org.grouplens.lenskit.core

Class LenskitRecommenderEngineLoader

java.lang.Object

org.grouplens.lenskit.core.LenskitRecommenderEngineLoader

public class LenskitRecommenderEngineLoader
extends Object

Load a pre-built recommender engine from a file.

Since:

2.1

Constructor Summary

Constructors

Constructor and Description
LenskitRecommenderEngineLoader()

Method Summary

Modifier and Type	Method and Description
LenskitRecommenderEngineLoader	addConfiguration(LenskitConfiguration config) Add a configuration to use when loading the configuration.
ClassLoader	getClassLoader() Get the configured class loader.
LenskitRecommenderEngine	load(File file) Load a recommender from a file.
LenskitRecommenderEngine	load(InputStream stream) Load a recommender engine from an input stream.
LenskitRecommenderEngineLoader	setClassLoader(ClassLoader classLoader) Set the class loader to use when reading the engine.
void	setCompressionMode(CompressionMode comp) Set the compression mode to use.
LenskitRecommenderEngineLoader	setValidationMode(EngineValidationMode mode) Set the validation mode for loading the recommender engine.

Methods inherited from class java.lang.Object

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

Constructor Detail

LenskitRecommenderEngineLoader

public LenskitRecommenderEngineLoader()

Method Detail

getClassLoader

public ClassLoader getClassLoader()

Get the configured class loader.

Returns:

The class loader that will be used when loading the engine.

setClassLoader

public LenskitRecommenderEngineLoader setClassLoader(ClassLoader classLoader)

Set the class loader to use when reading the engine.

Parameters:

classLoader - The class loader to use.

Returns:

The loader (for chaining).

addConfiguration

public LenskitRecommenderEngineLoader addConfiguration(LenskitConfiguration config)

Add a configuration to use when loading the configuration. The loaded graph will be post-processed to add in components bound by this configuration. If the engine was saved with excluded configurations, then this method should be used to provide configurations to reinstate any objects excluded from the saved model. The loader will throw an exception if the loaded model has any unresolved placeholders.

Parameters:

config - The configuration to add.

Returns:

The loader (for chaining).

setValidationMode

public LenskitRecommenderEngineLoader setValidationMode(EngineValidationMode mode)

Set the validation mode for loading the recommender engine. The default mode is EngineValidationMode.IMMEDIATE.

Parameters:

mode - The validation mode.

Returns:

The loader (for chaining).

setCompressionMode

public void setCompressionMode(CompressionMode comp)

Set the compression mode to use. The default is CompressionMode.AUTO.

Parameters:

comp - The compression mode.

load

public LenskitRecommenderEngine load(@WillClose
                                     InputStream stream)
                              throws IOException,
                                     RecommenderConfigurationException

Load a recommender engine from an input stream.

Note: this method is only capable of auto-detecting gzip-compressed data. If the compression mode is CompressionMode.AUTO, only gzip-compressed streams are supported. Set the compression mode manually if you are using XZ compression.

Parameters:

stream - The input stream.

Returns:

The deserialized recommender.

Throws:

IOException - if there is an error reading the input data.

RecommenderConfigurationException - if there is a configuration error with the deserialized recommender or the configurations applied to it.

load

public LenskitRecommenderEngine load(File file)
                              throws IOException,
                                     RecommenderConfigurationException

Load a recommender from a file.

Parameters:

file - The recommender model file to load.

Returns:

The recommender engine.

Throws:

IOException - if there is an error reading the input data.

RecommenderConfigurationException - if there is a configuration error with the deserialized recommender or the configurations applied to it.