fix opennlp version

additional debug messages
version with git commit

Author: livelace

pipeline_config.groovy

@@ -1,6 +1,7 @@
 def APP_NAME = "digator-opennlp"
 def APP_REPO = "https://github.com/livelace/digator-opennlp.git"
 def APP_VERSION = "${env.VERSION}-${env.OPENNLP_VERSION}"
+def JAR_VERSION = env.VERSION + '-${GIT_COMMIT_SHORT}-' + env.OPENNLP_VERSION
 
 
 libraries {
@@ -29,12 +30,12 @@ libraries {
         options = "--build-arg OPENNLP_VERSION=${env.OPENNLP_VERSION}"
     }
     maven {
-        options = "-Dquarkus.application.version=${APP_VERSION} -Dquarkus.package.type=uber-jar -Dopennlp.version=${env.OPENNLP_VERSION}"
+        options = "-Dquarkus.application.version=${JAR_VERSION} -Dquarkus.package.type=uber-jar -Dopennlp.version=${env.OPENNLP_VERSION}"
     }
     mattermost
     nexus {
         source = "target/digator-opennlp-1.0-SNAPSHOT-runner.jar"
-        destination = "dists-internal/${APP_NAME}/${APP_NAME}-${APP_VERSION}.jar"
+        destination = "dists-internal/${APP_NAME}/${JAR_VERSION}.jar"
     }
     sonarqube
     utils

pom.xml

@@ -19,7 +19,7 @@
     <quarkus.platform.group-id>io.quarkus</quarkus.platform.group-id>
     <quarkus.platform.version>1.10.5.Final</quarkus.platform.version>
     <surefire-plugin.version>2.22.1</surefire-plugin.version>
-    <opennlp.version>1.9.1</opennlp.version>
+    <opennlp.version>1.8.4</opennlp.version>
   </properties>
 
   <dependencyManagement>
@@ -72,7 +72,7 @@
     <dependency>
       <groupId>org.apache.opennlp</groupId>
       <artifactId>opennlp-tools</artifactId>
-      <version>1.9.3</version>
+      <version>${opennlp.version}</version>
     </dependency>
   </dependencies>
 

src/main/java/ru/livelace/digator_opennlp/NerModel.java

@@ -126,12 +126,16 @@ private JsonObject formatToLabelStudio(String text, String[] tokens, Span[] span
      * @param spans
      * @return
      */
-    private static String formatToOpennlp(String[] tokens, Span[] spans) {
+    private String formatToOpennlp(String[] tokens, Span[] spans) {
         for (Span span: spans) {
             tokens[span.getStart()] = String.format("<START:%s> %s", span.getType(), tokens[span.getStart()]);
             tokens[span.getEnd()-1] = String.format("%s <END>", tokens[span.getEnd()-1]);
         }
-        return String.join(" ", tokens);
+
+        var labeledString = String.join(" ", tokens);
+        logger.debug("labeled text: \"{}\"", labeledString);
+
+        return labeledString;
     }
 
     /**
@@ -153,8 +157,11 @@ public JsonObject label(String dataset, String lang, String type, String format,
         // Initialize model.
         if (models.get(modelSignature) == null) {
             try {
-                models.put(modelSignature, new Model(dataset, lang, type));
-                logger.debug("model has been loaded: {}/{}/{}/{}.bin", modelsPath, dataset, lang, type);
+                var model = new Model(dataset, lang, type);
+                models.put(modelSignature, model);
+                logger.debug("model has been loaded: {}/{}/{}/{}.bin, {}, {}",
+                        modelsPath, dataset, lang, type, model.getVersion(), model.getLanguage());
+
             } catch (Exception e) {
                 logger.error("cannot load model: {}", e.getMessage());
                 json.add(ERROR, e.getMessage());
@@ -221,8 +228,16 @@ public Span[] getLabel(String[] tokens) {
             return spans;
         }
 
+        public String getLanguage() {
+            return opennlpModel.getLanguage();
+        }
+
         public String[] getTokens(String text) {
             return opennlpTokenizer.tokenize(text);
         }
+
+        public String getVersion() {
+            return opennlpModel.getVersion().toString();
+        }
     }
 }