commit embeddings data, not binary jar

build.sh

@@ -0,0 +1,13 @@
+#! /bin/bash
+
+## download embeddings.txt
+# dvc pull embeddings.txt.dvc
+## or copy it of a previous build
+# unzip -j binary/graphseg.jar embeddings.txt
+
+cp embeddings.txt source/res/
+
+cd source
+mvn package
+mv target/graphseg-0.0.1-SNAPSHOT-jar-with-dependencies.jar ../binary/graphseg.jar
+

binary/graphseg.jar → embeddings.txt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8933d43d10f8ba885d4df38b3b0a2d1bf7796f7d367b7f9fef6f3925801987c5
-size 616819715
+oid sha256:18870b0a7516e4a72b44d3c226c242d2d846008967d8ce40b94c723a94d1a32b
+size 693432828

source/src/edu/uma/nlp/graphseg/preprocessing/StanfordAnnotator.java

@@ -37,7 +37,8 @@ public class StanfordAnnotator implements IAnnotator {
 		{
 			Properties props = new Properties();
 		    props.setProperty("annotators", stanfordAnnotatorsString);
-		    
+		    props.setProperty("ssplit.eolonly", "true");
+
 		    StanfordCoreNLP pipeline = new StanfordCoreNLP(props);
 		    
 		    edu.stanford.nlp.pipeline.Annotation docAnnotation = new edu.stanford.nlp.pipeline.Annotation(((Document)textUnit).getText());