zhangyong 4 月之前
父節點
當前提交
8c92a6f6bd

+ 7 - 0
etl-core/src/main/java/com/tzld/crawler/etl/service/impl/EtlServiceImpl.java

@@ -451,6 +451,13 @@ public class EtlServiceImpl implements EtlService {
                                         , !CollectionUtils.isEmpty(proxyPlatform) && proxyPlatform.contains(platfrm),
                                         proxyInfo);
                             }
+                            // 乐跑乐动
+                            else if (platform.equals("lepaoledongdijie") || platform.equals("lepaoledong")) {
+                                FileUtils.downloadLPLD(fileUrl, fpath,
+                                        !CollectionUtils.isEmpty(randomUaPlatform) && randomUaPlatform.contains(platfrm)
+                                        , !CollectionUtils.isEmpty(proxyPlatform) && proxyPlatform.contains(platfrm),
+                                        proxyInfo);
+                            }
                             else {
                                 String tempUrl = fileUrl;
                                 if (platform.equals("zhufuzhonglaonianrenruyijixiang")) {

+ 55 - 0
etl-core/src/main/java/com/tzld/crawler/etl/util/FileUtils.java

@@ -156,6 +156,61 @@ public class FileUtils {
         log.info("downloaded successfully [{}] to [{}]", fileUrl, filePath);
     }
 
+
+    public static void downloadLPLD(String fileUrl, String filePath, boolean useUa, boolean useProxy, Map<String, String> proxyInfo) throws Exception {
+        log.info("begin download [{}] to [{}] useUa [{}] useProxy [{}] proxyInfo[{}]", fileUrl, filePath, useUa, useProxy, proxyInfo);
+        URL url = new URL(fileUrl);
+        HttpURLConnection conn;
+        if (useProxy) {
+            String proxyUrl = proxyInfo.getOrDefault("url", "");
+            int port = Integer.parseInt(proxyInfo.getOrDefault("port", "0"));
+            String username = proxyInfo.getOrDefault("username", "");
+            String password = proxyInfo.getOrDefault("password", "");
+            // 创建代理服务的地址和端口
+            Proxy proxy = new Proxy(Proxy.Type.HTTP, new InetSocketAddress(proxyUrl, port));
+            Authenticator authenticator = new Authenticator() {
+                @Override
+                public PasswordAuthentication getPasswordAuthentication() {
+                    return (new PasswordAuthentication(username, password.toCharArray()));
+                }
+            };
+            Authenticator.setDefault(authenticator);
+            conn = (HttpURLConnection) url.openConnection(proxy);
+        } else {
+            conn = (HttpURLConnection) url.openConnection();
+        }
+        // 设置请求头, 添加了防盗链 referer
+        conn.setRequestProperty("Accept", "*/*");
+        conn.setRequestProperty("Accept-Encoding", "identity;q=1, *;q=0");
+        conn.setRequestProperty("Accept-Language", "zh-CN,zh;q=0.9");
+        conn.setRequestProperty("Range", "bytes=0-");
+
+//        conn.setRequestProperty("Host", "mpvideo.qpic.cn");
+//        conn.setRequestProperty("Origin", "https://mp.weixin.qq.com");
+//        conn.setRequestProperty("Referer", "https://mp.weixin.qq.com/");
+
+        if (useUa) {
+            conn.setRequestProperty("User-Agent", FakeUserAgent.getRandomUserAgent());
+        }
+        if (conn.getResponseCode() == HttpURLConnection.HTTP_FORBIDDEN) {
+            throw new CommonException(ExceptionEnum.URL_FORBIDDEN);
+        }
+        conn.setConnectTimeout(5000);
+        conn.setReadTimeout(5000);
+        log.info("download file size is {} of url [{}]", formatFileSize(conn.getContentLength()), fileUrl);
+
+        InputStream inputStream = conn.getInputStream();
+        FileOutputStream outputStream = new FileOutputStream(filePath);
+        byte[] buffer = new byte[4096];
+        int len;
+        while ((len = inputStream.read(buffer)) != -1) {
+            outputStream.write(buffer, 0, len);
+        }
+        inputStream.close();
+        outputStream.close();
+        log.info("downloaded successfully [{}] to [{}]", fileUrl, filePath);
+    }
+
     public static void downloadDY(String fileUrl, String filePath, boolean useUa, boolean useProxy, Map<String, String> proxyInfo) throws Exception {
         log.info("begin download [{}] to [{}] useUa [{}] useProxy [{}] proxyInfo[{}]", fileUrl, filePath, useUa, useProxy, proxyInfo);
         URL url = new URL(fileUrl);