resumesend.java

用ＪＡＶＡ语言实现ＪＯＢ ＨＵＮＴＥＲ的功能

package jobhunter;

import java.net.*;
import java.io.*;
import java.util.*;
import javax.mail.*;
import javax.mail.internet.*;
import javax.activation.*;


public class ResumeSend {
  public ResumeSend(int size) {
    this.keyQue = new Queue(size);
  }
  private String[] bakS = new String[]{".cn",".com.cn",".com",".net.cn",".gov.cn",".edu.cn",".net",".gov",".edu",".ac",".info",".jp",".hk",".tw",".uk"};
  public static Queue keyQue;
  public static int threads = 0;
  public static void main(String[] args) throws Exception {
    ResumeSend rssd = new ResumeSend(1000000);
    File f = new File("emailsbak");
    boolean onlyBak = false;
    if (f.exists()) {
      synchronized (f) {
        System.out.print("Checking Backup Emails....");
        try {
          RandomAccessFile raf = new RandomAccessFile("emailsbak", "r");
          String emLine;
          while ( (emLine = raf.readLine()) != null) {
            rssd.keyQue.enqueue(emLine);
          }
        }
        catch (Exception e) {
          e.printStackTrace();
        }
        System.out.println("BackFile Get " + rssd.keyQue.size() + " Emails.");
        f.deleteOnExit(); //how to delete timely???
      }

      onlyBak = true;
      if (rssd.keyQue.size() > 0) {
        System.out.println("Are you want to send only back emails?[Y/n]:");
        BufferedReader bin = new BufferedReader(new InputStreamReader(System.in));
        String answ = bin.readLine();
        if (answ.toLowerCase().equals("n")) {
          onlyBak = false;
        }
      }
    } //end exists

    if (!onlyBak) {
      System.out.println("Begin Collecting Email....");
      for(int i=0;i<args.length;i++){
          int loc = Integer.parseInt(args[i]);
          switch (loc) {
            case 5: {
              try{
                rssd.chinahr_job_search(5); //软件，北京
                System.out.println("ChinaHr 北京搜索成功！");
                System.out.println("Now Emails: "+keyQue.size());
              }catch(Exception e){
                System.out.println("ChinaHr 北京搜索失败！");
                e.printStackTrace(System.out);
              }
              break;
            }
            case 115: {
              try{
                rssd.chinahr_job_search(115); //软件，上海
                System.out.println("ChinaHr 上海搜索成功！");
                System.out.println("Now Emails: "+keyQue.size());
              }catch(Exception e){
                System.out.println("ChinaHr 上海搜索失败！");
                e.printStackTrace(System.out);
              }

              break;
            }
            case 125: {
              try{
                rssd.chinahr_job_search(125); //软件，深圳
                System.out.println("ChinaHr 深圳搜索成功！");
                System.out.println("Now Emails: "+keyQue.size());
              }catch(Exception e){
                System.out.println("ChinaHr 深圳搜索失败！");
                e.printStackTrace(System.out);
              }
              break;
            }
            case 100: {
              try{
                rssd.w51job_search("0100"); //软件 北京
                System.out.println("51job 北京搜索成功！");
                System.out.println("Now Emails: "+keyQue.size());
              }catch(Exception e){
                System.out.println("51job 北京搜索失败！");
                e.printStackTrace(System.out);
              }
              break;
            }
            case 200: {
              try{
                rssd.w51job_search("0200"); //软件，上海
                System.out.println("51job 上海搜索成功！");
                System.out.println("Now Emails: "+keyQue.size());
              }catch(Exception e){
                System.out.println("51job 上海搜索失败！");
                e.printStackTrace(System.out);
              }

              break;
            }
            case 400: {
              try{
                rssd.w51job_search("0400"); //软件，深圳
                System.out.println("51job 深圳搜索成功！");
                System.out.println("Now Emails: "+keyQue.size());
              }catch(Exception e){
                System.out.println("51job 深圳搜索失败！");
                e.printStackTrace(System.out);
              }
              break;
            }
            case 1: {
              try{
                rssd.zhaopin_job_search(); //智联招聘网，北京，上海，深圳
                System.out.println("智联招聘网 北京、上海、深圳搜索成功！");
                System.out.println("Now Emails: "+keyQue.size());
              }catch(Exception e){
                System.out.println("5智联招聘网 北京、上海、深圳搜索失败！");
                e.printStackTrace(System.out);
              }
              break;
            }
          } //switch
        }//end for
    }//end if
    System.out.println("GET " + keyQue.size() + " EMAILS! Now Sending...");
    SendThread sendT = new SendThread(10, 30);//每次最多连续发15封
    sendT.run();

