Add File

2025-09-04 14:09:28 +08:00
parent 064d1cc95a
commit fecfbae90d
1 changed files with 120 additions and 0 deletions
--- a/src/main/java/org/dromara/easyai/naturalLanguage/word/MyKeyWord.java
+++ b/src/main/java/org/dromara/easyai/naturalLanguage/word/MyKeyWord.java
@@ -0,0 +1,120 @@
+package org.dromara.easyai.naturalLanguage.word;
+
+import org.dromara.easyai.matrixTools.Matrix;
+import org.dromara.easyai.matrixTools.MatrixOperation;
+import org.dromara.easyai.config.RZ;
+import org.dromara.easyai.config.SentenceConfig;
+import org.dromara.easyai.entity.KeyWordForSentence;
+import org.dromara.easyai.entity.WordBack;
+import org.dromara.easyai.function.Tanh;
+import org.dromara.easyai.i.OutBack;
+import org.dromara.easyai.rnnNerveCenter.ModelParameter;
+import org.dromara.easyai.rnnNerveCenter.NerveManager;
+import org.dromara.easyai.rnnNerveEntity.SensoryNerve;
+
+import java.util.HashMap;
+import java.util.List;
+import java.util.Map;
+
+public class MyKeyWord extends MatrixOperation {
+    private final WordEmbedding wordEmbedding;
+    private final NerveManager typeNerveManager;
+    private final int nerveLength;
+    private final int maxWordLength;//最高句子长度
+
+    public MyKeyWord(SentenceConfig config, WordEmbedding wordEmbedding) throws Exception {
+        int vectorDimension = config.getWordVectorDimension();
+        int nerveLength = config.getKeyWordNerveDeep();
+        float studyPoint = config.getWeStudyPoint();
+        boolean isShowLog = config.isShowLog();
+        this.maxWordLength = config.getMaxWordLength();
+        typeNerveManager = new NerveManager(vectorDimension, vectorDimension, 2, nerveLength - 1, new Tanh(),
+                studyPoint, RZ.L1, 0.0001f);
+        typeNerveManager.initRnn(true, isShowLog);
+        this.wordEmbedding = wordEmbedding;
+        this.nerveLength = nerveLength;
+    }
+
+    public ModelParameter study(List<KeyWordForSentence> keyWordForSentenceList) throws Exception {//进行学习
+        for (KeyWordForSentence keyWordForSentence : keyWordForSentenceList) {
+            String word = keyWordForSentence.getSentence();
+            String keyWord = keyWordForSentence.getKeyWord();
+            myStudy(true, 1, word, keyWord, new HashMap<>());
+        }
+        return typeNerveManager.getModelParameter();
+    }
+
+    public void insertModel(ModelParameter modelParameter) throws Exception {
+        typeNerveManager.insertModelParameter(modelParameter);
+    }
+
+    public boolean isKeyWord(String word, long eventId) throws Exception {//是否存在关键词
+        boolean isKeyWord = false;
+        int key = myStudy(false, eventId, word, null, null);
+        if (key == 1) {
+            isKeyWord = true;
+        }
+        return isKeyWord;
+    }
+
+    private int myStudy(boolean isStudy, long eventId, String word, String keyWord, Map<Integer, Float> E) throws Exception {
+        int key = 0;
+        if (word.length() > 1) {
+            if (word.length() > maxWordLength) {
+                word = word.substring(0, maxWordLength);
+            }
+            WordBack wordBack = null;
+            int startIndex = -1;
+            int endIndex = -1;
+            if (!isStudy) {
+                wordBack = new WordBack();
+            } else if (keyWord != null && !keyWord.isEmpty()) {
+                startIndex = word.indexOf(keyWord);
+                endIndex = startIndex + keyWord.length() - 1;
+            }
+            int times = 1;
+            if (nerveLength < word.length()) {
+                times = word.length() - nerveLength + 1;
+            }
+            Matrix allFeature = wordEmbedding.getEmbedding(word, eventId, false).getFeatureMatrix();
+            for (int i = 0; i < times; i++) {
+                if (isStudy) {
+                    E.clear();
+                    if (i > endIndex || (i + nerveLength - 1) < startIndex) {//不相交
+                        E.put(2, 1f);
+                    } else {//相交
+                        E.put(1, 1f);
+                    }
+                } else {
+                    wordBack.clear();
+                }
+                Matrix feature;
+                if (times == 1) {
+                    feature = allFeature;//0,0,5,21-6 21
+                } else {
+                    feature = allFeature.getSonOfMatrix(i, 0, nerveLength, allFeature.getY());
+                }
+                List<Float> firstFeature = rowVectorToList(feature.getRow(0));//首行特征
+                studyNerve(eventId, typeNerveManager.getSensoryNerves(), firstFeature, feature, E, isStudy, wordBack);
+                if (!isStudy) {//需要返回结果
+                    key = wordBack.getId();
+                    if (key == 1) {
+                        break;
+                    }
+                }
+            }
+        }
+        return key;
+    }
+
+    private void studyNerve(long eventId, List<SensoryNerve> sensoryNerves, List<Float> featureList, Matrix rnnMatrix, Map<Integer, Float> E, boolean isStudy, OutBack convBack) throws Exception {
+        if (sensoryNerves.size() == featureList.size()) {
+            for (int i = 0; i < sensoryNerves.size(); i++) {
+                sensoryNerves.get(i).postMessage(eventId, featureList.get(i), isStudy, E, convBack, false, rnnMatrix);
+            }
+        } else {
+            throw new Exception("size not equals,feature size:" + featureList.size() + "," +
+                    "sensorySize:" + sensoryNerves.size());
+        }
+    }
+}