feat: 关键词提取

src/main/java/cn/com/poc/long_document/domain/keyword/Concepts.java

+package cn.com.poc.long_document.domain.keyword;
+
+/**
+ * @author alex.yao
+ * @date 2025/9/4
+ */
+public class Concepts {
+}

src/main/java/cn/com/poc/long_document/domain/keyword/KeyWord.java

+package cn.com.poc.long_document.domain.keyword;
+
+import java.util.List;
+
+/**
+ * @author alex.yao
+ * @date 2025/9/4
+ */
+public class KeyWord {
+
+    /**
+     * 关键词云图 [10个以内]
+     */
+    private List<String> keyWord;
+
+    /**
+     * 关键词提及频率 [10个以内]
+     */
+    private List<MentionFrequency> mentionFrequency;
+
+    /**
+     * 相关概念 [10个以内]
+     */
+    List<String> concepts;
+
+    public List<String> getKeyWord() {
+        return keyWord;
+    }
+
+    public void setKeyWord(List<String> keyWord) {
+        this.keyWord = keyWord;
+    }
+
+    public List<MentionFrequency> getMentionFrequency() {
+        return mentionFrequency;
+    }
+
+    public void setMentionFrequency(List<MentionFrequency> mentionFrequency) {
+        this.mentionFrequency = mentionFrequency;
+    }
+
+    public List<String> getConcepts() {
+        return concepts;
+    }
+
+    public void setConcepts(List<String> concepts) {
+        this.concepts = concepts;
+    }
+}

src/main/java/cn/com/poc/long_document/domain/keyword/MentionFrequency.java

+package cn.com.poc.long_document.domain.keyword;
+
+import java.util.List;
+
+/**
+ * @author alex.yao
+ * @date 2025/9/4
+ */
+public class MentionFrequency {
+
+    /**
+     * 关键词
+     */
+    private String keyword;
+
+    /**
+     * 关联句子
+     */
+    private List<String> sentences;
+
+    /**
+     * 出现频率
+     */
+    private Integer frequency;
+
+    public String getKeyword() {
+        return keyword;
+    }
+
+    public void setKeyword(String keyword) {
+        this.keyword = keyword;
+    }
+
+    public List<String> getSentences() {
+        return sentences;
+    }
+
+    public void setSentences(List<String> sentences) {
+        this.sentences = sentences;
+    }
+
+    public Integer getFrequency() {
+        return frequency;
+    }
+
+    public void setFrequency(Integer frequency) {
+        this.frequency = frequency;
+    }
+}

src/main/java/cn/com/poc/long_document/dto/LongTextSummaryDto.java

 package cn.com.poc.long_document.dto;
 
 import cn.com.poc.long_document.domain.MindMap;
+import cn.com.poc.long_document.domain.keyword.KeyWord;
 import com.fasterxml.jackson.annotation.JsonIgnoreProperties;
 import com.fasterxml.jackson.annotation.JsonInclude;
 
@@ -18,15 +19,15 @@ public class LongTextSummaryDto {
 
     private String corePoint;
 
-    private List<String> keyword;
+    private KeyWord keyword;
 
     private MindMap mindMap;
 
-    public List<String> getKeyword() {
+    public KeyWord getKeyword() {
         return keyword;
     }
 
-    public void setKeyword(List<String> keyword) {
+    public void setKeyword(KeyWord keyword) {
         this.keyword = keyword;
     }
 

src/main/java/cn/com/poc/meeting/cache/KeyWordCounter.java

@@ -38,6 +38,29 @@ public class KeyWordCounter {
         return sortedMap;
     }
 
+    /**
+     * 提取关键词与文本中匹配的句子
+     *
+     * @param content  文本
+     * @param keywords 关键词
+     */
+    public static Map<String, List<String>> keywordMatchSentence(String content, List<String> keywords) {
+        Map<String, List<String>> result = new LinkedHashMap<>();
+        // 将关键词转换为正则表达式模式（注意转义特殊字符）
+        for (String keyword : keywords) {
+            String regexPattern = Pattern.quote(keyword);
+            String sentenceRegex = "[^,，:：;；.。!?！？\\n]*" + regexPattern + "[^,，:：;；.。!?！？\\n]*[,，:：;；.。!?！？\\n]";
+            Pattern pattern = Pattern.compile(sentenceRegex, Pattern.CASE_INSENSITIVE);
+            Matcher matcher = pattern.matcher(content);
+            // 查找并输出匹配的句子
+            List<String> sentences = new ArrayList<>();
+            while (matcher.find()) {
+                sentences.add(matcher.group().trim());
+            }
+            result.put(keyword, sentences);
+        }
+        return result;
+    }
 
     /**
      * 缓存关键词计数

src/test/java/cn/com/poc/meeting/MeetingTest.java

 package cn.com.poc.meeting;
 
 import cn.com.yict.framemax.core.spring.SingleContextInitializer;
+import org.junit.Test;
 import org.junit.runner.RunWith;
 import org.springframework.test.context.ContextConfiguration;
 import org.springframework.test.context.junit4.SpringJUnit4ClassRunner;
 import org.springframework.test.context.web.WebAppConfiguration;
 
+import java.util.ArrayList;
+import java.util.List;
+import java.util.regex.Matcher;
+import java.util.regex.Pattern;
+
 @RunWith(SpringJUnit4ClassRunner.class)
 @ContextConfiguration(initializers = SingleContextInitializer.class)
 @WebAppConfiguration
 public class MeetingTest {
 
+    @Test
+    public void test() {
+        String keyword = "an example";
+        String text = "This is an example sentence. Another example? Yes, this is an example.";
+
+        // 将关键词转换为正则表达式模式（注意转义特殊字符）
+        String regexPattern = Pattern.quote(keyword);
+
+        // 构建匹配句子的正则表达式
+        // 匹配以任意标点符号或换行结尾的句子，包含关键词
+        // 支持中英文混合关键词匹配
+        String sentenceRegex = "[^,，:：;；.。!?！？\\n]*" + regexPattern + "[^,，:：;；.。!?！？\\n]*[,，:：;；.。!?！？\\n]";
+
+        Pattern pattern = Pattern.compile(sentenceRegex, Pattern.CASE_INSENSITIVE);
+        Matcher matcher = pattern.matcher(text);
 
+        // 查找并输出匹配的句子
+        while (matcher.find()) {
+            System.out.println("\"" + matcher.group().trim() + "\"");
+        }
+    }
 }