-Move StoryBookCreateFromEPubController method to a new Util class (C…

…loses elimu-ai#1825)

src/main/java/ai/elimu/util/AppConstants.java

@@ -0,0 +1,12 @@
+package ai.elimu.util;
+
+public class AppConstants {
+
+    public static class READING_LEVEL_CONSTANTS {
+        public static final String CHAPTER_COUNT_KEY = "chapter_count";
+        public static final String PARAGRAPH_COUNT_KEY = "paragraph_count";
+        public static final String WORD_COUNT_KEY = "word_count";
+        public static final String LEVEL = "LEVEL";
+    }
+
+}

src/main/java/ai/elimu/util/ml/ReadingLevelUtil.java

@@ -0,0 +1,41 @@
+package ai.elimu.util.ml;
+
+import ai.elimu.model.v2.enums.ReadingLevel;
+import org.pmml4s.model.Model;
+
+import java.util.Arrays;
+import java.util.Map;
+
+import static ai.elimu.util.AppConstants.READING_LEVEL_CONSTANTS.*;
+
+public class ReadingLevelUtil {
+
+    public static ReadingLevel predictReadingLevel(
+            int chapterCount,
+            int paragraphCount,
+            int wordCount,
+            String modelFilePath
+    ) {
+
+        Model model = Model.fromFile(modelFilePath);
+        Map<String, Double> features = Map.of(
+                CHAPTER_COUNT_KEY, (double) chapterCount,
+                PARAGRAPH_COUNT_KEY, (double) paragraphCount,
+                WORD_COUNT_KEY, (double) wordCount
+        );
+
+        Object[] valuesMap = Arrays.stream(model.inputNames())
+                .map(features::get)
+                .toArray();
+
+        Object[] results = model.predict(valuesMap);
+
+        Object result = results[0];
+        Double resultAsDouble = (Double) result;
+        int resultAsInteger = resultAsDouble.intValue();
+
+        String readingLevelAsString = LEVEL + resultAsInteger;
+        return ReadingLevel.valueOf(readingLevelAsString);
+
+    }
+}

src/main/java/ai/elimu/web/content/storybook/StoryBookCreateFromEPubController.java

@@ -23,6 +23,7 @@
 import ai.elimu.util.epub.EPubImageExtractionHelper;
 import ai.elimu.util.epub.EPubMetadataExtractionHelper;
 import ai.elimu.util.epub.EPubParagraphExtractionHelper;
+import ai.elimu.util.ml.ReadingLevelUtil;
 import ai.elimu.web.context.EnvironmentContextLoaderListener;
 import org.apache.commons.io.FileUtils;
 import org.apache.commons.io.IOUtils;
@@ -52,11 +53,7 @@
 import java.io.InputStream;
 import java.net.URI;
 import java.net.URL;
-import java.util.ArrayList;
-import java.util.Arrays;
-import java.util.Calendar;
-import java.util.List;
-import java.util.Map;
+import java.util.*;
 import java.util.zip.ZipEntry;
 import java.util.zip.ZipInputStream;
 
@@ -542,44 +539,19 @@ private void storeImageContributionEvent(Image image, HttpSession session, HttpS
     }
 
     private ReadingLevel predictReadingLevel(int chapterCount, int paragraphCount, int wordCount) {
-        logger.info("predictReadingLevel");
 
         // Load the machine learning model (https://github.com/elimu-ai/ml-storybook-reading-level)
-        String modelFilePath = getClass().getResource("step2_2_model.pmml").getFile();
-        logger.info("modelFilePath: " + modelFilePath);
-        org.pmml4s.model.Model model = org.pmml4s.model.Model.fromFile(modelFilePath);
-        logger.info("model: " + model);
-
-        // Prepare values (features) to pass to the model
-        Map<String, Double> values = Map.of(
-            "chapter_count", Double.valueOf(chapterCount),
-            "paragraph_count", Double.valueOf(paragraphCount),
-            "word_count", Double.valueOf(wordCount)
+
+        String modelFilePath = Objects.requireNonNull(getClass().getResource("step2_2_model.pmml")).getFile();
+
+        logger.info(
+                "Predicting reading level for chapter: {}, paragraph: {}, word: {}, modelPath: {} ",
+                chapterCount, paragraphCount, wordCount, modelFilePath
         );
-        logger.info("values: " + values);
-
-        // Make prediction
-        logger.info("Arrays.toString(model.inputNames()): " + Arrays.toString(model.inputNames()));
-        Object[] valuesMap = Arrays.stream(model.inputNames())
-                                   .map(values::get)
-                                   .toArray();
-        logger.info("valuesMap: " + valuesMap);
-        Object[] results = model.predict(valuesMap);
-        logger.info("results: " + results);
-        logger.info("Arrays.toString(results): " + Arrays.toString(results));
-        Object result = results[0];
-        logger.info("result: " + result);
-        logger.info("result.getClass().getSimpleName(): " + result.getClass().getSimpleName());
-        Double resultAsDouble = (Double) result;
-        logger.info("resultAsDouble: " + resultAsDouble);
-        Integer resultAsInteger = resultAsDouble.intValue();
-        logger.info("resultAsInteger: " + resultAsInteger);
-
-        // Convert from number to ReadingLevel enum (e.g. "LEVEL2")
-        String readingLevelAsString = "LEVEL" + resultAsInteger;
-        logger.info("readingLevelAsString: " + readingLevelAsString);
-        ReadingLevel readingLevel = ReadingLevel.valueOf(readingLevelAsString);
-        logger.info("readingLevel: " + readingLevel);
+
+        ReadingLevel readingLevel = ReadingLevelUtil.predictReadingLevel(chapterCount, paragraphCount, wordCount, modelFilePath);
+        logger.info("Predicted reading level: {}", readingLevel);
+
         return readingLevel;
     }
 }

