如何利用Java实现全文检索功能-成都快上网建站

快上网建站品牌

13518219792

如何利用Java实现全文检索功能

今天就跟大家聊聊有关如何利用Java实现全文检索功能，可能很多人都不太了解，为了让大家更加了解，小编给大家总结了以下内容，希望大家根据这篇文章可以有所收获。

创新互联是一家集网站建设,金东企业网站建设,金东品牌网站建设,网站定制,金东网站建设报价,网络营销,网络优化,金东网站推广为一体的创新建站企业，帮助传统企业提升企业形象加强企业竞争力。可充分满足这一群体相比中小企业更为丰富、高端、多元的互联网需求。同时我们时刻保持专业、时尚、前沿，时刻以成就客户成长自我，坚持不断学习、思考、沉淀、净化自己，让我们为更多的企业打造出实用型网站。

第一步，构建Trie树，定义Node类型：

/**
 * Created by zhaoyy on 2017/2/7.
 */
interface Node {

  char value();

  boolean exists();

  boolean isRoot();

  Node parent();

  Node childOf(char c);

  Node fail();

  void setFail(Node node);

  void setExists(boolean exists);

  void add(Node child);

  List children();
}

第二步，实现两种Node，如果词汇全是可打印的ASCII字符，就采用AsciiNode，否则（比如包含汉字），使用基于hash表的MapNode；这两种Node均集成自AbstractNode:

/**
 * Created by zhaoyy on 2017/2/8.
 */
abstract class AbstractNode implements Node {

  private static final char EMPTY = '\0';
  private final char value;
  private final Node parent;
  private boolean exists;
  private Node fail;

  public AbstractNode(Node parent, char value) {
    this.parent = parent;
    this.value = value;
    this.exists = false;
    this.fail = null;
  }

  public AbstractNode() {
    this(null, EMPTY);
  }


  private static String tab(int n) {
    StringBuilder builder = new StringBuilder();
    for (int i = 0; i < n; i++) {
      builder.append('\t');
    }
    return builder.toString();
  }

  private static String toString(Node node, int depth) {
    StringBuilder builder = new StringBuilder();
    String tab = tab(depth);
    Node fail = node.fail();
    Node parent = node.parent();
    builder
        .append(tab)
        .append('<')
        .append(node.value())
        .append(" exists=\"")
        .append(node.exists())
        .append('"')
        .append(" parent=\"")
        .append(parent == null ? "null" : parent.value())
        .append('"')
        .append(" fail=\"")
        .append(fail == null ? "null" : fail.value())
        .append("\">\r\n");
    for (Node child : node.children())
      builder.append(toString(child, depth + 1));
    builder
        .append(tab)
        .append("\r\n");

    return builder.toString();
  }

  @Override
  public char value() {
    return value;
  }


  @Override
  public boolean exists() {
    return exists;
  }

  @Override
  public boolean isRoot() {
    return value == EMPTY;
  }

  @Override
  public Node parent() {
    return parent;
  }

  @Override
  public Node fail() {
    return fail;
  }

  @Override
  public void setFail(Node node) {
    this.fail = node;
  }

  @Override
  public void setExists(boolean exists) {
    this.exists = exists;
  }

  @Override
  public String toString() {
    return toString(this, 0);
  }
}

/////////////////////////////////////////////////////////////////////////////////////////////

/**
 * Created by zhaoyy on 2017/2/8.
 */
final class AsciiNode extends AbstractNode implements Node {


  private static final char FROM = 32;
  private static final char TO = 126;
  private final Node[] children;


  public AsciiNode() {
    super();
    this.children = new Node[TO - FROM + 1];
  }

  public AsciiNode(Node parent, char value) {
    super(parent, value);
    this.children = new Node[TO - FROM + 1];
  }

  @Override
  public Node childOf(char c) {
    if (c >= FROM && c <= TO)
      return children[(int) c - FROM];
    else return null;
  }

  @Override
  public void add(Node child) {
    int index = (int) child.value();
    children[index - FROM] = child;
  }

  @Override
  public List children() {
    List nodes = new ArrayList();
    for (Node child : children)
      if (child != null)
        nodes.add(child);
    return nodes;
  }
}


//////////////////////////////////////////////////////////////////////////////////////////////

/**
 * Created by zhaoyy on 2017/2/8.
 */
final class MapNode extends AbstractNode implements Node {

  private final Map children;

  public MapNode() {
    super();
    this.children = new HashMap();
  }

  public MapNode(Node parent, char value) {
    super(parent, value);
    this.children = new HashMap();
  }

  @Override
  public Node childOf(char c) {
    return children.get(c);
  }

  @Override
  public void add(Node child) {
    children.put(child.value(), child);
  }

  @Override
  public List children() {
    List nodes = new ArrayList();
    nodes.addAll(children.values());
    return nodes;
  }
}

第三步，

首先定义一个Node构造器：

/**
 * Created by zhaoyy on 2017/2/8.
 */
public interface NodeMaker {

  Node make(Node parent, char value);

  Node makeRoot();
}

然后构建AC自动机，实现创建及查找方法：

/**
 * Created by zhaoyy on 2017/2/7.
 */
public final class WordTable {

  private final Node root;


  private WordTable(Collection<? extends CharSequence> words, NodeMaker maker) {
    Node root = buildTrie(words, maker);
    setFailNode(root);
    this.root = root;
  }

  public static WordTable compile(Collection<? extends CharSequence> words) {
    if (words == null || words.isEmpty())
      throw new IllegalArgumentException();
    final NodeMaker maker;
    if (isAscii(words))
      maker = new NodeMaker() {
        @Override
        public Node make(Node parent, char value) {
          return new AsciiNode(parent, value);
        }

