Revision: 15135
http://gate.svn.sourceforge.net/gate/?rev=15135&view=rev
Author: valyt
Date: 2012-01-19 15:52:00 +0000 (Thu, 19 Jan 2012)
Log Message:
-----------
The framework is basically there. Next we'll need some actual rules implemented.
Modified Paths:
--------------
futures/orthoref/src/gate/creole/coref/CorefBase.java
Added Paths:
-----------
futures/orthoref/build.xml
futures/orthoref/creole.xml
futures/orthoref/src/gate/creole/coref/Matcher.java
futures/orthoref/src/gate/creole/coref/OrthoRef.java
futures/orthoref/src/gate/creole/coref/Tagger.java
Added: futures/orthoref/build.xml
===================================================================
--- futures/orthoref/build.xml (rev 0)
+++ futures/orthoref/build.xml 2012-01-19 15:52:00 UTC (rev 15135)
@@ -0,0 +1,97 @@
+<project name="OrthoRef" basedir="." default="build" >
+ <property file="build.properties" />
+
+ <!-- Make environment variables available -->
+ <property environment="env" />
+
+ <!-- If environment variable GATE_HOME is set, use it for
+ gate.home (unless it was already set in build.properties -->
+ <condition property="gate.home" value="${env.GATE_HOME}">
+ <isset property="env.GATE_HOME"/>
+ </condition>
+
+ <!-- If gate.home is not set from build.properties or the
+ environment variable GATE_HOME, assume that this
+ plugin is in GATE Developer's plugin directory -->
+ <property name="gate.home" location="../.." />
+ <property name="src.dir" location="src" />
+ <property name="classes.dir" location="classes" />
+ <property name="lib.dir" location="lib" />
+ <property name="doc.dir" location="doc" />
+ <property name="javadoc.dir" location="${doc.dir}/javadoc" />
+ <property name="jar.location" location="orthoref.jar" />
+ <property name="plugin.name" value="OrthoRef" />
+
+
+ <!-- Path to compile - includes gate.jar and GATE/lib/*.jar -->
+ <path id="compile.classpath">
+ <pathelement location="${lib.dir}/*.jar" />
+ <pathelement location="${gate.home}/bin/gate.jar" />
+ <fileset dir="${gate.home}/lib">
+ <include name="**/*.jar" />
+ <include name="**/*.zip" />
+ </fileset>
+ </path>
+
+ <!-- create build directory structure -->
+ <target name="prepare">
+ <mkdir dir="${classes.dir}" />
+ </target>
+
+ <!-- compile the source -->
+ <target name="compile" depends="prepare">
+ <javac classpathref="compile.classpath"
+ srcdir="${src.dir}"
+ destdir="${classes.dir}"
+ debug="true"
+ debuglevel="lines,source"
+ encoding="UTF-8"
+ source="1.5" target="1.5" />
+ </target>
+
+ <!-- create the JAR file -->
+ <target name="jar" depends="compile" >
+ <jar destfile="${jar.location}"
+ update="false"
+ basedir="${classes.dir}" />
+ </target>
+
+ <!-- remove the generated .class files -->
+ <target name="clean.classes" >
+ <delete dir="${classes.dir}" />
+ </target>
+
+ <!-- Clean up - remove .class and .jar files -->
+ <target name="clean" depends="clean.classes" >
+ <delete file="${jar.location}" />
+ </target>
+
+ <!-- Build JavaDoc documentation -->
+ <target name="doc.prepare">
+ <mkdir dir="${javadoc.dir}" />
+ </target>
+
+ <target name="javadoc" depends="doc.prepare">
+ <javadoc destdir="${javadoc.dir}" packagenames="*"
+ classpathref="compile.classpath"
+ encoding="UTF-8"
+ windowtitle="${plugin.name} JavaDoc"
+ source="1.5">
+ <sourcepath>
+ <pathelement location="${src.dir}" />
+ </sourcepath>
+ <link href="http://java.sun.com/j2se/1.6.0/docs/api/"; />
+ <link href="http://gate.ac.uk/gate/doc/javadoc/"; />
+ </javadoc>
