Java tutorial
/* * Copyright 2010-2013 Amazon.com, Inc. or its affiliates. All Rights Reserved. * * Portions copyright 2006-2009 James Murty. Please see LICENSE.txt * for applicable license terms and NOTICE.txt for applicable notices. * * Licensed under the Apache License, Version 2.0 (the "License"). * You may not use this file except in compliance with the License. * A copy of the License is located at * * http://aws.amazon.com/apache2.0 * * or in the "license" file accompanying this file. This file is distributed * on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either * express or implied. See the License for the specific language governing * permissions and limitations under the License. */ package com.sina.scs; import java.io.BufferedOutputStream; import java.io.File; import java.io.FileInputStream; import java.io.FileOutputStream; import java.io.IOException; import java.io.OutputStream; import java.io.UnsupportedEncodingException; import java.net.MalformedURLException; import java.net.SocketException; import java.net.URL; import java.text.ParseException; import java.util.Arrays; import java.util.Date; import java.util.List; import javax.net.ssl.SSLProtocolException; import org.apache.commons.logging.Log; import org.apache.commons.logging.LogFactory; import com.sina.Request; import com.sina.SCSClientException; import com.sina.scs.model.S3Object; import com.sina.util.BinaryUtils; import com.sina.util.DateUtils; import com.sina.util.HttpUtils; import com.sina.util.Md5Utils; /** * General utility methods used throughout the AWS S3 Java client. */ public class ServiceUtils { private static final Log log = LogFactory.getLog(ServiceUtils.class); protected static final DateUtils dateUtils = new DateUtils(); public static Date parseIso8601Date(String dateString) throws ParseException { return dateUtils.parseIso8601Date(dateString); } public static String formatIso8601Date(Date date) { return dateUtils.formatIso8601Date(date); } public static Date parseRfc822Date(String dateString) throws ParseException { return dateUtils.parseRfc822Date(dateString); } public static String formatRfc822Date(Date date) { return dateUtils.formatRfc822Date(date); } /** * Returns true if the specified ETag was from a multipart upload. * * @param eTag * The ETag to test. * * @return True if the specified ETag was from a multipart upload, otherwise * false it if belongs to an object that was uploaded in a single * part. */ public static boolean isMultipartUploadETag(String eTag) { return eTag.contains("-"); } /** * Safely converts a string to a byte array, first attempting to explicitly * use our preferred encoding (UTF-8), and then falling back to the * platform's default encoding if for some reason our preferred encoding * isn't supported. * * @param s * The string to convert to a byte array. * * @return The byte array contents of the specified string. */ public static byte[] toByteArray(String s) { try { return s.getBytes(Constants.DEFAULT_ENCODING); } catch (UnsupportedEncodingException e) { log.warn("Encoding " + Constants.DEFAULT_ENCODING + " is not supported", e); return s.getBytes(); } } /** * Removes any surrounding quotes from the specified string and returns a * new string. * * @param s * The string to check for surrounding quotes. * * @return A new string created from the specified string, minus any * surrounding quotes. */ public static String removeQuotes(String s) { if (s == null) return null; s = s.trim(); if (s.startsWith("\"")) s = s.substring(1); if (s.endsWith("\"")) s = s.substring(0, s.length() - 1); return s; } /** * Converts the specified request object into a URL, containing all the * specified parameters, the specified request endpoint, etc. * * @param request * The request to convert into a URL. * @return A new URL representing the specified request. * * @throws AmazonClientException * If the request cannot be converted to a well formed URL. */ public static URL convertRequestToUrl(Request<?> request) { // To be backward compatible, this method by default does not // remove the leading slash in the request resource-path. return convertRequestToUrl(request, false); } /** * Converts the specified request object into a URL, containing all the * specified parameters, the specified request endpoint, etc. * * @param request * The request to convert into a URL. * @param removeLeadingSlashInResourcePath * Whether the leading slash in resource-path should be removed * before appending to the endpoint. * @return A new URL representing the specified request. * * @throws AmazonClientException * If the request cannot be converted to a well formed URL. */ public static URL convertRequestToUrl(Request<?> request, boolean removeLeadingSlashInResourcePath) { String resourcePath = HttpUtils.urlEncode(request.getResourcePath(), true); // Removed the padding "/" that was already added into the request's resource path. if (removeLeadingSlashInResourcePath && resourcePath.startsWith("/")) { resourcePath = resourcePath.substring(1); } // Some http client libraries (e.g. Apache HttpClient) cannot handle // consecutive "/"s between URL authority and path components. // So we escape "////..." into "/%2F%2F%2F...", in the same way as how // we treat consecutive "/"s in AmazonS3Client#presignRequest(...) String urlPath = "/" + resourcePath; urlPath = urlPath.replaceAll("(?