Skip to content

M
meta_crawler
提交 e7ad390c 作者: 张文库
操作

谷歌浏览器驱动复用

上级 a1b331fb
正在显示 包含 126 行增加 和 86 行删除
comm_crawler/src/main/java/com/zzsn/crawler/ChromeExecutor.java
package com.zzsn.crawler;
import org.openqa.selenium.NoSuchSessionException;
import org.openqa.selenium.SessionNotCreatedException;
import lombok.extern.slf4j.Slf4j;
import org.openqa.selenium.UnsupportedCommandException;
import org.openqa.selenium.WebDriverException;
import org.openqa.selenium.remote.*;
import org.openqa.selenium.remote.codec.w3c.W3CHttpCommandCodec;
import org.openqa.selenium.remote.codec.w3c.W3CHttpResponseCodec;
import org.openqa.selenium.remote.http.*;
import org.openqa.selenium.remote.http.HttpClient;
import org.openqa.selenium.remote.http.HttpRequest;
import org.openqa.selenium.remote.http.HttpResponse;
import java.io.IOException;
import java.net.URL;
......@@ -19,6 +19,7 @@ import static org.openqa.selenium.remote.DriverCommand.*;
* @author kethy
* @date 2022-07-23 18:33:50
*/
@Slf4j
public class ChromeExecutor extends HttpCommandExecutor {
private CommandCodec<HttpRequest> commandCodec;
private ResponseCodec<HttpResponse> responseCodec;
......@@ -39,17 +40,19 @@ public class ChromeExecutor extends HttpCommandExecutor {
public Response execute(Command command) throws IOException {
if (command.getSessionId() == null) {
if (QUIT.equals(command.getName())) {
return new Response();
return null;
}
if (!GET_ALL_SESSIONS.equals(command.getName()) && !NEW_SESSION.equals(command.getName())) {
throw new NoSuchSessionException("会话ID为空，请调用quit()退出后再使用驱动");
log.error("会话ID为空，请调用退出后再使用驱动...");
return null;
}
}
if (NEW_SESSION.equals(command.getName())) {
if (commandCodec != null) {
throw new SessionNotCreatedException("Session 已存在..");
log.error("Session 未创建...");
return null;
}
ProtocolHandshake handshake = new ProtocolHandshake();
......@@ -62,7 +65,8 @@ public class ChromeExecutor extends HttpCommandExecutor {
}
if (commandCodec == null || responseCodec == null) {
throw new WebDriverException("未定义命令或解码器。无法继续..");
log.error("未定义命令或解码器。无法继续...");
return null;
}
HttpRequest httpRequest = commandCodec.encode(command);
......@@ -81,10 +85,9 @@ public class ChromeExecutor extends HttpCommandExecutor {
return response;
} catch (UnsupportedCommandException e) {
if (e.getMessage() == null || "".equals(e.getMessage())) {
throw new UnsupportedOperationException("未接收到服务器信息。命令：" + command.getName(), e.getCause());
log.error("未接收到服务器返回信息。命令：" + command.getName());
}
throw e;
return null;
}
}
}
\ No newline at end of file
comm_crawler/src/main/java/com/zzsn/crawler/ReuseWebDriver.java
package com.zzsn.crawler;
import com.google.common.collect.ImmutableMap;
import lombok.extern.slf4j.Slf4j;
import org.openqa.selenium.Capabilities;
import org.openqa.selenium.MutableCapabilities;
import org.openqa.selenium.WebDriverException;
import org.openqa.selenium.net.UrlChecker;
import org.openqa.selenium.remote.Command;
import org.openqa.selenium.remote.DriverCommand;
import org.openqa.selenium.remote.RemoteWebDriver;
import org.openqa.selenium.remote.Response;
import org.openqa.selenium.remote.internal.WebElementToJsonConverter;
import java.io.IOException;
......@@ -19,44 +20,30 @@ import java.util.concurrent.TimeUnit;
import java.util.stream.Collectors;
import java.util.stream.Stream;
import static org.openqa.selenium.remote.CapabilityType.SUPPORTS_JAVASCRIPT;
/***
* 驱动复用
* @date 2022/7/23 18:18
* @date 2022/7/23 18:01
* @author andylau
*/
@Slf4j
public class ReuseWebDriver extends RemoteWebDriver {
private Capabilities capabilities;
private String serverUrl;
public ReuseWebDriver(String serverUrl, String sessionId) throws IOException {
super(new ChromeExecutor(new URL(serverUrl)), null);
this(serverUrl);
super.setSessionId(sessionId);
this.serverUrl = serverUrl;
connectTest(sessionId);
}
@Override
protected void startSession(Capabilities capabilities) {
private ReuseWebDriver(String serverUrl) throws MalformedURLException {
