refactored sentiment trainer to be able to use files from tests

Author: ikwattro

.gitignore

@@ -1,3 +1,4 @@
 target/
 *iml
 dependency-reduced-pom.xml
+**/.DS_Store

pom.xml

@@ -96,36 +96,42 @@
       <artifactId>slf4j-simple</artifactId>
       <version>1.7.21</version>
     </dependency>
+
     <dependency>
       <groupId>junit</groupId>
       <artifactId>junit</artifactId>
       <version>4.12</version>
       <scope>test</scope>
     </dependency>
+
     <dependency>
       <groupId>com.graphaware.neo4j</groupId>
       <artifactId>runtime</artifactId>
       <version>${graphaware.version}</version>
       <scope>test</scope>
     </dependency>
+
     <dependency>
       <groupId>com.graphaware.neo4j</groupId>
       <artifactId>server</artifactId>
       <version>${graphaware.version}</version>
       <scope>test</scope>
     </dependency>
+
     <dependency>
       <groupId>com.graphaware.neo4j</groupId>
       <artifactId>tests</artifactId>
       <version>${graphaware.version}</version>
       <scope>test</scope>
     </dependency>
+
     <dependency>
       <groupId>com.graphaware.neo4j</groupId>
       <artifactId>resttest</artifactId>
       <scope>test</scope>
       <version>${resttest.version}</version>
     </dependency>
+
     <dependency>
       <groupId>com.graphaware.neo4j</groupId>
       <artifactId>nlp</artifactId>
@@ -134,6 +140,13 @@
       <scope>test</scope>
     </dependency>
 
+    <dependency>
+      <groupId>com.sun.jersey</groupId>
+      <artifactId>jersey-server</artifactId>
+      <version>1.19.1</version>
+      <scope>test</scope>
+    </dependency>
+
   </dependencies>
 
 

src/main/java/com/graphaware/nlp/processor/opennlp/OpenNLPPipeline.java

@@ -364,10 +364,22 @@ else if (alg.toLowerCase().equals("sentiment")) {
             SentimentModelTool sentModel = new SentimentModelTool(fileTrain, modelId, lang, params);
             sentModel.train();
             result = sentModel.validate();
-            sentModel.saveModel(fileOut);
+            String[] dirPathSplit = fileTrain.split(File.separator);
+            String fileOutToUse = null;
+            if (dirPathSplit.length > 2) {
+                StringBuilder sb = new StringBuilder("");
+                for (int i = 0; i < dirPathSplit.length -2; ++i) {
+                    sb.append(dirPathSplit[i]).append(File.separator);
+                }
+                fileOutToUse = sb.toString() + fileOut;
+            } else {
+                fileOutToUse = fileOut;
+            }
+            System.out.println("Saving model to " + fileOutToUse);
+            sentModel.saveModel(fileOutToUse);
             // incorporate this model to the OpenNLPPipeline
             if (sentModel.getModel() != null) {
-                customSentimentModels.put(newKey, fileOut);
+                customSentimentModels.put(newKey, fileOutToUse);
                 //sentimentDetectors.put(newKey, new DocumentCategorizerME((DoccatModel) sentModel.getModel()));
             }
         } else {

src/main/java/com/graphaware/nlp/processor/opennlp/model/OpenNLPGenericModelTool.java

@@ -211,7 +211,7 @@ class ImprovisedInputStreamFactory implements InputStreamFactory {
             }
             try {
                 if (this.inputSourceStr.startsWith("file://")) {
-                    this.inputSourceFile = new File(new URI(this.inputSourceStr));
+                    this.inputSourceFile = new File(new URI(this.inputSourceStr.replace("file://", "")));
                 } else if (this.inputSourceStr.startsWith("/")) {
                     this.inputSourceFile = new File(this.inputSourceStr);
                 }

src/test/java/com/graphaware/nlp/processor/opennlp/model/CustomSentimentModelIntegrationTest.java