+ </target>
+
+ <!-- Targets used by the main GATE build file:
+ build: build the plugin - just calls "jar" target
+ test : run the unit tests - there aren't any
+ distro.prepare: remove intermediate files that shouldn't be in the
+ distribution
+ -->
+ <target name="build" depends="jar" />
+ <target name="test" />
+ <target name="distro.prepare" depends="clean.classes" />
+</project>
Property changes on: futures/orthoref/build.xml
___________________________________________________________________
Added: svn:eol-style
+ native
Added: futures/orthoref/creole.xml
===================================================================
--- futures/orthoref/creole.xml (rev 0)
+++ futures/orthoref/creole.xml 2012-01-19 15:52:00 UTC (rev 15135)
@@ -0,0 +1,3 @@
+<CREOLE-DIRECTORY>
+ <JAR SCAN="true">orthoref.jar</JAR>
+</CREOLE-DIRECTORY>
Modified: futures/orthoref/src/gate/creole/coref/CorefBase.java
===================================================================
--- futures/orthoref/src/gate/creole/coref/CorefBase.java 2012-01-19
12:21:27 UTC (rev 15134)
+++ futures/orthoref/src/gate/creole/coref/CorefBase.java 2012-01-19
15:52:00 UTC (rev 15135)
@@ -1,7 +1,7 @@
/*
* OrthoRef.java
*
- * Copyright (c) 1995-2010, The University of Sheffield. See the file
+ * Copyright (c) 1995-2012, The University of Sheffield. See the file
* COPYRIGHT.txt in the software or at http://gate.ac.uk/gate/COPYRIGHT.txt
*
* This file is part of GATE (see http://gate.ac.uk/), and is free
@@ -9,6 +9,8 @@
* Version 2, June 1991 (in the distribution as file licence.html,
* and also available at http://gate.ac.uk/gate/licence.html).
*
+ * Valentin Tablan 18 Jan 2012
+ *
* $Id$
*/
package gate.creole.coref;
@@ -16,17 +18,24 @@
import java.util.ArrayList;
import java.util.Collections;
import java.util.HashMap;
+import java.util.HashSet;
+import java.util.LinkedList;
import java.util.List;
import java.util.Map;
import java.util.Set;
import gate.Annotation;
+import gate.Gate;
+import gate.Resource;
import gate.Utils;
import gate.creole.AbstractLanguageAnalyser;
import gate.creole.ExecutionException;
+import gate.creole.ResourceInstantiationException;
+import gate.creole.metadata.CreoleParameter;
import gate.creole.metadata.CreoleResource;
+import gate.creole.metadata.RunTime;
-@CreoleResource (comment = "An orthographic coreferencer")
+@CreoleResource
/**
* Base class for coreferencers. This class has no internal state, so it safe
to
* use in a multi-threaded environment. The same may not be true about
@@ -35,17 +44,171 @@
public abstract class CorefBase extends AbstractLanguageAnalyser {
/**
+ *
+ */
+ private static final long serialVersionUID = -4069411446200124751L;
+
+ /**
* Which annotation types are seen as anaphors
*/
private Set<String> anaphorAnnotationTypes;
private String annotationSetName;
- public abstract Set<String> anaphorTags(Annotation anaphor);
+ private List<Tagger> taggers;
- public abstract boolean matches(Annotation anaphor, Annotation antecedent);
+ private List<Matcher> matchers;
+
+ private List<String> packageNames;
+ private List<String> taggerClassNames;
+
+ private List<String> matcherClassNames;
+
+
+ public List<String> getPackageNames() {
+ return packageNames;
+ }
+
+ @CreoleParameter (defaultValue = "gate.creole.coref",
+ comment="<html>Java <b>packages</b> that should be searched for Tagger "
+
+ "and Matcher implementations.</html>")
+ public void setPackageNames(List<String> packageNames) {
