验证码识别，爬虫永远的话题~

用打码兔总体的体验就是单线程速度太慢~

import java.io.IOException;

import java.net.MalformedURLException;

import java.util.Date;

import org.apache.log4j.Logger;

import org.jsoup.Jsoup;

import org.jsoup.nodes.Document;

import org.jsoup.select.Elements;

import com.gargoylesoftware.htmlunit.BrowserVersion;

import com.gargoylesoftware.htmlunit.FailingHttpStatusCodeException;

import com.gargoylesoftware.htmlunit.WebClient;

import com.gargoylesoftware.htmlunit.html.HtmlButton;

import com.gargoylesoftware.htmlunit.html.HtmlForm;

import com.gargoylesoftware.htmlunit.html.HtmlPage;

import com.gargoylesoftware.htmlunit.html.HtmlSubmitInput;

import cn.smy.dama2.Dama2Web;

import cn.smy.dama2.Dama2Web.DecodeResult;

import cn.smy.dama2.Dama2Web.ReadBalanceResult;

/**

* @Title: main.java

* @Package

* @Description: TODO(用一句话描述该文件做什么)

* @author A18ccms A18ccms_gmail_com

* @date 2017年2月15日 下午3:42:00

* @version V1.0

*/

/**

 * @ClassName: main

 * @Description: TODO

 * @author zeze

 * @date 2017年2月15日 下午3:42:00

 *

 */

public class main {

    private static Logger logger = Logger.getLogger(main.class);

    private static final long serialVersionUID = 1325980466616825****;

    private static Dama2Web dama2 = new Dama2Web(46****, "41c5a58de6********d23b67f61645e3a7", "***", "****");

    private static int id;

    private static long nd = 1000 * 24 * 60 * 60;

    private static long nh = 1000 * 60 * 60;

    private static long nm = 1000 * 60;

    private static long ns = 1000;

    // 获得两个时间的毫秒时间差异

    private static Date nowDate;

    private static Date endDate;

    private static long diff;// getTime返回的是一个long型的毫秒数

    // 计算差多少分钟

    private static long min;

    // 计算差多少秒//输出结果

    private static long sec;

    // 计算多少毫秒

    private static long ms;

    public static void main(String[] agrs) {

        String emailAccount = "asd@qq.com";

        for (int i = 0; i < 10; i++) {

            nowDate = new Date();

            emailAccount = "asd" + i + "@qq.com";

            if(i==0)

                emailAccount="asd@qq.com";

            int statusCode=checkEbayAccount(emailAccount);

            if(statusCode==0){

                System.out.println(emailAccount + " 该邮箱号不是ebay账号");

            }else if(statusCode==1){

                System.out.println(emailAccount + " 该账号是eBay账号!");

            }else if(statusCode==101){

                System.out.println("打码错误!");

                statusCode=checkEbayAccount(emailAccount);

                while(statusCode==101){

                    statusCode=checkEbayAccount(emailAccount);

                }

            }else{

                System.out.println(statusCode);

            }

            endDate = new Date();

            diff = endDate.getTime() - nowDate.getTime();

            min = diff % nd % nh / nm;

            sec = diff % nd % nh % nm / ns;

            ms = diff % nd % nh % nm % ns;

            System.out.println(min + "分钟" + sec + "秒" + ms + "毫秒");

        }

    }

    // 验证邮箱是否为eBay账号

    public static int checkEbayAccount(String emailAccount) {

        System.out.println("开始验证账号：" + emailAccount);

        WebClient webClient = new WebClient(BrowserVersion.FIREFOX_17);

        HtmlPage page = null;

        try {

            page = webClient.getPage("http://fyp.ebay.com/");

        } catch (FailingHttpStatusCodeException e) {

            logger.error(e);

        } catch (MalformedURLException e) {

            logger.error(e);

        } catch (IOException e) {

            logger.error(e);

        }

        HtmlForm form = page.getForms().get(1);

        form.getInputByName("input").setValueAttribute(emailAccount);

        HtmlButton button = (HtmlButton) form.getElementsByTagName("button").get(0);

        try {

            page = button.click();

        } catch (IOException e1) {

            logger.error(e1);

