added a callback for unknown words

1 files changed, 35 insertions, 0 deletions

diff --git a/src/runtime/java/org/grammaticalframework/pgf/UnknownLiteralCallback.java b/src/runtime/java/org/grammaticalframework/pgf/UnknownLiteralCallback.java
new file mode 100644
index 000000000..1cd5806b6
--- /dev/null
+++ b/src/runtime/java/org/grammaticalframework/pgf/UnknownLiteralCallback.java
@@ -0,0 +1,35 @@
+package org.grammaticalframework.pgf;
+
+import java.util.Collections;
+import java.util.Iterator;
+
+public class UnknownLiteralCallback implements LiteralCallback {
+	private Concr concr;
+
+	public UnknownLiteralCallback(PGF pgf, Concr concr) {
+		this.concr = concr;
+	}
+
+	public CallbackResult match(int lin_idx, String sentence, int offset) {
+		if (!Character.isUpperCase(sentence.charAt(offset))) {
+			int start_offset = offset;
+			while (offset < sentence.length() &&
+			       !Character.isWhitespace(sentence.charAt(offset))) {
+				offset++;
+			}
+			int end_offset = offset;
+			String word = sentence.substring(start_offset,end_offset);
+			
+			if (concr.lookupMorpho(word).size() == 0) {
+				Expr expr = new Expr("MkSymb", new Expr(word));
+				return new CallbackResult(new ExprProb(expr, 0), end_offset);
+			}
+		}
+
+		return null;
+	}
+	
+	public Iterator<TokenProb> predict(int lin_idx, String prefix) {
+		return Collections.<TokenProb>emptyList().iterator();
+	}
+}