+ this.packageNames = packageNames;
+ }
+
+ public List<String> getTaggerClassNames() {
+ return taggerClassNames;
+ }
+
+ @CreoleParameter (comment =
+ "<html>List of <b>class names</b< of Tagger implementations that " +
+ "should be instantiated and used.<br>" +
+ "Non fully-qualified names will be searched for in the provided " +
+ "package names.</html>")
+ public void setTaggerClassNames(List<String> taggerClassNames) {
+ this.taggerClassNames = taggerClassNames;
+ }
+
+ public List<String> getMatcherClassNames() {
+ return matcherClassNames;
+ }
+
+ @CreoleParameter (comment =
+ "<html>List of <b>class names</b> of Matcher implementations that " +
+ "should be instantiated and used.<br>" +
+ "Non fully-qualified names will be searched for in the provided " +
+ "package names.</html>")
+ public void setMatcherClassNames(List<String> matcherClassNames) {
+ this.matcherClassNames = matcherClassNames;
+ }
+
+ public Set<String> getAnaphorAnnotationTypes() {
+ return anaphorAnnotationTypes;
+ }
+
+ @CreoleParameter(comment = "Annotation types that should be considered",
+ defaultValue = "Person;Location;Organization")
+ @RunTime
+ public void setAnaphorAnnotationTypes(Set<String> anaphorAnnotationTypes) {
+ this.anaphorAnnotationTypes = anaphorAnnotationTypes;
+ }
+
+ public String getAnnotationSetName() {
+ return annotationSetName;
+ }
+
+ @CreoleParameter(comment="The name of the annotation set to be processed",
+ defaultValue="")
+ @RunTime
+ public void setAnnotationSetName(String annotationSetName) {
+ this.annotationSetName = annotationSetName;
+ }
+
@Override
+ public Resource init() throws ResourceInstantiationException {
+ if(packageNames == null) packageNames = Collections.EMPTY_LIST;
+ if(taggerClassNames == null) throw new ResourceInstantiationException(
+ "No tagger class names provided!");
+ if(matcherClassNames == null) throw new ResourceInstantiationException(
+ "No matcher class names provided!");
+ // create the list of taggers
+ taggers = new LinkedList<Tagger>();
+ for(String taggerClassName : taggerClassNames) {
+ Class<Tagger> taggerClass = null;
+ if(taggerClassName.indexOf('.') >= 0) {
+ // fully qualified name
+ try {
+ taggerClass = (Class<Tagger>)Class.forName(taggerClassName, true,
+ Gate.getClassLoader());
+ } catch(ClassNotFoundException e) {
+ throw new ResourceInstantiationException(e);
+ }
+ } else {
+ // search packages
+ for(String packageName : packageNames) {
+ try {
+ taggerClass = (Class<Tagger>)Class.forName(
+ packageName + "." + taggerClassName, true,
Gate.getClassLoader());
+ break;
+ } catch(ClassNotFoundException e) {
+ // ignore
+ }
+ }
+ }
+ if(taggerClass == null) {
+ throw new ResourceInstantiationException(
+ "Could not find class for tagger class name \"" + taggerClassName +
+ "\".");
+ }
+ // create an instance
+ try {
+ taggers.add(taggerClass.newInstance());
+ } catch(Exception e) {
+ throw new ResourceInstantiationException(e);
+ }
+ }
+
+ // create the list of matchers
+ matchers = new LinkedList<Matcher>();
+ for(String matcherClassName : matcherClassNames) {
+ Class<Matcher> matcherClass = null;
+ if(matcherClassName.indexOf('.') >= 0) {
+ // fully qualified name
+ try {
+ matcherClass = (Class<Matcher>)Class.forName(matcherClassName, true,
+ Gate.getClassLoader());
+ } catch(ClassNotFoundException e) {
+ throw new ResourceInstantiationException(e);
+ }
+ } else {
+ // search packages
+ for(String packageName : packageNames) {
+ try {
