Return-Path: 
 <commits-return-3984-apmail-ctakes-commits-archive=ctakes.apache.org@ctakes.apache.org>
X-Original-To: apmail-ctakes-commits-archive@www.apache.org
Delivered-To: apmail-ctakes-commits-archive@www.apache.org
Received: from mail.apache.org (hermes.apache.org [140.211.11.3])
	by minotaur.apache.org (Postfix) with SMTP id 6A23B172A7
	for <apmail-ctakes-commits-archive@www.apache.org>;
 Fri, 13 Mar 2015 16:23:21 +0000 (UTC)
Received: (qmail 64996 invoked by uid 500); 13 Mar 2015 16:23:15 -0000
Delivered-To: apmail-ctakes-commits-archive@ctakes.apache.org
Received: (qmail 64925 invoked by uid 500); 13 Mar 2015 16:23:15 -0000
Mailing-List: contact commits-help@ctakes.apache.org; run by ezmlm
Precedence: bulk
List-Help: <mailto:commits-help@ctakes.apache.org>
List-Unsubscribe: <mailto:commits-unsubscribe@ctakes.apache.org>
List-Post: <mailto:commits@ctakes.apache.org>
List-Id: <commits.ctakes.apache.org>
Reply-To: dev@ctakes.apache.org
Delivered-To: mailing list commits@ctakes.apache.org
Received: (qmail 64909 invoked by uid 99); 13 Mar 2015 16:23:15 -0000
Received: from eris.apache.org (HELO hades.apache.org) (140.211.11.105)
    by apache.org (qpsmtpd/0.29) with ESMTP; Fri, 13 Mar 2015 16:23:15 +0000
Received: from hades.apache.org (localhost [127.0.0.1])
	by hades.apache.org (ASF Mail Server at hades.apache.org) with ESMTP id
 DA045AC0323
	for <commits@ctakes.apache.org>; Fri, 13 Mar 2015 16:23:14 +0000 (UTC)
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Subject: svn commit: r1666501 [2/2] -
 /ctakes/sandbox/ctakes-coref-cleartk/src/main/java/org/apache/ctakes/coreference/ae/
Date: Fri, 13 Mar 2015 16:23:14 -0000
To: commits@ctakes.apache.org
From: tmill@apache.org
X-Mailer: svnmailer-1.0.9
Message-Id: <20150313162314.DA045AC0323@hades.apache.org>

Added: ctakes/sandbox/ctakes-coref-cleartk/src/main/java/org/apache/ctakes/coreference/ae/PersonChainAnnotator.java
URL: http://svn.apache.org/viewvc/ctakes/sandbox/ctakes-coref-cleartk/src/main/java/org/apache/ctakes/coreference/ae/PersonChainAnnotator.java?rev=1666501&view=auto
==============================================================================
--- ctakes/sandbox/ctakes-coref-cleartk/src/main/java/org/apache/ctakes/coreference/ae/PersonChainAnnotator.java (added)
+++ ctakes/sandbox/ctakes-coref-cleartk/src/main/java/org/apache/ctakes/coreference/ae/PersonChainAnnotator.java Fri Mar 13 16:23:14 2015
@@ -0,0 +1,177 @@
+package org.apache.ctakes.coreference.ae;
+
+import java.util.ArrayList;
+import java.util.List;
+
+import org.apache.ctakes.dependency.parser.util.DependencyUtility;
+import org.apache.ctakes.typesystem.type.relation.CollectionTextRelation;
+import org.apache.ctakes.typesystem.type.syntax.ConllDependencyNode;
+import org.apache.ctakes.typesystem.type.syntax.WordToken;
+import org.apache.ctakes.typesystem.type.textsem.Markable;
+import org.apache.uima.analysis_engine.AnalysisEngineDescription;
+import org.apache.uima.analysis_engine.AnalysisEngineProcessException;
+import org.apache.uima.fit.component.JCasAnnotator_ImplBase;
+import org.apache.uima.fit.factory.AnalysisEngineFactory;
+import org.apache.uima.fit.util.JCasUtil;
+import org.apache.uima.jcas.JCas;
+import org.apache.uima.jcas.cas.EmptyFSList;
+import org.apache.uima.jcas.cas.FSList;
+import org.apache.uima.jcas.cas.NonEmptyFSList;
+import org.apache.uima.resource.ResourceInitializationException;
+
+public class PersonChainAnnotator extends JCasAnnotator_ImplBase {
+
+  @Override
+  public void process(JCas jcas) throws AnalysisEngineProcessException {
+    NonEmptyFSList ptList = new NonEmptyFSList(jcas);
+    ptList.setHead(null);
+    NonEmptyFSList weList = new NonEmptyFSList(jcas);
+    weList.setHead(null);
+    NonEmptyFSList drList = new NonEmptyFSList(jcas);
+    drList.setHead(null);
+    List<NonEmptyFSList> otherDrs = new ArrayList<>();
+    
+    List<WordToken> words = new ArrayList<>(JCasUtil.select(jcas, WordToken.class));
+    for(int i = 0; i < words.size(); i++){
+      WordToken word = words.get(i);
+      String text = word.getCoveredText();
+      if(word.getPartOfSpeech().startsWith("PRP")){
+        if(text.equalsIgnoreCase("I") || text.equalsIgnoreCase("me") || text.equalsIgnoreCase("my")){
+          Markable drMention = new Markable(jcas, word.getBegin(), word.getEnd());
+          addToList(jcas, drList, drMention);
+        }else if(text.equalsIgnoreCase("we") || text.equalsIgnoreCase("us") || text.equalsIgnoreCase("our")){
+          Markable weMention = new Markable(jcas, word.getBegin(), word.getEnd());
+          addToList(jcas, weList, weMention);
+        }else if(text.equalsIgnoreCase("it")){
+          // do nothing
+        }else{
