so that it can be recreated later using another DAOFactory.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

LenskitRecommenderEngine

public LenskitRecommenderEngine(@Nonnull
                                DAGNode<Component,Dependency> graph,
                                boolean instantiable)

Build an engine encapsulating a dependency graph. You generally do not want to use this - use LenskitRecommenderEngineBuilder instead.

Parameters:

graph - The graph.

instantiable - true if the recommender can be instantiated as-is, false otherwise.

Method Detail

load

public static LenskitRecommenderEngine load(java.io.File file)
                                     throws java.io.IOException,
                                            RecommenderConfigurationException

Create a new recommender engine by reading a previously serialized engine from the given file. The new engine will be identical to the old except it will use the new DAOFactory. It is assumed that the file was created by using write(OutputStream). Classes will be loaded using a default class loader.

Parameters:

file - The file from which to load the engine.

Returns:

The loaded recommender engine.

Throws:

java.io.IOException - If there is an error reading from the file.

RecommenderConfigurationException - If the configuration cannot be used.

load

@Deprecated
public static LenskitRecommenderEngine load(java.io.File file,
                                                        java.lang.ClassLoader loader)
                                                 throws java.io.IOException,
                                                        RecommenderConfigurationException

Deprecated. Use LenskitRecommenderEngineLoader for sophisticated loading.

Create a new recommender engine by reading a previously serialized engine from the given file. The new engine will be identical to the old except it will use the new DAOFactory. It is assumed that the file was created by using write(OutputStream).

Parameters:

file - The file from which to load the engine.

loader - The class loader to load from (null to use a default class loader).

Returns:

The loaded recommender engine.

Throws:

java.io.IOException - If there is an error reading from the file.

RecommenderConfigurationException - If the configuration cannot be used.

load

public static LenskitRecommenderEngine load(java.io.InputStream input)
                                     throws java.io.IOException,
                                            RecommenderConfigurationException

Create a new recommender engine by reading a previously serialized engine from the given input stream. The new engine will be identical to the old. It is assumed that the file was created by using write(OutputStream). Classes will be loaded using a default class loader.

Parameters:

input - The stream from which to load the engine.

Returns:

The loaded recommender engine.

Throws:

java.io.IOException - If there is an error reading from the file.

RecommenderConfigurationException - If the configuration cannot be used.

write

public void write(@Nonnull
                  java.io.File file)
           throws java.io.IOException

Write the state of this recommender engine to the given file so that it can be recreated later using another DAOFactory. This uses default object serialization so if the factory has a PicoContainer or session bindings containing non-serializable types, this will fail.

Parameters:

file - The file to write the rec engine to.

Throws:

java.io.IOException - if there is an error serializing the engine.

See Also:

write(OutputStream)

write

public void write(@Nonnull
                  java.io.File file,
                  CompressionMode compressed)
           throws java.io.IOException

Write the state of this recommender engine to the given file so that it can be recreated later using another DAOFactory. This uses default object serialization so if the factory has a PicoContainer or session bindings containing non-serializable types, this will fail.

Parameters:

file - The file to write the rec engine to.

compressed - Whether to compress the output file.

Throws:

java.io.IOException - if there is an error serializing the engine.

See Also:

write(OutputStream)

write

public void write(@Nonnull @WillClose
                  java.io.OutputStream stream)
           throws java.io.IOException

Write the state of this recommender engine to the given stream so that it can be recreated later using another DAOFactory. This uses default object serialization so if the factory has session bindings containing non-serializable types, this will fail.

Parameters:

stream - The file to write the rec engine to.

Throws:

java.io.IOException - if there is an error serializing the engine.

See Also:

load(InputStream)

createRecommender

@Deprecated
public LenskitRecommender createRecommender()

Deprecated. Use createRecommender(DataAccessObject)

Create a recommender.

Specified by:

createRecommender in interface RecommenderEngine

Returns:

The recommender

createRecommender

public LenskitRecommender createRecommender(LenskitConfiguration config)
                                     throws RecommenderConfigurationException

Construct a recommender with some additional configuration. This can be used to do things like add data source configuration on a per-recommender, rather than per-engine, basis.

Parameters:

config - The configuration to adjust the recommender.

Returns:

The constructed recommender.

Throws:

RecommenderConfigurationException - if there is an error configuring the recommender.

createRecommender

public LenskitRecommender createRecommender(@WillNotClose
                                            DataAccessObject dao)
                                     throws RecommenderBuildException

Create a LensKit recommender.

Parameters:

dao - The data access object

Returns:

The constructed recommender.

Throws:

RecommenderBuildException

isInstantiable

public boolean isInstantiable()

Query whether this engine is instantiable. Instantiable recommenders have all their placeholders removed and are ready to instantiate.

Returns:

true if the recommender is instantiable.

getGraph

@Nonnull
public DAGNode<Component,Dependency> getGraph()

Get the dependency graph of the recommender engine.

Returns:

The dependency graph.

getComponent

@Nullable
public <T> T getComponent(java.lang.Class<T> type)

Get the component of a particular type, if one is already instantiated. This is useful to extract pre-built models from serialized recommender engines, for example.

Parameters:

type - The required component type.

T - The required component type.

Returns:

The component instance, or null if no instance can be retreived (either because no such component is configured, or it is not yet instantiated).

build

public static LenskitRecommenderEngine build(LenskitConfiguration config)
                                      throws RecommenderBuildException

Build a LensKit recommender engine from a configuration. The resulting recommender is independent of any subsequent modifications to the configuration.

Parameters:

config - The configuration.

Returns:

The recommender engine.

Throws:

RecommenderBuildException

build

public static LenskitRecommenderEngine build(LenskitConfiguration config,
                                             DataAccessObject dao)
                                      throws RecommenderBuildException

Build a LensKit recommender engine from a configuration. The resulting recommender is independent of any subsequent modifications to the configuration.

Parameters:

config - The configuration.

dao - The data access object

Returns:

The recommender engine.

Throws:

RecommenderBuildException

newBuilder

public static LenskitRecommenderEngineBuilder newBuilder()

Create a new recommender engine builder.

Returns:

A new recommender engine builder.

newLoader

public static LenskitRecommenderEngineLoader newLoader()

Create a new recommender engine loader.

Returns:

A new recommender engine loader.