[incubator-nlpcraft] branch master_test updated: WIP

[aradzinski]

This is an automated email from the ASF dual-hosted git repository.

aradzinski pushed a commit to branch master_test
in repository https://gitbox.apache.org/repos/asf/incubator-nlpcraft.git


The following commit(s) were added to refs/heads/master_test by this push:
     new 3694de2  WIP
3694de2 is described below

commit 3694de29fcb1369f2236b0ed6f7e3d929ffcbf2f
Author: Aaron Radzinski <aradzin...@datalingvo.com>
AuthorDate: Thu Dec 16 19:47:44 2021 -0800

    WIP
---
 .../token/parser/opennlp/NCOpenNlpTokenParser.java | 28 ++++++++++------------
 .../token/parser/opennlp/impl/NCOpenNlpImpl.scala  | 14 +++++------
 2 files changed, 20 insertions(+), 22 deletions(-)

diff --git 
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/parser/opennlp/NCOpenNlpTokenParser.java
 
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/parser/opennlp/NCOpenNlpTokenParser.java
index 8f3e155..91378e7 100644
--- 
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/parser/opennlp/NCOpenNlpTokenParser.java
+++ 
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/parser/opennlp/NCOpenNlpTokenParser.java
@@ -17,10 +17,8 @@
 
 package org.apache.nlpcraft.internal.nlp.token.parser.opennlp;
 
-import org.apache.nlpcraft.NCRequest;
-import org.apache.nlpcraft.NCToken;
-import org.apache.nlpcraft.NCTokenParser;
-
+import org.apache.nlpcraft.*;
+import org.apache.nlpcraft.internal.nlp.token.parser.opennlp.impl.*;
 import java.io.File;
 import java.io.InputStream;
 import java.util.List;
@@ -37,26 +35,26 @@ import java.util.List;
  *
  */
 public class NCOpenNlpTokenParser implements NCTokenParser {
-    private final NCOpenNlpTokenParserImpl impl;
+    private final NCOpenNlpImpl impl;
 
     /**
      *
-     * @param tokenizer
-     * @param tagger
-     * @param lemmatizer
+     * @param tokModel
+     * @param posModel
+     * @param lemmaDic
      */
-    public NCOpenNlpTokenParser(File tokenizer, File tagger, File lemmatizer) {
-        this.delegate = NCOpenNlpTokenParserImpl.apply(tokenizer, tagger, 
lemmatizer);
+    public NCOpenNlpTokenParser(File tokModel, File posModel, File lemmaDic) {
+        // TODO
     }
 
     /**
      *
-     * @param tokenizer
+     * @param tokMdlPath
      * @param tagger
      * @param lemmatizer
      */
-    public NCOpenNlpTokenParser(String tokenizer, String tagger, String 
lemmatizer) {
-        this.delegate = NCOpenNlpTokenParserImpl.apply(tokenizer, tagger, 
lemmatizer);
+    public NCOpenNlpTokenParser(String tokMdlPath, String tagger, String 
lemmatizer) {
+       // TODO
     }
 
     /**
@@ -66,12 +64,12 @@ public class NCOpenNlpTokenParser implements NCTokenParser {
      * @param lemmatizer
      */
     public NCOpenNlpTokenParser(InputStream tokenizer, InputStream tagger, 
InputStream lemmatizer) {
-        this.delegate = NCOpenNlpTokenParserImpl.apply(tokenizer, tagger, 
lemmatizer);
+        impl = NCOpenNlpImpl.apply(tokenizer, tagger, lemmatizer);
     }
 
     @Override
     public List<NCToken> parse(NCRequest req) {
-        return delegate.parse(req);
+        return impl.parse(req);
     }
 
     /**
diff --git 
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/parser/opennlp/impl/NCOpenNlpImpl.scala
 
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/parser/opennlp/impl/NCOpenNlpImpl.scala
index bd4c272..6bc721f 100644
--- 
a/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/parser/opennlp/impl/NCOpenNlpImpl.scala
+++ 
b/nlpcraft/src/main/scala/org/apache/nlpcraft/internal/nlp/token/parser/opennlp/impl/NCOpenNlpImpl.scala
@@ -29,14 +29,14 @@ import scala.jdk.CollectionConverters.*
 
 /**
   *
-  * @param tokIn
-  * @param tagIn
-  * @param lemmaIn
+  * @param tokModelIn
+  * @param posModelIn
+  * @param lemmaDicIn
   */
-class NCOpenNlpImpl(tokIn: InputStream, tagIn: InputStream, lemmaIn: 
InputStream):
-    private val tokenizer = new TokenizerME(new TokenizerModel(tokIn))
-    private val tagger = new POSTaggerME(new POSModel(tagIn))
-    private val lemmatizer = new DictionaryLemmatizer(lemmaIn)
+class NCOpenNlpImpl(tokModelIn: InputStream, posModelIn: InputStream, 
lemmaDicIn: InputStream):
+    private val tokenizer = new TokenizerME(new TokenizerModel(tokModelIn))
+    private val tagger = new POSTaggerME(new POSModel(posModelIn))
+    private val lemmatizer = new DictionaryLemmatizer(lemmaDicIn)
     private val stemmer = new PorterStemmer
 
     /**