X-Git-Url: https://git.wpitchoune.net/gitweb/?p=pnews.git;a=blobdiff_plain;f=war%2Fsrc%2Fmain%2Fjava%2Fpnews%2FNER.java;h=ac34c0812faf0f24f71392b497ac6fc86ea5d608;hp=f8238c1d7ce51869727304c65b9336fb5b08ca13;hb=c6f5722d9572158f0d561819f080f8ccf4e3c7d9;hpb=56c07f5de3319eb61182b7100855801644538e6f diff --git a/war/src/main/java/pnews/NER.java b/war/src/main/java/pnews/NER.java index f8238c1..ac34c08 100644 --- a/war/src/main/java/pnews/NER.java +++ b/war/src/main/java/pnews/NER.java @@ -14,19 +14,19 @@ public class NER { private static final String CLASS_NAME = NER.class.getName(); private static final Logger LOG = Logger.getLogger(CLASS_NAME); - public static String[] classify(String str) throws ClassCastException, ClassNotFoundException, IOException { + public static List classify(String str, List entities) throws ClassCastException, ClassNotFoundException, IOException { CRFClassifier classifier; List> out; String cat, w; - List entities; final String FUNCTION_NAME = "classify"; LOG.entering(CLASS_NAME, FUNCTION_NAME, str); + + OpenNLP.classify(str, entities); classifier = CRFClassifier.getDefaultClassifier(); out = classifier.classify(str); - entities = new ArrayList<>(); for (List labels: out) for (CoreLabel l: labels) { cat = l.getString(AnswerAnnotation.class); @@ -35,12 +35,17 @@ public class NER { entities.add(w); } + entities.remove("CNET"); + entities.remove("Read More"); + entities.remove("New"); + entities.remove("App"); + LOG.exiting(CLASS_NAME, FUNCTION_NAME, entities); - return entities.toArray(new String[0]); + return entities; } public static void main(String[] args) throws Exception { - classify("I live in Washington."); + classify("I live in Washington.", new ArrayList<>()); } } \ No newline at end of file