de.julielab.jcore.ae.jnet.uima

Class ConsistencyPreservation

java.lang.Object

de.julielab.jcore.ae.jnet.uima.ConsistencyPreservation

public class ConsistencyPreservation
extends Object

Field Summary

Fields

Modifier and Type	Field and Description
static String	MODE_ACRO2FULL
static String	MODE_FULL2ACRO
static String	MODE_STRING String matches will be expanded to token boundaries
static String	MODE_STRING_TOKEN_BOUNDARIES If set, only create new annotations if the matched string begins and ends exactly with token borders.

Constructor Summary

Constructors

Constructor and Description
ConsistencyPreservation(String modesString) builds the modes used during consistency preservation from a string which is a coma-separated list of modes.

Method Summary

Modifier and Type	Method and Description
void	acroMatch(org.apache.uima.jcas.JCas aJCas, Set<String> entityMentionClassnames)
void	stringMatch(org.apache.uima.jcas.JCas aJCas, TreeSet<String> entityMentionClassnames, double confidenceThresholdForConsistencyPreservation) consistency presevation based on (exact) string matching.
String	toString()

Methods inherited from class java.lang.Object

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

Field Detail

MODE_ACRO2FULL

public static final String MODE_ACRO2FULL

See Also:

Constant Field Values

MODE_FULL2ACRO

public static final String MODE_FULL2ACRO

See Also:

Constant Field Values

MODE_STRING

public static final String MODE_STRING

String matches will be expanded to token boundaries

See Also:

Constant Field Values

MODE_STRING_TOKEN_BOUNDARIES

public static final String MODE_STRING_TOKEN_BOUNDARIES

If set, only create new annotations if the matched string begins and ends exactly with token borders. This avoids partial token matches which are then expanded to the whole token. Should be used for full texts.

See Also:

Constant Field Values

Constructor Detail

ConsistencyPreservation

public ConsistencyPreservation(String modesString)
                        throws org.apache.uima.resource.ResourceInitializationException

builds the modes used during consistency preservation from a string which is a coma-separated list of modes.

Parameters:

tring - coma-separated list of modes to be used

Throws:

org.apache.uima.analysis_engine.AnalysisEngineProcessException

org.apache.uima.resource.ResourceInitializationException

Method Detail

acroMatch

public void acroMatch(org.apache.uima.jcas.JCas aJCas,
                      Set<String> entityMentionClassnames)
               throws org.apache.uima.analysis_engine.AnalysisEngineProcessException

Throws:

org.apache.uima.analysis_engine.AnalysisEngineProcessException

stringMatch

public void stringMatch(org.apache.uima.jcas.JCas aJCas,
                        TreeSet<String> entityMentionClassnames,
                        double confidenceThresholdForConsistencyPreservation)
                 throws org.apache.uima.analysis_engine.AnalysisEngineProcessException

consistency presevation based on (exact) string matching. If string was annotated once as entity, all other occurrences of this string get the same label. For mode: _string_ TODO: more intelligent (voting) mechanism needed to avoid false positives

Parameters:

aJCas -

entityMentionClassnames -

confidenceThresholdForConsistencyPreservation -

Throws:

org.apache.uima.analysis_engine.AnalysisEngineProcessException

toString

public String toString()

Overrides:

toString in class Object