        }

        if (page.asText().indexOf("Select how you want to reset your password") != -1) {

//            System.out.println(emailAccount + " 该账号是eBay账号!");

            return 1;

        }

        while (page.asText().indexOf("Security Measure") != -1) {

            Document doc = Jsoup.parse(page.asXml());

            Elements imgSrc = doc.getElementsByTag("iframe");

            String imgUrl = imgSrc.attr("src");

            System.out.println("验证码图片链接：" + imgUrl);

            String code = getCode(imgUrl);

            // 提交验证码

            form = page.getForms().get(0);

            form.getInputByName("tokenText").setValueAttribute(code);

            HtmlSubmitInput input = (HtmlSubmitInput) form.getElementsByTagName("input").get(5);

            try {

                page = input.click();

            } catch (IOException e1) {

                logger.error(e1);

            }

            if (page.asText().indexOf("the verification code you entered doesn't match against the image") != -1) {

//                System.out.println("打码错误！");

                dama2.reportError(id);

                return 101;

            }

            // 再次提交邮箱

            form = page.getForms().get(1);

            form.getInputByName("input").setValueAttribute(emailAccount);

            button = (HtmlButton) form.getElementsByTagName("button").get(0);

            try {

                page = button.click();

            } catch (IOException e1) {

                logger.error(e1);

            }

            if (page.asText().indexOf("Security Measure") != -1){// 如果还是验证码页面

                System.out.println("提交还是验证码页面!");

                continue;

            }

            if (page.asText().indexOf("Oops, that's not a match. Try again?") != -1) {

//                System.out.println(emailAccount + " 该邮箱号不是ebay账号");

                return 0;

            } else if (page.asText().indexOf("Select how you want to reset your password") != 1) {

//                System.out.println(emailAccount + " 该账号是eBay账号!");

                return 1;

            } else {

                System.out.println(page.asText());

                return 2;

            }

        }

        return 3;

    }

    // 打码兔获取验证码

    public static String getCode(String imgUrl) {

        // 打码兔

        int type = 6;

        int timeout = 30;

        ReadBalanceResult balanceResult = dama2.getBalance();

        // System.out.println(balanceResult);

        DecodeResult res = dama2.decodeUrlAndGetResult(imgUrl, type, timeout);

        String s;

        if (res.ret >= 0) {

            id = res.ret;

            s = "success: result=" + res.result + "; id=" + res.ret;

            System.out.println(s);

        } else {

            s = "failed: ret = " + res.ret + "; desc=" + res.desc;

            System.err.println(s);

        }

        return res.result;

    }

}

测试结果如下：

用多线程测试，明显快多了

package test;

import java.io.IOException;

import java.net.MalformedURLException;

import java.util.ArrayList;

import java.util.Date;

import java.util.concurrent.Callable;

import java.util.concurrent.ExecutorService;

import java.util.concurrent.Executors;

import java.util.concurrent.Future;

import org.apache.log4j.Logger;

import org.jsoup.Jsoup;

import org.jsoup.nodes.Document;

import org.jsoup.select.Elements;

import com.gargoylesoftware.htmlunit.BrowserVersion;

import com.gargoylesoftware.htmlunit.FailingHttpStatusCodeException;

import com.gargoylesoftware.htmlunit.WebClient;

import com.gargoylesoftware.htmlunit.html.HtmlButton;

import com.gargoylesoftware.htmlunit.html.HtmlForm;

import com.gargoylesoftware.htmlunit.html.HtmlPage;

import com.gargoylesoftware.htmlunit.html.HtmlSubmitInput;

import cn.smy.dama2.Dama2Web;

import cn.smy.dama2.Dama2Web.DecodeResult;

import cn.smy.dama2.Dama2Web.ReadBalanceResult;

/***

 *

 * @ClassName: EbayMultiplyThreadCheck

 * @Description: TODO

 * @author zeze

 * @date 2017年2月16日 上午8:49:46

 *

 */

public class EbayMultiplyThreadCheck {

    private static int threadNum = 30;

    private static long nd = 1000 * 24 * 60 * 60;

    private static long nh = 1000 * 60 * 60;

    private static long nm = 1000 * 60;

    private static long ns = 1000;

    private static Date nowDate;

    private static Date endDate;

    private static long diff;

    private static long min;

    private static long sec;

    private static long ms;

    public static void main(String[] args) {

        nowDate = new Date();

