selenium 实现爬虫获取数据

selenium 是一款比较好用,模拟人工操作浏览器的一款工具。本是做自动化测试的,但是由于项目中要进行项目的爬取,但是由于项目的复杂度有高,本人采用selenim 进行自动化获取数据。

import org.openqa.selenium.By;
import org.openqa.selenium.WebDriver;
import org.openqa.selenium.WebElement;
import org.openqa.selenium.chrome.ChromeDriver;
import org.openqa.selenium.chrome.ChromeOptions;

import java.util.Arrays;
import java.util.List;
import java.util.concurrent.TimeUnit;

public class PPs {

    public static void main(String[] args) throws InterruptedException {
        String[] data =  {"党参","三七","黄连"};
        ChromeOptions chromeOptions = new ChromeOptions();
//        chromeOptions.addArguments("--headless","--no-sandbox","--disable-gpu","--window-size=1290,1080");
        chromeOptions.addArguments("--headless");
        System.setProperty("webdriver.chrome.driver", "C:\Program Files\Google\Chrome\Application\chromedriver.exe");
        WebDriver driver = new ChromeDriver();
        driver.manage().timeouts().implicitlyWait(10, TimeUnit.SECONDS);


        List<String> strList = Arrays.asList(data);
        for (String s: strList) {
            getData(driver,s);
        }
        driver.close();
    }

    private static void getData(WebDriver driver,String data) throws InterruptedException {
        driver.get("https://www.yt1998.com");
        WebElement top_search = driver.findElement(new By.ById("fltext"));
        String typeText = top_search.getText();
        if(!"今日价格".equals(typeText)){top_search.click();
             WebElement type = driver.findElement(new By.ByXPath("//*[@id="fltype"]/ul/li[1]/a"));
             type.click();
        }
        WebElement kw = driver.findElement(new By.ByXPath("//*[@id="kw"]"));
        kw.sendKeys(data);
        WebElement su = driver.findElement(new By.ByXPath("//*[@id="su"]"));
        su.click();

        WebElement dd = driver.findElement(new By.ByXPath("/html/body/div[3]/div/div[2]/dl[1]/dd"));
        List<WebElement> aList = dd.findElements(new By.ByTagName("a"));
        for (int i = 2; i <= aList.size(); i++) {
            WebElement webElement = driver.findElement(new By.ByXPath("/html/body/div[3]/div/div[2]/dl[1]/dd/a["+i+"]"));
          //  System.out.println(webElement.getText());
            webElement.click();
            extracted(driver);
//            Thread.sleep(1000);
        }
    }

    /**
     * 获取当前的市场的数据
     * @param driver
     */
    private static void extracted(WebDriver driver) {
        WebElement priceList = driver.findElement(new By.ById("priceList"));
        List<WebElement> trList = priceList.findElements(new By.ByTagName("tr"));
        for (WebElement tr :trList) {
            List<WebElement> tdList = tr.findElements(new By.ByTagName("td"));
            String font1 = tdList.get(0).findElement(new By.ByTagName("font")).getText();
            String font2 = tdList.get(1).findElement(new By.ByTagName("font")).getText();
            String font3 = tdList.get(2).findElement(new By.ByTagName("font")).getText();
            String font4 = tdList.get(3).findElement(new By.ByTagName("font")).getText();
            String font5 = tdList.get(4).findElement(new By.ByTagName("font")).getText();
            String font6 = tdList.get(5).findElement(new By.ByTagName("font")).getText();
            String font7 = tdList.get(6).findElement(new By.ByTagName("font")).getText();
            String font8 = tdList.get(7).findElement(new By.ByTagName("font")).getText();
            String font9 = tdList.get(8).findElement(new By.ByTagName("font")).getText();
            String font10 = tdList.get(9).findElement(new By.ByTagName("font")).getText();
            String font11 = tdList.get(10).findElement(new By.ByTagName("font")).getText();
            System.out.print(font1+"	");
            System.out.print(font2+"	");
            System.out.print(font3+"	");
            System.out.print(font4+"	");
            System.out.print(font5+"	");
            System.out.print(font6+"	");
            System.out.print(font7+"	");
            System.out.print(font8+"	");
            System.out.print(font9+"	");
            System.out.print(font10+"	");
            System.out.println(font11+"	");
        }
    }
}
经验分享 程序员 微信小程序 职场和发展