+ matcherClass = (Class<Matcher>)Class.forName(
+ packageName + "." + matcherClassName, true,
Gate.getClassLoader());
+ break;
+ } catch(ClassNotFoundException e) {
+ // ignore
+ }
+ }
+ }
+ if(matcherClass == null) {
+ throw new ResourceInstantiationException(
+ "Could not find class for matcher class name \"" + matcherClassName
+
+ "\".");
+ }
+ // create an instance
+ try {
+ matchers.add(matcherClass.newInstance());
+ } catch(Exception e) {
+ throw new ResourceInstantiationException(e);
+ }
+ }
+ return super.init();
+ }
+
+ @Override
public void execute() throws ExecutionException {
fireProgressChanged(1);
//collect all candidates
@@ -55,27 +218,47 @@
document.getAnnotations(annotationSetName)).get(anaphorAnnotationTypes));
Collections.sort(anaphors, Utils.OFFSET_COMPARATOR);
- Map<String, Annotation> tag2candidate = new HashMap<String, Annotation>();
- Map<Annotation, String> anaphor2chain = new HashMap<Annotation, String>();
+ Map<String, Integer> tag2antecedent = new HashMap<String, Integer>();
+ Map<Integer, String> anaphor2chain = new HashMap<Integer, String>();
+ Map<String, List<Annotation>> chain2anaphors =
+ new HashMap<String, List<Annotation>>();
// for each candidate, check the hash map
- for(Annotation anaphor : anaphors) {
- Set<String> newTags = anaphorTags(anaphor);
+ for(int anaphorIdx = 0; anaphorIdx < anaphors.size(); anaphorIdx++) {
+ Set<String> newTags = new HashSet<String>();
+ for(Tagger tagger : taggers){
+ newTags.addAll(tagger.tags(anaphors, anaphorIdx));
+ }
for(String tag : newTags) {
- Annotation candidate = tag2candidate.get(tag);
- if(candidate != null && matches(anaphor, candidate)) {
- // we have a match
- String chainLabel = anaphor2chain.get(candidate);
- if(chainLabel == null) {
- // first match for this chain
- chainLabel = Utils.contentFor(document, candidate).toString();
- anaphor2chain.put(candidate, chainLabel);
+ Integer candidateIdx = tag2antecedent.get(tag);
+ if(candidateIdx != null) {
+ for(Matcher matcher : matchers) {
+ if(matcher.matches(anaphors, candidateIdx, anaphorIdx)) {
+ // we have a match
+ String chainLabel = anaphor2chain.get(candidateIdx);
+ if(chainLabel == null) {
+ // first match for this chain
+ chainLabel = Utils.contentFor(document,
+ anaphors.get(candidateIdx)).toString();
+ anaphor2chain.put(candidateIdx, chainLabel);
+ }
+ anaphor2chain.put(anaphorIdx, chainLabel);
+ List<Annotation> chain = chain2anaphors.get(chainLabel);
+ if(chain == null) {
+ chain = new LinkedList<Annotation>();
+ chain2anaphors.put(chainLabel, chain);
+ }
+ chain.add(anaphors.get(anaphorIdx));
+ // no need to look for more
+ break;
+ }
}
- anaphor2chain.put(anaphor, chainLabel);
}
}
}
+ // create the annotation features
+ //TODO
Added: futures/orthoref/src/gate/creole/coref/Matcher.java
===================================================================
--- futures/orthoref/src/gate/creole/coref/Matcher.java
(rev 0)
+++ futures/orthoref/src/gate/creole/coref/Matcher.java 2012-01-19 15:52:00 UTC
(rev 15135)
@@ -0,0 +1,26 @@
+/*
+ * AnaphorMatcher.java
+ *
+ * Copyright (c) 1995-2012, The University of Sheffield. See the file
+ * COPYRIGHT.txt in the software or at http://gate.ac.uk/gate/COPYRIGHT.txt
+ *
+ * This file is part of GATE (see http://gate.ac.uk/), and is free
+ * software, licenced under the GNU Library General Public License,
+ * Version 2, June 1991 (in the distribution as file licence.html,
+ * and also available at http://gate.ac.uk/gate/licence.html).