        ExecutorService exec = Executors.newFixedThreadPool(threadNum);

        ArrayList<Future<Integer>> results = new ArrayList<Future<Integer>>();

        for (int i = 0; i < threadNum; i++) {

            String email = "asd" + i + "@qq.com";

            if (i == 0)

                email = "asd@qq.com";

            results.add(exec.submit(new CheckEbayAccount(email)));

        }

        boolean isDone = false;

        while (!isDone) {

            isDone = true;

            for (Future<Integer> future : results) {

                if (!future.isDone()) {

                    isDone = false;

                    try {

                        Thread.sleep(1000);

                    } catch (InterruptedException e) {

                    }

                    break;

                }

            }

        }

        exec.shutdown();

        endDate = new Date();

        diff = endDate.getTime() - nowDate.getTime();

        min = diff % nd % nh / nm;

        sec = diff % nd % nh % nm / ns;

        ms = diff % nd % nh % nm % ns;

        System.out.println(min + "分钟" + sec + "秒" + ms + "毫秒");

    }

}

class CheckEbayAccount implements Callable<Integer> {

    private String email;

    private static Logger logger = Logger.getLogger(CheckEbayAccount.class);

    private static Dama2Web dama2 = new Dama2Web(****, "41c5a58de68ebe2*******", "***", "****");

    private static int id;

    public CheckEbayAccount(String email) {

        this.email = email;

    }

    @Override

    public Integer call() {

        System.out.println(Thread.currentThread().getName() + " 开始验证账号：" + email);

        WebClient webClient = new WebClient(BrowserVersion.FIREFOX_17);

        HtmlPage page = null;

        try {

            page = webClient.getPage("http://fyp.ebay.com/");

        } catch (FailingHttpStatusCodeException e) {

            logger.error(e);

        } catch (MalformedURLException e) {

            logger.error(e);

        } catch (IOException e) {

            logger.error(e);

        }

        HtmlForm form = page.getForms().get(1);

        form.getInputByName("input").setValueAttribute(email);

        HtmlButton button = (HtmlButton) form.getElementsByTagName("button").get(0);

        try {

            page = button.click();

        } catch (IOException e1) {

            logger.error(e1);

        }

        if (page.asText().indexOf("Select how you want to reset your password") != -1) {

            System.out.println(Thread.currentThread().getName() + " " + email + " 该账号是eBay账号!");

            return 1;

        } else if (page.asText().indexOf("Oops, that's not a match. Try again?") != -1) {

            System.out.println(Thread.currentThread().getName() + " " + email + " 该邮箱号不是ebay账号");

            return 0;

        }

        while (page.asText().indexOf("Security Measure") != -1) {

            Document doc = Jsoup.parse(page.asXml());

            Elements imgSrc = doc.getElementsByTag("iframe");

            String imgUrl = imgSrc.attr("src");

            System.out.println(Thread.currentThread().getName() + " " + "验证码图片链接：" + imgUrl);

            String code = getCode(imgUrl);

            // 提交验证码

            form = page.getForms().get(0);

            form.getInputByName("tokenText").setValueAttribute(code);

            HtmlSubmitInput input = (HtmlSubmitInput) form.getElementsByTagName("input").get(5);

            try {

                page = input.click();

            } catch (IOException e1) {

                System.out.println(Thread.currentThread().getName() + " " + e1);

            }

            while (page.asText().indexOf("Sorry") != -1) {

                System.out.println(Thread.currentThread().getName() + " 打码错误！重试");

                dama2.reportError(id);

                doc = Jsoup.parse(page.asXml());

                imgSrc = doc.getElementsByTag("iframe");

                imgUrl = imgSrc.attr("src");

                System.out.println(Thread.currentThread().getName() + " " + "验证码图片链接：" + imgUrl);

                code = getCode(imgUrl);