+          Markable ptMention = new Markable(jcas, word.getBegin(), word.getEnd());
+          addToList(jcas, ptList, ptMention);
+        }
+      }else if(text.equalsIgnoreCase("dr.")){      
+        Markable drMention = getDoctorMarkable(jcas, word); //new Markable(jcas, word.getBegin(), words.get(i+1).getEnd());
+        addToList(jcas, getCorrectDoctor(jcas, drMention, otherDrs), drMention);
+      }else if(text.equalsIgnoreCase("mrs.") || text.equalsIgnoreCase("mr.") || text.equalsIgnoreCase("ms.")){
+        // TODO - smarter logic for Dr. Firstname Lastname
+        Markable ptMention = new Markable(jcas, word.getBegin(), words.get(i+1).getEnd());
+        addToList(jcas, ptList, ptMention);
+      }else if(text.equalsIgnoreCase("patient")){
+        Markable ptMention = new Markable(jcas, word.getBegin(), word.getEnd());
+        addToList(jcas, ptList, ptMention);
+      }
+    }
+    
+    for(NonEmptyFSList otherDr : otherDrs){
+      if(otherDr.getHead() != null){
+        if(otherDr.getTail() != null){
+          endList(jcas, otherDr);
+          CollectionTextRelation drChain = new CollectionTextRelation(jcas);
+          drChain.setMembers(otherDr);
+          drChain.addToIndexes();
+        }
+      }
+    }
+    
+    if(drList.getHead() != null && drList.getTail() != null){
+      endList(jcas, drList);
+      CollectionTextRelation drChain = new CollectionTextRelation(jcas);
+      drChain.setMembers(drList);
+      drChain.addToIndexes();
+    }
+    if(ptList.getHead() != null && ptList.getTail() != null){
+      endList(jcas, ptList);
+      CollectionTextRelation ptChain = new CollectionTextRelation(jcas);
+      ptChain.setMembers(ptList);
+      ptChain.addToIndexes();
+    }
+    if(weList.getHead() != null && weList.getTail() != null){
+      endList(jcas, weList);
+      CollectionTextRelation weChain = new CollectionTextRelation(jcas);
+      weChain.setMembers(weList);
+      weChain.addToIndexes();
+    }
+  }
+
+  public static AnalysisEngineDescription createAnnotatorDescription() throws ResourceInitializationException {
+    return AnalysisEngineFactory.createEngineDescription(PersonChainAnnotator.class);
+  }
+
+  private static void addToList(JCas jcas, NonEmptyFSList list, Markable arg){
+    arg.addToIndexes();
+    if(list.getHead() == null){
+      // first list element:
+      list.setHead(arg);
+    }else{
+      // subsequent list elements:
+      NonEmptyFSList cur = list;
+      while(cur.getTail() != null){
+        cur = (NonEmptyFSList)cur.getTail();
+      }
+      NonEmptyFSList tail = new NonEmptyFSList(jcas);
+      tail.setHead(arg);
+      cur.setTail(tail);
+      tail.addToIndexes();
+    }
+  }
+  
+  private static void endList(JCas jcas, NonEmptyFSList list){
+    NonEmptyFSList cur = list;
+    while(cur.getTail() != null){
+      cur = (NonEmptyFSList)cur.getTail();
+    }
+    EmptyFSList tail = new EmptyFSList(jcas);
+    cur.setTail(tail);
+    tail.addToIndexes();
+  }
+  
+  private static NonEmptyFSList getCorrectDoctor(JCas jcas, Markable mention, List<NonEmptyFSList> drLists){
+    NonEmptyFSList correctDr = null;
+    if(mention.getCoveredText().length() < 5){
+      if(drLists.size() > 0){
+        correctDr = drLists.get(0);
+      }
+    }else{
+      String nameText = mention.getCoveredText().substring(4);
+      for(NonEmptyFSList drList : drLists){
+        FSList curNode = drList;
+        do{
+          String otherName = ((Markable)((NonEmptyFSList)curNode).getHead()).getCoveredText();
+          if(otherName.length() >= 5){
+            otherName = otherName.substring(4);
+            if(otherName.contains(nameText) || nameText.contains(otherName)){
+              correctDr = drList;
+            }
+          }
+          curNode = ((NonEmptyFSList)curNode).getTail();
+        }while(curNode instanceof NonEmptyFSList);
+        if(correctDr != null) break;
+      }
+    }
+    if(correctDr == null){
+      correctDr = new NonEmptyFSList(jcas);
+      correctDr.setHead(null);
+      drLists.add(correctDr);
+    }
+    return correctDr;
+  }
+  
+  private static Markable getDoctorMarkable(JCas jcas, WordToken drToken){
+    Markable markable = null;
+    
+    ConllDependencyNode nnpHead = DependencyUtility.getDependencyNode(jcas, drToken);
+    try{
+    while(nnpHead != null && nnpHead.getHead() != null && nnpHead.getHead().getId() != 0 && nnpHead.getHead().getPostag().equals("NNP")){
+      nnpHead = nnpHead.getHead();
+    }
+    }catch(NullPointerException e){
+      System.err.print(".");
+    }
+    
+    int start = drToken.getBegin();
+    int end = nnpHead.getEnd();
+    if(end < start) end = drToken.getEnd();
+    
+    markable = new Markable(jcas, start, end);    
+    return markable;
+  }
+}