opennlp.tools.util.model

Class BaseModel

java.lang.Object

opennlp.tools.util.model.BaseModel

All Implemented Interfaces:

ArtifactProvider

Direct Known Subclasses:

ChunkerModel, CorefModel, DoccatModel, ParserModel, POSModel, SentenceModel, TokenizerModel, TokenNameFinderModel

public abstract class BaseModel
extends Object
implements ArtifactProvider

This model is a common based which can be used by the components model classes. TODO: Provide sub classes access to serializers already in constructor

Field Summary

Fields

Modifier and Type	Field and Description
protected Map<String,Object>	artifactMap
protected static String	FACTORY_NAME
protected static String	MANIFEST_ENTRY
protected BaseToolFactory	toolFactory
static String	TRAINING_CUTOFF_PROPERTY
static String	TRAINING_EVENTHASH_PROPERTY
static String	TRAINING_ITERATIONS_PROPERTY

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	BaseModel(String componentName, File modelFile)
protected	BaseModel(String componentName, InputStream in) Initializes the current instance.
protected	BaseModel(String componentName, String languageCode, Map<String,String> manifestInfoEntries) Initializes the current instance.
protected	BaseModel(String componentName, String languageCode, Map<String,String> manifestInfoEntries, BaseToolFactory factory) Initializes the current instance.
protected	BaseModel(String componentName, URL modelURL)

Method Summary

Modifier and Type	Method and Description
protected void	checkArtifactMap() Checks the artifact map.
protected static Map<String,ArtifactSerializer>	createArtifactSerializers()
protected void	createArtifactSerializers(Map<String,ArtifactSerializer> serializers) Registers all ArtifactSerializer for their artifact file name extensions.
<T> T	getArtifact(String key) Gets an artifact by name
protected ArtifactSerializer	getArtifactSerializer(String resourceName)
protected Class<? extends BaseToolFactory>	getDefaultFactory() Sub-classes should override this method if their module has a default BaseToolFactory sub-class.
String	getLanguage() Retrieves the language code of the material which was used to train the model or x-unspecified if non was set.
String	getManifestProperty(String key) Retrieves the value to the given key from the manifest.properties entry.
Version	getVersion() Retrieves the OpenNLP version which was used to create the model.
boolean	isLoadedFromSerialized() Indicates if this provider was loaded from serialized.
void	serialize(OutputStream out) Serializes the model to the given OutputStream.
protected void	setManifestProperty(String key, String value) Sets a given value for a given key to the manifest.properties entry.
protected void	validateArtifactMap() Validates the parsed artifacts.

Methods inherited from class java.lang.Object

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

Field Detail

MANIFEST_ENTRY

protected static final String MANIFEST_ENTRY

See Also:

Constant Field Values

FACTORY_NAME

protected static final String FACTORY_NAME

See Also:

Constant Field Values

TRAINING_CUTOFF_PROPERTY

public static final String TRAINING_CUTOFF_PROPERTY

See Also:

Constant Field Values

TRAINING_ITERATIONS_PROPERTY

public static final String TRAINING_ITERATIONS_PROPERTY

See Also:

Constant Field Values

TRAINING_EVENTHASH_PROPERTY

public static final String TRAINING_EVENTHASH_PROPERTY

See Also:

Constant Field Values

artifactMap

protected final Map<String,Object> artifactMap

toolFactory

protected BaseToolFactory toolFactory

Constructor Detail

BaseModel

protected BaseModel(String componentName,
                    String languageCode,
                    Map<String,String> manifestInfoEntries,
                    BaseToolFactory factory)

Initializes the current instance. The sub-class constructor should call the method checkArtifactMap() to check the artifact map is OK.

Sub-classes will have access to custom artifacts and serializers provided by the factory.

Parameters:

componentName - the component name

languageCode - the language code

manifestInfoEntries - additional information in the manifest

factory - the factory

BaseModel

protected BaseModel(String componentName,
                    String languageCode,
                    Map<String,String> manifestInfoEntries)

Initializes the current instance. The sub-class constructor should call the method checkArtifactMap() to check the artifact map is OK.

Parameters:

componentName - the component name

languageCode - the language code

manifestInfoEntries - additional information in the manifest

BaseModel

protected BaseModel(String componentName,
                    InputStream in)
             throws IOException,
                    InvalidFormatException

Initializes the current instance.

Parameters:

componentName - the component name

in - the input stream containing the model

Throws:

IOException

InvalidFormatException

BaseModel

protected BaseModel(String componentName,
                    File modelFile)
             throws IOException,
                    InvalidFormatException

Throws:

IOException

InvalidFormatException

BaseModel

protected BaseModel(String componentName,
                    URL modelURL)
             throws IOException,
                    InvalidFormatException

Throws:

IOException

InvalidFormatException

Method Detail

getDefaultFactory

protected Class<? extends BaseToolFactory> getDefaultFactory()

Sub-classes should override this method if their module has a default BaseToolFactory sub-class.

Returns:

the default BaseToolFactory for the module, or null if none.

getArtifactSerializer

protected ArtifactSerializer getArtifactSerializer(String resourceName)

createArtifactSerializers

protected static Map<String,ArtifactSerializer> createArtifactSerializers()

createArtifactSerializers

protected void createArtifactSerializers(Map<String,ArtifactSerializer> serializers)

Registers all ArtifactSerializer for their artifact file name extensions. The registered ArtifactSerializer are used to create and serialize resources in the model package. Override this method to register custom ArtifactSerializers. Note: Subclasses should generally invoke super.createArtifactSerializers at the beginning of this method. This method is called during construction.

Parameters:

serializers - the key of the map is the file extension used to lookup the ArtifactSerializer.

validateArtifactMap

protected void validateArtifactMap()
                            throws InvalidFormatException

Validates the parsed artifacts. If something is not valid subclasses should throw an InvalidFormatException. Note: Subclasses should generally invoke super.validateArtifactMap at the beginning of this method.

Throws:

InvalidFormatException

checkArtifactMap

protected void checkArtifactMap()

Checks the artifact map.

A subclass should call this method from a constructor which accepts the individual artifact map items, to validate that these items form a valid model.

If the artifacts are not valid an IllegalArgumentException will be thrown.

getManifestProperty

public final String getManifestProperty(String key)

Retrieves the value to the given key from the manifest.properties entry.

Specified by:

getManifestProperty in interface ArtifactProvider

Parameters:

key -

Returns:

the value

setManifestProperty

protected final void setManifestProperty(String key,
                                         String value)

Sets a given value for a given key to the manifest.properties entry.

Parameters:

key -

value -

getLanguage

public final String getLanguage()

Retrieves the language code of the material which was used to train the model or x-unspecified if non was set.

Specified by:

getLanguage in interface ArtifactProvider

Returns:

the language code of this model

getVersion

public final Version getVersion()

Retrieves the OpenNLP version which was used to create the model.

Returns:

the version

serialize

public final void serialize(OutputStream out)
                     throws IOException

Serializes the model to the given OutputStream.

Parameters:

out - stream to write the model to

Throws:

IOException

getArtifact

public <T> T getArtifact(String key)

Description copied from interface: ArtifactProvider

Gets an artifact by name

Specified by:

getArtifact in interface ArtifactProvider

isLoadedFromSerialized

public boolean isLoadedFromSerialized()

Description copied from interface: ArtifactProvider

Indicates if this provider was loaded from serialized. It is useful, for example, while validating artifacts: you can skip the time consuming ones if they where already validated during the serialization.

Specified by:

isLoadedFromSerialized in interface ArtifactProvider

Returns:

true if this model was loaded from serialized