<=/)/", "%2F"); String urlString = request.getEndpoint() + urlPath; boolean firstParam = true; for (String param : request.getParameters().keySet()) { if (firstParam) { urlString += "?"; firstParam = false; } else { urlString += "&"; } String value = request.getParameters().get(param); urlString += param + "=" + HttpUtils.urlEncode(value, false); } try { return new URL(urlString); } catch (MalformedURLException e) { throw new SCSClientException("Unable to convert request to well formed URL: " + e.getMessage(), e); } } /** * Returns a new string created by joining each of the strings in the * specified list together, with a comma between them. * * @param strings * The list of strings to join into a single, comma delimited * string list. * @return A new string created by joining each of the strings in the * specified list together, with a comma between strings. */ public static String join(List<String> strings) { String result = ""; boolean first = true; for (String s : strings) { if (!first) result += ", "; result += s; first = false; } return result; } /** * Downloads an S3Object, as returned from * {@link SCSClient#getObject(com.amazonaws.services.s3.model.GetObjectRequest)}, * to the specified file. * * @param s3Object * The S3Object containing a reference to an InputStream * containing the object's data. * @param destinationFile * The file to store the object's data in. * @param performIntegrityCheck * Boolean valuable to indicate whether do the integrity check or not * */ public static void downloadObjectToFile(S3Object s3Object, File destinationFile, boolean performIntegrityCheck) { // attempt to create the parent if it doesn't exist File parentDirectory = destinationFile.getParentFile(); if (parentDirectory != null && !parentDirectory.exists()) { parentDirectory.mkdirs(); } OutputStream outputStream = null; try { outputStream = new BufferedOutputStream(new FileOutputStream(destinationFile)); byte[] buffer = new byte[1024 * 10]; int bytesRead; while ((bytesRead = s3Object.getObjectContent().read(buffer)) > -1) { outputStream.write(buffer, 0, bytesRead); } } catch (IOException e) { try { s3Object.getObjectContent().abort(); } catch (IOException abortException) { log.warn("Couldn't abort stream", e); } throw new SCSClientException("Unable to store object contents to disk: " + e.getMessage(), e); } finally { try { outputStream.close(); } catch (Exception e) { } try { s3Object.getObjectContent().close(); } catch (Exception e) { } } byte[] clientSideHash = null; byte[] serverSideHash = null; try { // Multipart Uploads don't have an MD5 calculated on the service side if (ServiceUtils.isMultipartUploadETag(s3Object.getObjectMetadata().getETag()) == false) { clientSideHash = Md5Utils.computeMD5Hash(new FileInputStream(destinationFile)); serverSideHash = BinaryUtils.fromHex(s3Object.getObjectMetadata().getETag()); } } catch (Exception e) { log.warn("Unable to calculate MD5 hash to validate download: " + e.getMessage(), e); } if (performIntegrityCheck && clientSideHash != null && serverSideHash != null && !Arrays.equals(clientSideHash, serverSideHash)) { throw new SCSClientException("Unable to verify integrity of data download. " + "Client calculated content hash didn't match hash calculated by Amazon S3. " + "The data stored in '" + destinationFile.getAbsolutePath() + "' may be corrupt."); } } /** * Interface for the task of downloading object from S3 to a specific file, * enabling one-time retry mechanism after integrity check failure * on the downloaded file. */ public interface RetryableS3DownloadTask { /** * User defines how to get the S3Object from S3 for this RetryableS3DownloadTask. * * @return * The S3Object containing a reference to an InputStream * containing the object's data. */ public S3Object getS3ObjectStream(); /** * User defines whether integrity check is needed for this RetryableS3DownloadTask. * * @return * Boolean value indicating whether this task requires integrity check * after downloading the S3 object to file. */ public boolean needIntegrityCheck(); } /** * Gets an object stored in S3 and downloads it into the specified file. * This method includes the one-time retry mechanism after integrity check failure * on the downloaded file. It will also return immediately after getting null valued * S3Object (when getObject request does not meet the specified constraints). * * @param file * The file to store the object's data in. * @param safeS3DownloadTask * The implementation of SafeS3DownloadTask interface which allows user to * get access to all the visible variables at the calling site of this method. */ public static S3Object retryableDownloadS3ObjectToFile(File file, RetryableS3DownloadTask retryableS3DownloadTask) { boolean hasRetried = false; boolean needRetry; S3Object s3Object; do { needRetry = false; s3Object = retryableS3DownloadTask.getS3ObjectStream(); if (s3Object == null) return null; try { ServiceUtils.downloadObjectToFile(s3Object, file, retryableS3DownloadTask.needIntegrityCheck()); } catch (SCSClientException ace) { // Determine whether an immediate retry is needed according to the captured AmazonClientException. // (There are three cases when downloadObjectToFile() throws AmazonClientException: // 1) SocketException or SSLProtocolException when writing to disk (e.g. when user aborts the download) // 2) Other IOException when writing to disk // 3) MD5 hashes don't match // The current code will retry the download only when case 2) or 3) happens. if (ace.getCause() instanceof SocketException || ace.getCause() instanceof SSLProtocolException) { throw ace; } else { needRetry = true; if (hasRetried) throw ace; else { log.info("Retry the download of object " + s3Object.getKey() + " (bucket " + s3Object.getBucketName() + ")", ace); hasRetried = true; } } } finally { try { s3Object.getObjectContent().abort(); } catch (IOException e) { } } } while (needRetry); return s3Object; } }