        @Override
        public Node makeRoot() {
          return new AsciiNode();
        }
      };
    else maker = new NodeMaker() {
      @Override
      public Node make(Node parent, char value) {
        return new MapNode(parent, value);
      }

      @Override
      public Node makeRoot() {
        return new MapNode();
      }
    };
    return new WordTable(words, maker);
  }

  private static boolean isAscii(Collection<? extends CharSequence> words) {
    for (CharSequence word : words) {
      int len = word.length();
      for (int i = 0; i < len; i++) {
        int c = (int) word.charAt(i);
        if (c < 32 || c > 126)
          return false;
      }
    }
    return true;
  }

  private static Node buildTrie(Collection<? extends CharSequence> sequences, NodeMaker maker) {
    Node root = maker.makeRoot();
    for (CharSequence sequence : sequences) {
      int len = sequence.length();
      Node current = root;
      for (int i = 0; i < len; i++) {
        char c = sequence.charAt(i);
        Node node = current.childOf(c);
        if (node == null) {
          node = maker.make(current, c);
          current.add(node);
        }
        current = node;
        if (i == len - 1)
          node.setExists(true);
      }
    }
    return root;
  }

  private static void setFailNode(final Node root) {
    root.setFail(null);
    Queue queue = new LinkedList();
    queue.add(root);
    while (!queue.isEmpty()) {
      Node parent = queue.poll();
      Node temp;
      for (Node child : parent.children()) {
        if (parent.isRoot())
          child.setFail(root);
        else {
          temp = parent.fail();
          while (temp != null) {
            Node node = temp.childOf(child.value());
            if (node != null) {
              child.setFail(node);
              break;
            }
            temp = temp.fail();
          }
          if (temp == null)
            child.setFail(root);
        }
        queue.add(child);
      }
    }
  }


  public boolean findAnyIn(CharSequence cs) {
    int len = cs.length();
    Node node = root;
    for (int i = 0; i < len; i++) {
      Node next = node.childOf(cs.charAt(i));
      if (next == null) {
        next = node.fail();
        if (next == null) {
          node = root;
          continue;
        }
      }
      if (next.exists())
        return true;
    }

    return false;
  }

  public List search(CharSequence cs) {
    if (cs == null || cs.length() == 0)
      return Collections.emptyList();
    List result = new ArrayList();
    int len = cs.length();
    Node node = root;
    for (int i = 0; i < len; i++) {
      Node next = node.childOf(cs.charAt(i));
      if (next == null) {
        next = node.fail();
        if (next == null) {
          node = root;
          continue;
        }
      }
      if (next.exists()) {
        MatchInfo info = new MatchInfo(i, next);
        result.add(info);
        node = root;
        continue;
      }
      node = next;
    }
    return result;
  }

  @Override
  public String toString() {
    return root.toString();
  }
}

定义一个保存查找结果的实体：

/**
 * Created by zhaoyy on 2017/2/7.
 */
public final class MatchInfo {

  private final int index;
  private final String word;

  public MatchInfo(int index, String word) {
    this.index = index;
    this.word = word;
  }

  public MatchInfo(int index, Node node) {
    StringBuilder builder = new StringBuilder();
    while (node != null) {
      if (!node.isRoot())
        builder.append(node.value());
      node = node.parent();
    }
    String word = builder.reverse().toString();
    this.index = index + 1 - word.length();
    this.word = word;
  }


  public int getIndex() {
    return index;
  }

  public String getWord() {
    return word;
  }

  @Override
  public String toString() {
    return index + ":" + word;
  }
}

第四步，调用Demo：

public static void main(String[] args) {
    List list = Arrays.asList("say", "her", "he", "she", "shr", "alone");
    WordTable table = WordTable.compile(list);
    System.out.println(table);
    System.out.println(table.search("1shesaynothingabouthislivinghimalone"));
  }

以下是输出结果：

< exists="false" parent="null" fail="null">
 
 
  
  
 
 
  
  
  
  
 
 
 
 
  
  
 
 
 
 
  
  
   
   
  
  
 
 


[1:she, 4:say, 31:alone]

看完上述内容，你们对如何利用Java实现全文检索功能有进一步的了解吗？如果还想了解更多知识或者相关内容，请关注创新互联行业资讯频道，感谢大家的支持。


分享文章：如何利用Java实现全文检索功能
浏览路径：http://kswjz.com/article/gecgdo.html
扫二维码与项目经理沟通

我们在微信上24小时期待你的声音

解答本文疑问/技术咨询/运营咨询/技术建议/互联网交流

其他资讯

行业动态

企业网站建设的重要性！

现在虽然是移动互联网时代，但企业网站依然重要，包含PC站点，移动站。可以说企业网站关系企业的未来发展和前途，尤其对中小企业更是如此，一些中小企业老板，对自己的名片很在乎，因为这是个门面。...

服务项目

联系吧 在百度地图上找到我们

电话：13518219792

如遇占线或暂未接听请拨：136xxx98888

业务咨询 技术咨询 售后服务
网站设计
高端网站设计推广
成都网站设计
重庆网站设计
响应式网站设计
网站制作
手机网站制作
成都网站制作
达州网站制作
成都网站制作
联系我们
电话：13518219792
邮箱：631063699@qq.com
地址：成都青羊区锦天国际1002号
网址：www.kswjz.com
网站建设
成都企业网站建设公司
重庆企业网站建设
成都做网站建设公司
成都网站建设

微信二维码

Copyright © 2002-2023 www.kswjz.com 快上网建站品牌 QQ：244261566 版权所有 备案号：蜀ICP备19037934号