@@ -0,0 +1,19 @@
+package com.graphaware.nlp.processor.opennlp.model;
+
+import com.graphaware.nlp.NLPIntegrationTest;
+import org.junit.Test;
+
+import static org.junit.Assert.*;
+
+public class CustomSentimentModelIntegrationTest extends NLPIntegrationTest {
+
+    @Test
+    public void testTrainCustomModelWithProcedure() {
+        String p = getClass().getClassLoader().getResource("import/sentiment_tweets.train").getPath();
+        String q = "CALL ga.nlp.processor.train({textProcessor: \"com.graphaware.nlp.processor.opennlp.OpenNLPTextProcessor\", modelIdentifier: \"component-en\", alg: \"sentiment\", inputFile: \""+p+"\" , lang: \"en\"})";
+        executeInTransaction(q, (result -> {
+            assertTrue(result.hasNext());
+        }));
+    }
+
+}

src/test/resources/import/sentiment_tweets.train

@@ -0,0 +1,100 @@
+3	Watching a nice movie
+1	The painting is ugly, will return it tomorrow...
+3	One of the best soccer games, worth seeing it
+3	Very tasty, not only for vegetarians
+3	Super party!
+1	Too early to travel..need a coffee
+1	Damn..the train is late again...
+1	Bad news, my flight just got cancelled.
+3	Happy birthday mr. president
+3	Just watch it. Respect.
+3	Wonderful sunset.
+3	Bravo, first title in 2014!
+1	Had a bad evening, need urgently a beer.
+1	I put on weight again
+3	On today's show we met Angela, a woman with an amazing story
+3	I fell in love again
+1	I lost my keys
+3	On a trip to Iceland
+3	Happy in Berlin
+1	I hate Mondays
+3	Love the new book I reveived for Christmas
+1	He killed our good mood
+3	I am in good spirits again
+3	This guy creates the most awesome pics ever 
+1	The dark side of a selfie.
+3	Cool! John is back!
+3	Many rooms and many hopes for new residents
+1	False hopes for the people attending the meeting
+3	I set my new year's resolution
+1	The ugliest car ever!
+1	Feeling bored
+1	Need urgently a pause
+3	Nice to see Ana made it
+3	My dream came true
+1	I didn't see that one coming
+1	Sorry mate, there is no more room for you
+1	Who could have possibly done this?
+3	I won the challenge
+1	I feel bad for what I did		
+3	I had a great time tonight
+3	It was a lot of fun
+3	Thank you Molly making this possible
+1	I just did a big mistake
+3	I love it!!
+1	I never loved so hard in my life
+1	I hate you Mike!!
+1	I hate to say goodbye
+3	Lovely!
+3	Like and share if you feel the same
+1	Never try this at home
+1	Don't spoil it!
+3	I love rock and roll
+1	The more I hear you, the more annoyed I get
+3	Finnaly passed my exam!
+3	Lovely kittens
+1	I just lost my appetite
+1	Sad end for this movie
+1	Lonely, I am so lonely
+3	Beautiful morning
+3	She is amazing
+3	Enjoying some time with my friends
+3	Special thanks to Marty
+3	Thanks God I left on time
+3	Greateful for a wonderful meal
+3	So happy to be home
+1	Hate to wait on a long queue		
+1	No cab available
+1	Electricity outage, this is a nightmare
+1	Nobody to ask about directions
+3	Great game!
+3	Nice trip
+3	I just received a pretty flower
+3	Excellent idea
+3	Got a new watch. Feeling happy
+1	I feel sick
+1	I am very tired
+3	Such a good taste 
+1	Such a bad taste
+3	Enjoying brunch
+1	I don't recommend this restaurant
+3	Thank you mom for supporting me
+1	I will never ever call you again
+1	I just got kicked out of the contest
+3	Smiling
+1	Big pain to see my team loosing
+1	Bitter defeat tonight
+1	My bike was stollen
+3	Great to see you!
+1	I lost every hope for seeing him again
+3	Nice dress!
+3	Stop wasting my time
+3	I have a great idea
+3	Excited to go to the pub
+3	Feeling proud
+3	Cute bunnies
+1	Cold winter ahead
+1	Hopless struggle..
+1	Ugly hat
+3	Big hug and lots of love
+3	I hope you have a wonderful celebration