src/test/java/ai/elimu/util/ml/ReadingLevelUtilTest.java

@@ -0,0 +1,59 @@
+package ai.elimu.util.ml;
+
+import ai.elimu.model.v2.enums.ReadingLevel;
+import org.junit.jupiter.api.Test;
+
+import java.io.IOException;
+
+import static org.junit.jupiter.api.Assertions.assertEquals;
+import static org.junit.jupiter.api.Assertions.assertThrows;
+
+public class ReadingLevelUtilTest {
+
+    @Test
+    public void testPredictReadingLevel_Level1() {
+
+        String modelFilePath = "src/test/resources/ai/elimu/util/reading_level/model1.pmml";
+        int chapterCount = 5;
+        int paragraphCount = 20;
+        int wordCount = 100;
+
+        ReadingLevel result = ReadingLevelUtil.predictReadingLevel(chapterCount, paragraphCount, wordCount, modelFilePath);
+        assertEquals(ReadingLevel.LEVEL1, result, "Expected ReadingLevel to be LEVEL1, but got: " + result);
+
+    }
+
+    @Test
+    public void testPredictReadingLevel_Level2() {
+
+        String modelFilePath = "src/test/resources/ai/elimu/util/reading_level/model1.pmml";
+        int chapterCount = 12;
+        int paragraphCount = 22;
+        int wordCount = 250;
+
+        ReadingLevel result = ReadingLevelUtil.predictReadingLevel(chapterCount, paragraphCount, wordCount, modelFilePath);
+        assertEquals(ReadingLevel.LEVEL2, result, "Expected ReadingLevel to be LEVEL2, but got: " + result);
+
+    }
+
+    @Test
+    public void testPredictReadingLevel_Level3() {
+
+        String modelFilePath = "src/test/resources/ai/elimu/util/reading_level/model1.pmml";
+        int chapterCount = 12;
+        int paragraphCount = 25;
+        int wordCount = 350;
+
+        ReadingLevel result = ReadingLevelUtil.predictReadingLevel(chapterCount, paragraphCount, wordCount, modelFilePath);
+        assertEquals(ReadingLevel.LEVEL3, result, "Expected ReadingLevel to be LEVEL3, but got: " + result);
+
+    }
+
+    @Test
+    public void testPredictReadingLevel_InvalidModelFile() {
+
+        assertThrows(IOException.class, () -> {
+            ReadingLevelUtil.predictReadingLevel(1, 1, 1, "invalidPath");
+        }, "Expected IOException when loading an invalid model file path");
+    }
+}

src/test/resources/ai/elimu/util/reading_level/model1.pmml

@@ -0,0 +1,41 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<PMML version="4.4" xmlns="http://www.dmg.org/PMML-4_4">
+    <Header>
+        <Application name="Decision Tree Regressor" version="1.0"/>
+    </Header>
+    <DataDictionary numberOfFields="4">
+        <DataField name="chapter_count" optype="continuous" dataType="integer"/>
+        <DataField name="paragraph_count" optype="continuous" dataType="integer"/>
+        <DataField name="word_count" optype="continuous" dataType="integer"/>
+        <DataField name="reading_level" optype="continuous" dataType="double"/>
+    </DataDictionary>
+    <TreeModel functionName="regression" algorithmName="DecisionTree" missingValueStrategy="none">
+        <MiningSchema>
+            <MiningField name="chapter_count"/>
+            <MiningField name="paragraph_count"/>
+            <MiningField name="word_count"/>
+            <MiningField name="reading_level" usageType="target"/>
+        </MiningSchema>
+        <Node score="1.0">
+            <True/>
+            <Node score="1.0">
+                <SimplePredicate field="chapter_count" operator="lessThan" value="10"/>
+                <Node score="1.0">
+                    <SimplePredicate field="paragraph_count" operator="lessOrEqual" value="20"/>
+                </Node>
+                <Node score="2.0">
+                    <SimplePredicate field="paragraph_count" operator="greaterThan" value="20"/>
+                </Node>
+            </Node>
+            <Node score="2.0">
+                <SimplePredicate field="chapter_count" operator="greaterOrEqual" value="10"/>
+                <Node score="2.0">
+                    <SimplePredicate field="word_count" operator="lessThan" value="300"/>
+                </Node>
+                <Node score="3.0">
+                    <SimplePredicate field="word_count" operator="greaterOrEqual" value="300"/>
+                </Node>
+            </Node>
+        </Node>
+    </TreeModel>
+</PMML>