super(new ChromeExecutor(new URL(serverUrl)), new MutableCapabilities());
}
@Override
public Capabilities getCapabilities() {
return capabilities;
}
protected void startSession(Capabilities capabilities) {
/**
* 判断浏览器是否支持运行JS脚本
*
* @author andylau
* @date 2022/7/24 10:08
*/
private boolean isJavascriptDisabled() {
return !capabilities.is(SUPPORTS_JAVASCRIPT);
}
/***
......@@ -66,9 +53,6 @@ public class ReuseWebDriver extends RemoteWebDriver {
*/
@Override
public Object executeScript(String script, Object... args) {
if (isJavascriptDisabled()) {
throw new UnsupportedOperationException("必须使用支持执行javascript的WebDriver的底层实例..");
}
// 替换引号
script = script.replaceAll("\"", "\\\"");
List<Object> convertedArgs = Stream.of(args).map(new WebElementToJsonConverter()).collect(Collectors.toList());
......@@ -79,13 +63,10 @@ public class ReuseWebDriver extends RemoteWebDriver {
/***
* 再打开的浏览器异步执行JS脚本
* @author andylau
* @date 2022/7/24 10:06
* @date 2022/7/23 18:06
*/
@Override
public Object executeAsyncScript(String script, Object... args) {
if (isJavascriptDisabled()) {
throw new UnsupportedOperationException("必须使用支持执行javascript的WebDriver的底层实例..");
}
script = script.replaceAll("\"", "\\\"");
List<Object> convertedArgs = Stream.of(args).map(new WebElementToJsonConverter()).collect(Collectors.toList());
Map<String, ?> params = ImmutableMap.of("script", script, "args", convertedArgs);
......@@ -93,37 +74,51 @@ public class ReuseWebDriver extends RemoteWebDriver {
}
/***
* 连接测试
* session 连接测试
* @author andylau
* @date 2022/7/24 10:03
* @date 2022/7/23 18:03
*/
private void connectTest(String sessionId) throws IOException {
if (!sessionId.isEmpty()) {
super.setSessionId(sessionId);
private boolean connectTest() {
Command command = new Command(super.getSessionId(), DriverCommand.STATUS);
try {
Response response = getCommandExecutor().execute(command);
if (response != null && 0 == response.getStatus()) {
return true;
}
} catch (Exception e) {
log.error("Session 连接失败...");
}
return false;
}
Command command = new Command(super.getSessionId(), DriverCommand.STATUS);
getCommandExecutor().execute(command);
this.capabilities = new MutableCapabilities();
/***
* 判断是否连接失败并退出
* @author andylau
* @date 2022/7/25 14:25
*/
public boolean connectTestFail() {
boolean flag = !this.connectTest();
// 失败时退出服务
if (flag) {
try {
this.quit();
} catch (Exception e) {
// do nothing
}
}
return flag;
}
@Override
public void quit() {
super.quit();
// 关闭DriverService，避免无法关闭DriverService，导致出现过多Driver进程。
WebDriverException throwe = null;
try {
URL killUrl = new URL(serverUrl + "/shutdown");
new UrlChecker().waitUntilUnavailable(3, TimeUnit.SECONDS, killUrl);
} catch (MalformedURLException e) {
throwe = new WebDriverException(e);
} catch (UrlChecker.TimeoutException e) {
throwe = new WebDriverException("等待驱动程序服务关闭时超时..", e);
}
if (throwe != null) {
throw throwe;
} catch (Exception e) {
log.error("驱动退出异常...");
}
}
}
comm_crawler/src/main/java/com/zzsn/test/ChromeTest.java
package com.zzsn.test;
import com.alibaba.fastjson.JSON;
import com.zzsn.crawler.ChromeDriverPool;
import com.zzsn.crawler.ReuseWebDriver;
import com.zzsn.generation.Constants;
import com.zzsn.job.JedisUtil;
import com.zzsn.util.DriverUtil;
import lombok.extern.slf4j.Slf4j;
import org.openqa.selenium.chrome.ChromeDriver;
import org.openqa.selenium.chrome.ChromeDriverService;
import org.openqa.selenium.remote.HttpCommandExecutor;
import org.openqa.selenium.remote.SessionId;
import java.io.IOException;
import java.util.HashMap;
import java.util.Map;
import java.util.concurrent.CountDownLatch;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;
@Slf4j
public class ChromeTest {
// public static void main(String[] args) throws Exception {
// //模拟并发
......@@ -87,33 +96,17 @@ public class ChromeTest {
// latch.countDown();//开炮
// }
public static void main(String[] args) {
// 初始化一个chrome浏览器实例
ReuseWebDriver driver = null;
try {
driver = new ReuseWebDriver("http://localhost:63714", "fdaa2ca38a077a604dffdca6ecc5df1d");