  }


  public void chinahr_job_search(int loc) throws Exception{
    String baseUrl = "http://searchjob.chinahr.com/";
    String head = "GET /jobsearch.asp?page=";
    String body ="&mid=&order=0 HTTP/1.1"+  // /jobsearch.asp HTTP/1.1"+
    "\r\n"+
    "Accept-Language: zh-cn"+
    "\r\n"+
    "Accept: image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, application/vnd.ms-powerpoint, application/vnd.ms-excel, application/msword, application/x-shockwave-flash, */*"+
    "\r\n"+
    "Host: searchjob.chinahr.com"+
    "\r\n"+
    "User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.0; (R1 1.1))"+
    "\r\n"+
    "Connection: Close"+
    "\r\n"+
    "Cache-Control: no-cache"+
    "\r\n"+
    "Cookie: SrchJob=page=";
    String cook ="&where=+FROM+searchjob++WHERE++and+occ%5Fid+%3D+601+and+job%5Fid+in+%28Select+job%5Fid+From+searchjobloc+Where+searchjobloc%2Ejob%5Fid+%3D+searchjob%2Ejob%5Fid++and+loc%5Fid+in+%28"+loc+"%29%29+and+job%5Ftitle+like+%27%25%C8%ED%BC%FE%25%27+and+update%5Fdate+%3E%3D+DATEADD%28day%2C+%2D7%2C+getdate%28%29%29+; JobAgent="+
    "\r\n"+
    "\r\n";

    int str_p=1;
    int end_p=2;
    while(str_p<end_p)
    {
      Socket sk1 = new Socket("211.100.5.249",80);
      BufferedOutputStream os = new BufferedOutputStream(sk1.getOutputStream());
      HTTPInputStream is = new HTTPInputStream(sk1.getInputStream());
      BufferedReader reader = new BufferedReader(new InputStreamReader(is));
      os.write( (head+str_p+body+(str_p-1)+cook).getBytes());
      os.flush();
      String temp_line,line;
      int j_p = 0;
      int u_p = 0;
      while ( (temp_line = reader.readLine()) != null) {
       // System.out.println(temp_line);
        //<A HREF="jobsearch.asp?page=6&mid=&order=0">
        line = temp_line;
        while( (u_p = line.indexOf("jobsearch.asp?page=")) > 0) {
          String s1 = line.substring(u_p + 19);
          line = s1;
          int p_p = s1.indexOf("&");
          if ( p_p > 100) {
            continue;
          }
          String sPg = s1.substring(0, p_p);
          int newP = (new Integer(sPg)).intValue();
          if (newP < end_p) {
            continue;
          }
          end_p = newP+1;
        }
        while ( (j_p = line.indexOf("Javascript:view(")) > 0) {
          //Javascript:view('22203241092220128201','2002-5-13')
          String s1 = line.substring(j_p + 17);
          line =s1;
          int com = s1.indexOf(",");
          int rk = s1.indexOf(")");
          String lNum = s1.substring(0, com - 1);
          String dAte = s1.substring(com + 2, rk - 1);
          int fH = dAte.lastIndexOf("-");
          String YM = dAte.substring(0, fH);
          String DY = dAte.substring(fH + 1);
          String emUrl = baseUrl + "jobs/" + YM + "/" + DY + "/" + lNum +
              ".htm";
          try{
            URL_email_search(emUrl);
          }catch(Exception e){
            System.out.println(emUrl+" Failed!");
            e.printStackTrace(System.out);
          }
        }
      }//end while
      str_p++;
    }
  }

