简单的58爬虫一份

package spider_test; import java.util.ArrayList; import java.util.regex.Matcher; import java.util.regex.Pattern; public class Jobs { public String Information; public String Company; public String Job; public String Number; public String Experience; public String Education; public String Location; public String Wages; public String Others; public String Url; public String Detail; public Jobs(){ Information=""; Company=""; Job=""; Number=""; Experience=""; Education=""; Location=""; Wages=""; Others=""; } public String toString() { return writeString(); //return ("公司："+ Company +"职位"+Job+"招聘人数"+Number+"工作经验"+Experience+"学历"+Education+"工作地点"+Location+"工资"+Wages+"其他待遇"+Others); } /*public String Getinformation(String content,String aim,String end){ String result=""; int flag1,flag2; flag1=content.indexOf(aim); flag2=content.indexOf(end); System.out.println(content); System.out.println(flag1); System.out.println(flag2); result=content.substring(flag1,flag2); return result; }*/ public static String Getinformation(String content,String aim){ Pattern urlPattern = Pattern.compile(aim); Matcher urlMatcher = urlPattern.matcher(content); // 问题和链接要均能匹配到 boolean isFind = urlMatcher .find(); return urlMatcher.group(0); } public String writeString() { // 拼接写入本地的字符串 String result = ""; /*Company=Getinformation(Information,"<h2.*</h2>"); result += "\r\n"+"公司:" + Company+" "; Job=Getinformation(Information,"<span>招聘职位：.*?</li>"); result += "招聘职位:" + Job+" "; Number=Getinformation(Information,"招聘人数：.*?</li>"); Experience=Getinformation(Information,"工作经验：.*?</li>"); Education=Getinformation(Information,"学历要求：.*?</li>"); Location=Getinformation(Information,"工作地点：.*?</li>"); Wages=Getinformation(Information,"转正工资：.*?<p>"); //Others=Getinformation(Information,"fl.*?<p>"); */ Url=Getinformation(Information,"http.*?'"); //System.out.println(Url); /*result += "招聘人数:" + Number+" "; result += "工作经验:" + Experience+" "; result += "学历要求:" + Education+" "; result += "工作地点:" + Location+" "; result += "转正工资" + Wages+" ";*/ result += "转正工资" + Url+" "; Detail=Getmoredetail.Getdetail(Url); //result += "其他条件" + Others+" "; //result += "属性：" + property + "\r\n"; //result += "链接：" + Url + "\r\n"; //for (int i = 0; i < elements.size(); i++) { // result += "回答" + i + "：" + answers.get(i) + "\r\n\r\n"; // } // result += "\r\n\r\n\r\n\r\n"; // 将其中的html标签进行筛选 //result = result.replaceAll("<br>", ""); //result = result.replaceAll("<.*?>",""); // result = result.replaceAll("\\s*"," "); //result+="\n"; //result = result.replaceAll("[",""); //result = result.replaceAll("]",""); return result; } }