+ *
+ * Valentin Tablan 19 Jan 2012
+ *
+ * $Id$
+ */
+package gate.creole.coref;
+
+import gate.Annotation;
+
+import java.util.List;
+
+public interface Matcher {
+
+ public boolean matches(List<Annotation> anaphors, int antecedent,
+ int anaphor);
+}
Property changes on: futures/orthoref/src/gate/creole/coref/Matcher.java
___________________________________________________________________
Added: svn:keywords
+ Id
Added: svn:eol-style
+ native
Added: futures/orthoref/src/gate/creole/coref/OrthoRef.java
===================================================================
--- futures/orthoref/src/gate/creole/coref/OrthoRef.java
(rev 0)
+++ futures/orthoref/src/gate/creole/coref/OrthoRef.java 2012-01-19
15:52:00 UTC (rev 15135)
@@ -0,0 +1,28 @@
+package gate.creole.coref;
+
+import gate.creole.metadata.CreoleParameter;
+import gate.creole.metadata.CreoleResource;
+
+import java.util.List;
+
+@CreoleResource (name="OrthoRef", comment = "An orthographic coreferencer")
+public class OrthoRef extends CorefBase {
+
+ /**
+ *
+ */
+ private static final long serialVersionUID = -1723494592444418868L;
+
+ @Override
+ @CreoleParameter(defaultValue="tagger1;tagger2")
+ public void setTaggerClassNames(List<String> taggerClassNames) {
+ super.setTaggerClassNames(taggerClassNames);
+ }
+
+ @Override
+ @CreoleParameter(defaultValue="matcher1;matcher2")
+ public void setMatcherClassNames(List<String> matcherClassNames) {
+ super.setMatcherClassNames(matcherClassNames);
+ }
+
+}
Property changes on: futures/orthoref/src/gate/creole/coref/OrthoRef.java
___________________________________________________________________
Added: svn:keywords
+ Id
Added: svn:eol-style
+ native
Added: futures/orthoref/src/gate/creole/coref/Tagger.java
===================================================================
--- futures/orthoref/src/gate/creole/coref/Tagger.java
(rev 0)
+++ futures/orthoref/src/gate/creole/coref/Tagger.java 2012-01-19 15:52:00 UTC
(rev 15135)
@@ -0,0 +1,27 @@
+/*
+ * AnaphorTagger.java
+ *
+ * Copyright (c) 1995-2012, The University of Sheffield. See the file
+ * COPYRIGHT.txt in the software or at http://gate.ac.uk/gate/COPYRIGHT.txt
+ *
+ * This file is part of GATE (see http://gate.ac.uk/), and is free
+ * software, licenced under the GNU Library General Public License,
+ * Version 2, June 1991 (in the distribution as file licence.html,
+ * and also available at http://gate.ac.uk/gate/licence.html).
+ *
+ * Valentin Tablan 19 Jan 2012
+ *
+ * $Id$
+ */
+package gate.creole.coref;
+
+import gate.Annotation;
+
+import java.util.List;
+import java.util.Set;
+
+public interface Tagger {
+
+ public Set<String> tags(List<Annotation> anaphors, int anaphor);
+
+}
Property changes on: futures/orthoref/src/gate/creole/coref/Tagger.java
___________________________________________________________________
Added: svn:keywords
+ Id
Added: svn:eol-style
+ native
This was sent by the SourceForge.net collaborative development platform, the
world's largest Open Source development site.
------------------------------------------------------------------------------
Keep Your Developer Skills Current with LearnDevNow!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-d2d
_______________________________________________
GATE-cvs mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/gate-cvs
