Package opennlp.tools.ml.libsvm.doccat

Class SvmDoccatModel

java.lang.Object

opennlp.tools.ml.libsvm.doccat.SvmDoccatModel

All Implemented Interfaces:

Serializable

public class SvmDoccatModel
extends Object
implements Serializable

A model for SVM-based document categorization. This model wraps a zlibsvm SvmModel together with the feature vocabulary, category label mappings, corpus statistics, and configuration required for classification.

Persistence uses Java object serialization via serialize(OutputStream) and deserialize(InputStream). Reads are guarded by an ObjectInputFilter that allow-lists only the classes reachable from a legitimate SvmDoccatModel graph and bounds graph depth, references, and array length. Foreign payloads are rejected with InvalidClassException before being materialised.

Treat the filter as defense-in-depth, not as a license to deserialize from untrusted sources: callers should still ensure the input stream originates from a location they trust.

See Also:

• DocumentCategorizerSVM
• Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static final record	SvmDoccatModel.DeserializationLimits	Resource limits applied to the ObjectInputFilter used by deserialize(InputStream, DeserializationLimits).

Method Summary

Modifier and Type	Method	Description
static SvmDoccatModel	deserialize(InputStream in)	Deserializes a SvmDoccatModel from the given InputStream using default resource limits.
static SvmDoccatModel	deserialize(InputStream in, SvmDoccatModel.DeserializationLimits limits)	Deserializes a SvmDoccatModel from the given InputStream using the supplied resource limits.
Map<String,Integer>	getCategoryToIndex()
SvmDoccatConfiguration	getConfiguration()
Map<Integer,Double>	getFeatureMaxValues()
Map<Integer,Double>	getFeatureMinValues()
Map<String,Integer>	getFeatureVocabulary()
Map<String,Double>	getIdfValues()
Map<Integer,String>	getIndexToCategory()
String	getLanguageCode()
int	getNumberOfCategories()
de.hhn.mi.domain.SvmModel	getSvmModel()
void	serialize(OutputStream out)	Serializes this model to the given OutputStream using Java object serialization.

Methods inherited from class java.lang.Object

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

Method Details

getSvmModel

public de.hhn.mi.domain.SvmModel getSvmModel()

Returns:

The underlying SvmModel.

getFeatureVocabulary

public Map<String,Integer> getFeatureVocabulary()

Returns:

An unmodifiable mapping from feature strings to their numeric indices.

getIndexToCategory

public Map<Integer,String> getIndexToCategory()

Returns:

An unmodifiable mapping from numeric category labels to category names.

getCategoryToIndex

public Map<String,Integer> getCategoryToIndex()

Returns:

An unmodifiable mapping from category names to numeric labels.

getIdfValues

public Map<String,Double> getIdfValues()

Returns:

An unmodifiable mapping from feature strings to their IDF values.

getFeatureMinValues

public Map<Integer,Double> getFeatureMinValues()

Returns:

An unmodifiable mapping from feature index to its minimum value in the training corpus.

getFeatureMaxValues

public Map<Integer,Double> getFeatureMaxValues()

Returns:

An unmodifiable mapping from feature index to its maximum value in the training corpus.

getConfiguration

public SvmDoccatConfiguration getConfiguration()

Returns:

The SvmDoccatConfiguration used for this model.

getLanguageCode

public String getLanguageCode()

Returns:

The ISO language code associated with this model.

getNumberOfCategories

public int getNumberOfCategories()

Returns:

The number of categories in this model.

serialize

public void serialize(OutputStream out)
               throws IOException

Serializes this model to the given OutputStream using Java object serialization. The resulting stream can be read back with deserialize(InputStream).

Parameters:

out - The OutputStream to write to. Must not be null.

Throws:

IOException - Thrown if IO errors occurred during serialization.

IllegalArgumentException - if out is null.

deserialize

public static SvmDoccatModel deserialize(InputStream in)
                                  throws IOException,
ClassNotFoundException

Deserializes a SvmDoccatModel from the given InputStream using default resource limits.

The stream is filtered via an ObjectInputFilter that allow-lists only the classes required to reconstruct an SvmDoccatModel, plus resource limits on graph depth, references, and array length. Foreign payloads are rejected with InvalidClassException before ObjectInputStream.readObject() returns.

Callers should still treat this method as defense-in-depth: only invoke it on streams from trusted sources.

If the default limits are too tight for an unusually large model — for example, a model with more than 10000000L entries in a single map — use deserialize(InputStream, DeserializationLimits) to supply higher limits. The class allow-list is intentionally not configurable; loosening it would defeat the purpose of the filter.

Parameters:

in - The InputStream to read from. Must not be null.

Returns:

A valid SvmDoccatModel instance.

Throws:

IOException - Thrown if IO errors occurred during deserialization, including InvalidClassException when the stream contains a class outside the allow-list or exceeds a resource limit.

ClassNotFoundException - Thrown if required classes are not found.

IllegalArgumentException - if in is null.

deserialize

public static SvmDoccatModel deserialize(InputStream in,
 SvmDoccatModel.DeserializationLimits limits)
                                  throws IOException,
ClassNotFoundException

Deserializes a SvmDoccatModel from the given InputStream using the supplied resource limits.

Use this overload when the default limits reject a legitimate model — for example, models trained with very large feature vocabularies. The class allow-list applied to the stream is the same as for deserialize(InputStream); only the numeric limits change.

Parameters:

in - The InputStream to read from. Must not be null.

limits - The SvmDoccatModel.DeserializationLimits to apply. Must not be null.

Returns:

A valid SvmDoccatModel instance.

Throws:

IOException - Thrown if IO errors occurred during deserialization, including InvalidClassException when the stream contains a class outside the allow-list or exceeds one of the supplied limits.

ClassNotFoundException - Thrown if required classes are not found.

IllegalArgumentException - if in or limits is null.