  public void URL_email_search(String sUrl) throws Exception{
    final String fUrl = sUrl;
    Thread emsearch = new Thread(){
      public void run(){
        try {
          ResumeSend.threads++;
          URL url = new URL(fUrl);
          BufferedReader reader = null;
          boolean tryCon = true;
          while(tryCon){
            try{
              HttpURLConnection urlc = (HttpURLConnection) url.openConnection();
              reader = new BufferedReader(
                  new InputStreamReader(urlc.getInputStream()));
              tryCon = false;
            }catch(Exception e){
              int trys=1+(int)(10*Math.random());
              tryCon = true;
              e.printStackTrace(System.out);
              System.out.println("Connect Failed! Let us wait "+trys+" minutes for "+fUrl+" again!");
              this.sleep(1000*60*trys);
            }
          }
          String temp_line, line;
          int m_p = 0;
          while ( (temp_line = reader.readLine()) != null) {
            line = temp_line;
            while ( (m_p = line.indexOf("mailto:")) > 0) {
              String s1 = line.substring(m_p + 7);
              line = s1;
              int ed1 = s1.indexOf("?");
              int ed2 = s1.indexOf(">");
              int ed3 = ed1 < 0 ? ed2 :
                  (ed2 < 0 ? ed1 : ( (ed1 < ed2) ? ed1 : ed2));
              if (ed3 < 1)
                ed3 = Math.min(40, s1.length());
              String lem = s1.substring(0, ed3).toLowerCase();
              //int end = s1.indexOf("?");
              int bak_p = 100;
              int bak_i = 0;
              for (int j = 0; j < bakS.length; j++) {
                int tt = 0;
                tt = lem.indexOf(bakS[j]);
                if ( (tt > 0) && (tt < bak_p)) {
                  bak_p = tt;
                  bak_i = j;
                }

              }
              String email;
              if(bak_p<100){
                email = lem.substring(0, bak_p) + bakS[bak_i];
              }else{
                email = lem.substring(0, ed3);
              }
              if(email.equals("[e-mail]")) continue;
              synchronized(keyQue){
                keyQue.add(email);
              }
            }
          }
        }catch(Exception e){
          e.printStackTrace(System.out);
        }
        ResumeSend.threads--;
      }
    };
    emsearch.start();
  }

  private static final String safeCharList = "$-_.+!*'(),";
  public static String decode(String s)  {
      if (s.equals("")) return "";
      int len = s.length();
      StringBuffer buf = new StringBuffer();

      for (int i = 0; i < len; i++) {
          char c = s.charAt(i);
          if (Character.isLetterOrDigit(c))
              buf.append(c);
          else if (c == '+')
              buf.append(' ');
          else if (safeCharList.indexOf(c) != -1)
              buf.append(c);
          else if (c == '%') {
              if (i >= len - 2) {
                      return null;
              } else {
                  char[] hexChars = new char[2];

                  hexChars[0] = s.charAt(++i);
                  hexChars[1] = s.charAt(++i);

                  String hexval = new String(hexChars);
                  try {
                      long read = hexToLong(hexval);
                      if (read == 0)
                          return null;
                      buf.append(new Character((char) read));
                  }
                  catch (NumberFormatException nfe) {
                      return null;
                  }
              }
          }
          else
              buf.append(c);
              // throw new URLEncodedFormatException(s);
      }
      return buf.toString();
  }
  public static final long hexToLong(String hex) throws NumberFormatException {
      int len = hex.length();
      if (len > 16)
          throw new NumberFormatException();

      long l = 0;
      for (int i=0; i<len; i++) {
          l<<=4;
          int c = Character.digit(hex.charAt(i), 16);
          if (c < 0)
              throw new NumberFormatException();
          l |= c;
      }
      return l;
  }


// Convenience class for reading client requests
 class HTTPInputStream extends FilterInputStream {
 public HTTPInputStream(InputStream in) {
  super(in);
 }
 // Get a line
 public String readLine() throws IOException {
  StringBuffer result=new StringBuffer();
  boolean finished = false;
  boolean cr = false;
  do {
   int ch = -1;
   ch = read();
   if(ch==-1) return result.toString();
   result.append((char) ch);
   if(cr && ch==10){
    result.setLength(result.length()-2);
    return result.toString();
   }
   if(ch==13) cr = true;
   else cr=false;