com.waku.mmdataextract.IMEI.java Source code

Introduction

Here is the source code for com.waku.mmdataextract.IMEI.java

Source

/*
 * IMEI.java
 * Created on 2011-2-27; Project to Colt2010; $Id: IMEI.java 309 2013-04-25 16:38:44Z tristan $
 * 
 * Copyright (c) 2011, Xu Brothers and/or its affiliates. All rights reserved.
 * Xu Brothers PROPRIETARY/CONFIDENTIAL. Use is subject to license terms.
 */

package com.waku.mmdataextract;

import java.io.File;
import java.io.FileWriter;
import java.io.InputStream;
import java.util.List;

import org.dom4j.Document;
import org.dom4j.Element;
import org.dom4j.io.SAXReader;
import org.openqa.selenium.By;
import org.openqa.selenium.WebDriver;
import org.openqa.selenium.WebElement;
import org.openqa.selenium.htmlunit.HtmlUnitDriver;

import com.waku.common.http.HttpCaller;

/**
 * @versin $Rev: 309 $, $Date: 2013-04-26 00:38:44 +0800 (, 26  2013) $
 * @author Jin
 */

public class IMEI {

    private static final String PRODUCT_DETAIL_URL = "http://shouji.gd.chinamobile.com/gdmobile/productDetail/";
    private static final String SEARCH_ACTION_PER_BRD = "searchImeiAction.do?flag=typebrand&typeid=1&brandid=";
    private static final String SEARCH_ACTION_PER_PAGE = "searchImeiAction.do?flag=brand&currentPage=";

    private static String urlFor(int i, String searchvalue) {
        return PRODUCT_DETAIL_URL + SEARCH_ACTION_PER_PAGE + i + "&searchvalue=" + searchvalue;
    }

    private static String urlFor(String bId) {
        return PRODUCT_DETAIL_URL + SEARCH_ACTION_PER_BRD + bId;
    }

    @SuppressWarnings("unchecked")
    public static void main(String[] args) throws Exception {
        WebDriver driver = new HtmlUnitDriver();
        InputStream in = Thread.currentThread().getContextClassLoader().getResourceAsStream("Brand.xml");
        List<Element> brands = new SAXReader().read(in).selectNodes("/brand/option");
        for (Element brand : brands) {
            String bId = brand.attributeValue("value");
            FileWriter fw = new FileWriter(new File("output/" + brand.getText() + ".csv"));
            try {
                Document doc = HttpCaller.httpGetAsXMLDoc(urlFor(bId));
                List<Element> products = doc.selectNodes("/products/product");
                for (Element e : products) {
                    String qString = e.attributeValue("id");
                    driver.get(urlFor(1, qString));
                    int pageNumber = Integer
                            .parseInt(driver.findElements(By.xpath("//span[@class='font_b2']")).get(1).getText());
                    if (pageNumber >= 1) {
                        for (int i = 1; i <= pageNumber; i++) {
                            readForEachPage(fw, driver, urlFor(i, qString));
                        }
                    }
                }
                fw.close();
            } catch (RuntimeException e) {
                System.out.println("Skipped " + brand.getText());
                fw.write("Skipped!");
                fw.close();
            }
        }
    }

    private static void readForEachPage(FileWriter fw, WebDriver driver, String url) throws Exception {
        try {
            driver.get(url);
            WebElement table = driver.findElement(By.xpath("//table[@bgcolor='BFBEC3']"));
            List<WebElement> items = table.findElements(By.xpath("./tbody/tr"));
            for (int i = 1; i < items.size() - 1; i++) {
                List<WebElement> cols = items.get(i).findElements(By.tagName("td"));
                for (WebElement col : cols) {
                    System.out.print(col.getText() + ",");
                    fw.write(col.getText() + ",");
                }
                System.out.println();
                fw.write("\n");
            }
        } catch (Exception e) {
            System.out.println("---------Retry--------------------------");
            System.out.println(e.getMessage());
            System.out.println("---------Retry---------------------------");
            readForEachPage(fw, driver, url);
        }
    }

}