关键词查找【Aho-Corasick 算法】

【全程干货】程序员必备算法！AC自动机算法敏感词匹配算法！动画演示讲解，看完轻松掌握，面试官都被你唬住！！_哔哩哔哩_bilibili

著名的多模匹配算法

引入依赖：

<dependency>
   <groupId>org.ahocorasick</groupId>
   <artifactId>ahocorasick</artifactId>
   <version>0.6.3</version>
</dependency>

public List<String> findKeywordsInText(String text, List<String> keywords) {
   Trie trie = Trie.builder()
         .addKeywords(keywords)
         .build();
 
   long start = System.currentTimeMillis();
   Collection<Emit> emits = trie.parseText(text);
   logger.info("trie: " + (System.currentTimeMillis() - start));
 
   List<String> matchedKeywords = new ArrayList<>();
   for (Emit emit : emits) {
      matchedKeywords.add(emit.getKeyword());
   }
 
   return matchedKeywords;
}

main demo

public static void main(String[] args) {
   String BASE_URL = "https://www.baidu.com/s?wd=";
 
   String text = "abcd";
   List<String> wordList = new ArrayList<>();
   wordList.add("ab");
   wordList.add("bc");
   wordList.add("cd");
   wordList.add("a");
   wordList.add("bcd");
   wordList.add("abc");
 
   // 构建 Trie 时忽略大小写
 Trie trie = Trie.builder()
         .ignoreCase()
         .addKeywords(wordList)
         .build();
 
   Collection<Emit> emits = trie.parseText(text);
 
   // 按照关键词长度降序排序
 List<Emit> sortedEmits = new ArrayList<>(emits);
   sortedEmits.sort((e1, e2) -> e2.getKeyword().length() - e1.getKeyword().length());
 
   // 移除开始位置在前面词的开始和结束中间的关键词
 Set<String> processedKeywords = new HashSet<>();
   List<Emit> filteredEmits = new ArrayList<>();
   for (Emit emit : sortedEmits) {
      String keyword = emit.getKeyword();
      boolean shouldAdd = true;
      for (Emit filteredEmit : filteredEmits) {
         if (emit.getStart() >= filteredEmit.getStart() && emit.getStart() <= filteredEmit.getEnd()) {
            shouldAdd = false;
            break;
         }
      }
      if (shouldAdd) {
         filteredEmits.add(emit);
         processedKeywords.add(keyword);
      }
   }
 
   // 使用TreeMap来保持emit的start位置顺序
 TreeMap<Integer, Emit> firstOccurrences = new TreeMap<>();
   for (Emit emit : filteredEmits) {
      firstOccurrences.putIfAbsent(emit.getStart(), emit);
   }
 
   StringBuilder result = new StringBuilder(text);
   int offset = 0;
   processedKeywords.clear();
   for (Emit emit : firstOccurrences.values()) {
      String keyword = emit.getKeyword();
      if (!processedKeywords.contains(keyword)) {
         int startIndex = emit.getStart() + offset;
         int endIndex = emit.getEnd() + 1 + offset;
         String link = "<a href=\"" + BASE_URL + keyword + "\" class='cnt-inner-link'>" + keyword + "</a>";
         result.replace(startIndex, endIndex, link);
         offset += link.length() - keyword.length();
         processedKeywords.add(keyword);
      }
   }
   System.out.println(result.toString());
}