                // 提交验证码

                form = page.getForms().get(0);

                form.getInputByName("tokenText").setValueAttribute(code);

                input = (HtmlSubmitInput) form.getElementsByTagName("input").get(5);

                try {

                    page = input.click();

                } catch (IOException e1) {

                    logger.error(e1);

                }

            }

            // 再次提交邮箱

            form = page.getForms().get(1);

            form.getInputByName("input").setValueAttribute(email);

            button = (HtmlButton) form.getElementsByTagName("button").get(0);

            try {

                page = button.click();

            } catch (IOException e1) {

                logger.error(e1);

            }

            if (page.asText().indexOf("Security Measure") != -1) {// 如果还是验证码页面

                System.out.println(Thread.currentThread().getName() + " 提交还是验证码页面!");

                continue;

            }

            if (page.asText().indexOf("Oops, that's not a match. Try again?") != -1) {

                System.out.println(Thread.currentThread().getName() + " " + email + " 该邮箱号不是ebay账号");

                return 0;

            } else if (page.asText().indexOf("Select how you want to reset your password") != 1) {

                System.out.println(Thread.currentThread().getName() + " " + email + " 该账号是eBay账号!");

                return 1;

            } else {

                System.out.println(Thread.currentThread().getName() + " " + page.asText());

                return 2;

            }

        }

        System.out.println(Thread.currentThread().getName() + " " + page.asText());

        return 3;

    }

    // 打码兔获取验证码

    public static String getCode(String imgUrl) {

        // 打码兔

        int type = 6;

        int timeout = 30;

        ReadBalanceResult balanceResult = dama2.getBalance();

        // System.out.println(balanceResult);

        DecodeResult res = dama2.decodeUrlAndGetResult(imgUrl, type, timeout);

        String s;

        if (res.ret >= 0) {

            id = res.ret;

            s = "[打码结果=" + res.result + "] [id=" + res.ret + "] " + balanceResult;

            System.out.println(Thread.currentThread().getName() + " " + s);

        } else {

            while (res.result == null) {

                s = "打码失败,重试: ret = " + res.ret + "; desc=" + res.desc;

                System.out.println(Thread.currentThread().getName() + " " + s);

                dama2.reportError(id);

                res = dama2.decodeUrlAndGetResult(imgUrl, type, timeout);

                if (res.ret >= 0) {

                    id = res.ret;

                    s = "[打码结果=" + res.result + "] [id=" + res.ret + "] " + balanceResult;

                    System.out.println(Thread.currentThread().getName() + " " + s);

                }

            }

        }

        return res.result;

    }

}

测试30个账号，平均每个3秒

最新文章

  1. openerp7 时区问题
  2. Quick Apps for Sharepoint小型BI解决方案
  3. mysql max_allowed_packet查询和修改
  4. PAT (Basic Level) Practise：1031. 查验身份证
  5. JAVA基础学习day20--IO流二-缓冲流、字节流
  6. 使用Spring Aop验证方法参数是否合法
  7. request.getParameter与request.getAttribute()
  8. 求两个数的最大公约数（Euclid算法）
  9. Android自定义Toast宽度无法设置问题解决
  10. Storm并发度和Grouping方式
  11. 【重学计算机】操作系统D6章：并发程序设计
  12. vi命令设置行号
  13. ue4 StringTable
  14. 记一下vue.js事件的修饰等问题
  15. cmder使用简介
  16. 潭州课堂25班：Ph201805201 django 项目 第四十四课 项目部署 （课堂笔记）
  17. 补充照片：某基同学使用Bing词典
  18. 用图来教你怎样用Photoshop蓝底转换红底
  19. whith ~ as 用法
  20. 利用crontab每天定时备份MySQL数据库

热门文章

  1. java 实现文本格式转换
  2. 一段测试try...catch运行时间的代码
  3. 初学者的React全家桶完整实例
  4. document.body、document.documentElement和window获取视窗大小的差别
  5. 架构-LAMP特级学习(网站服务器监控)
  6. 重写lucene.net的分词器支持3.0.3.0版本
  7. pylot 学习笔记
  8. mysql-5.7 group commit 详解
  9. python中if __name__ == &#39;__main__&#39;的说明
  10. How To Setup Apache Hadoop On CentOS