} catch (Exception e) {
System.out.println("Session连接失败，重新打开浏览器驱动...");
ChromeDriver chromeDriver = DriverUtil.reconnectDriver();
String serverUrl = DriverUtil.getServerUrl(chromeDriver);
String sessionId = chromeDriver.getSessionId().toString();
try {
driver = new ReuseWebDriver(serverUrl, sessionId);
} catch (IOException e1) {
System.out.println("获取驱动连接失败！！！");
// 可复用驱动使用Demo
public static void main(String[] args) throws Exception {
ReuseWebDriver driver = DriverUtil.getChromeDriver();
if (driver == null) {
// 从缓存取出SessionId为空才时，驱动会返回null，可参考工具类重新设置缓存
log.error("获取浏览器驱动失败，请检查SessionId缓存是否存在...");
return;
}
}
// 最大化窗口
// driver.manage().window().maximize();
// 设置隐性等待时间
// driver.manage().timeouts().implicitlyWait(3, TimeUnit.SECONDS);
// get()打开一个站点
try {
// 测试打开bing
driver.get("https://www.bing.com");
// getTitle()获取当前页面title的值
System.out.println("当前打开页面的标题是： " + driver.getTitle());
......@@ -123,8 +116,14 @@ public class ChromeTest {
System.out.println(((HttpCommandExecutor) driver.getCommandExecutor()).getAddressOfRemoteServer());
driver.executeScript("alert(\"hello,this is an alert!\")");
// 关闭并退出浏览器
// driver.quit();
} catch (Exception e) {
// 驱动突然崩溃（可手动关闭驱动触发该场景）， response在短时间内会返回正常请求码从而导致
// driver.get("https://www.bing.com") 获取页面信息异常
// 这里直接退出，不在重新获取，下次获取链接时在重新启动驱动服务
log.error("获取浏览器驱动异常，驱动重启中...");
driver.quit();
}
}
}
comm_crawler/src/main/java/com/zzsn/util/DriverUtil.java
package com.zzsn.util;
import com.alibaba.fastjson.JSON;
import com.zzsn.crawler.ReuseWebDriver;
import com.zzsn.generation.Constants;
import com.zzsn.job.JedisUtil;
import lombok.extern.slf4j.Slf4j;
import org.openqa.selenium.chrome.ChromeDriver;
import org.openqa.selenium.chrome.ChromeDriverService;
import org.openqa.selenium.remote.HttpCommandExecutor;
import java.net.URL;
import java.util.HashMap;
import java.util.Map;
/**
* @author andylau
* @version 1.0
* @date 2022/7/23 17:14
**/
@Slf4j
public class DriverUtil {
/***
......@@ -19,7 +26,7 @@ public class DriverUtil {
* @author andylau
* @date 2022/7/23 17:15
*/
public static ChromeDriver reconnectDriver() {
private static ChromeDriver reconnectDriver() {
System.setProperty("webdriver.chrome.driver", Constants.CHROMEDRIVE);
ChromeDriverService service = ChromeDriverService.createDefaultService();
// 重新初始化一个chrome浏览器实例
......@@ -31,9 +38,46 @@ public class DriverUtil {
* @author andylau
* @date 2022/7/23 17:15
*/
public static String getServerUrl(ChromeDriver driver) {
private static String getServerUrl(ChromeDriver driver) {
HttpCommandExecutor commandExecutor = (HttpCommandExecutor) driver.getCommandExecutor();
URL server = commandExecutor.getAddressOfRemoteServer();
return server.toString();
}
@SuppressWarnings("all")
public static ReuseWebDriver connectChrome(String sessionId, String serverUrl) throws Exception {
if (serverUrl == null || "".equals(serverUrl) || sessionId == null || "".equals(sessionId)) {
log.error("未获取到驱动服务地址、sessionId");
return null;
}
ReuseWebDriver driver = new ReuseWebDriver(serverUrl, sessionId);
if (driver.connectTestFail()) {
// 若驱动返回错误码，重新创建驱动服务并缓存
ChromeDriver chromeDriver = DriverUtil.reconnectDriver();
serverUrl = DriverUtil.getServerUrl(chromeDriver);
sessionId = chromeDriver.getSessionId().toString();
Map<String, String> map = new HashMap<>(2);
map.put("sessionId", sessionId);
map.put("serverUrl", serverUrl);
// 缓存浏览器驱动信息
JedisUtil.setString("SELENIUM_DRIVER_CACHE", JSON.toJSONString(map), -1);
driver = new ReuseWebDriver(serverUrl, sessionId);
}
return driver;
}
/***
* 获取当前活动的谷歌浏览器驱动
* @author andylau
* @date 2022/7/25 15:07
*/
@SuppressWarnings("all")
public static ReuseWebDriver getChromeDriver() throws Exception {
String cacheInfo = JedisUtil.getString("SELENIUM_DRIVER_CACHE");
Map<String, String> map = JSON.parseObject(cacheInfo, Map.class);
String sessionId = map.get("sessionId");
String serverUrl = map.get("serverUrl");
return connectChrome(sessionId, serverUrl);
}
}
Markdown 格式
0% 或
您添加了 0 人 到此讨论。请谨慎行事。
请先完成此评论的编辑！
请 注册 或者 后发表评论