⭐ 欢迎来到虫虫下载站! | 📦 资源下载 📁 资源专辑 ℹ️ 关于我们
⭐ 虫虫下载站

📄 indextypefilter.java

📁 java写的搜索引擎网络爬虫 java写的搜索引擎网络爬虫 java写的搜索引擎网络爬虫
💻 JAVA
字号:
/* 
 * FileName IndexTypeFilter.java
 * Create Time 2006-5-16 16:51:43
 * Author shiwei
 * Descript 判断是否属于首页类型
 * Version 
 */

package com.snoics.reptile.regex.filter;

import java.util.Iterator;
import java.util.List;
import java.util.regex.Pattern;

import com.snoics.base.util.StringClass;
import com.snoics.base.util.regex.Regex;
import com.snoics.reptile.system.common.Common;
import com.snoics.reptile.system.common.CommonObject;

public class IndexTypeFilter implements IRegexFilter{
	private String url="";
	private Regex regex=new Regex();
	private CommonObject commonObject=new CommonObject();
	
	/**
	 * 获取当前URL
	 * @return Returns the url.
	 */
	public String getUrl() {
		return url;
	}
	/**
	 * 设置当前URL
	 * @param url The url to set.
	 */
	public void setUrl(String url) {
		this.url = url;
	}
	
	/**
	 * 判断是否属于首页类型
	 * @return boolean
	 */
	public boolean filter() {
		if(url.indexOf("/")>-1) {
			url=StringClass.getLastString(url,"/");
		}
		if(url.equals("")) {
			url=" ";
		}
		url=StringClass.getPreString(url,"?");
		String indexUrlFlags=commonObject.getConfigInfo(Common.CONFIGFILE_NODE_INDEXPAGE);
		String startPageFlags=commonObject.getConfigInfo(Common.CONFIGFILE_NODE_STARTPAGE);
		List indexpagelist=null;
		List startPageList=null;
		indexpagelist=StringClass.getInterString(Common.STRING_SEPARATE_FLAG,indexUrlFlags);
		startPageList=StringClass.getInterString(Common.STRING_SEPARATE_FLAG,startPageFlags);
		boolean flag=false;

		if((startPageList!=null)&&(startPageList.size()>0)){
			Iterator it=startPageList.iterator();
			while(it.hasNext()){
				String startPageFlag=(String)it.next();
				if(regex.find(url,startPageFlag,Pattern.CASE_INSENSITIVE)) {
					return true;
				}
			}
		}
		
		if((indexpagelist!=null)&&(indexpagelist.size()>0)){
			Iterator iterator=indexpagelist.iterator();
			while(iterator.hasNext()) {
				String indexUrlFlag=(String)iterator.next();
				if(regex.find(url,indexUrlFlag,Pattern.CASE_INSENSITIVE)) {
					return true;
				}
			}
		}
		return flag;
	}
	

	
	public static void main(String[] args) {
		IndexTypeFilter indexTypeFilter=new IndexTypeFilter();
		String url="HTTPs://fjdsklfjkdslfjldsf/inde.html";
		indexTypeFilter.setUrl(url);
		System.out.println(indexTypeFilter.filter());
		
		url="index.jsp";
		indexTypeFilter.setUrl(url);
		System.out.println(indexTypeFilter.filter());
		
		url="/index.jspa";
		indexTypeFilter.setUrl(url);
		System.out.println(indexTypeFilter.filter());
		
		url="/inDEx.jspa";
		indexTypeFilter.setUrl(url);
		System.out.println(indexTypeFilter.filter());
		
		url="/inDExa.jspa";
		indexTypeFilter.setUrl(url);
		System.out.println(indexTypeFilter.filter());

		url="/";
		indexTypeFilter.setUrl(url);
		System.out.println(indexTypeFilter.filter());
		
		url="  ";
		indexTypeFilter.setUrl(url);
		System.out.println(indexTypeFilter.filter());
	}
	
}

⌨️ 快捷键说明

复制代码 Ctrl + C
搜索代码 Ctrl + F
全屏模式 F11
切换主题 Ctrl + Shift + D
显示快捷键 ?
增大字号 Ctrl + =
减小字号 Ctrl + -