Package org.eclipse.jgit.transport

Class AmazonS3

java.lang.Object

org.eclipse.jgit.transport.AmazonS3

public class AmazonS3
extends Object

A simple HTTP REST client for the Amazon S3 service.

This client uses the REST API to communicate with the Amazon S3 servers and read or write content through a bucket that the user has access to. It is a very lightweight implementation of the S3 API and therefore does not have all of the bells and whistles of popular client implementations.

Authentication is always performed using the user's AWSAccessKeyId and their private AWSSecretAccessKey.

Optional client-side encryption may be enabled if requested. The format is compatible with jets3t, a popular Java based Amazon S3 client library. Enabling encryption can hide sensitive data from the operators of the S3 service.

Constructor Summary

Constructors

Constructor	Description
AmazonS3(Properties props)	Create a new S3 client for the supplied user information.

Method Summary

Modifier and Type	Method	Description
OutputStream	beginPut(String bucket, String key, ProgressMonitor monitor, String monitorTask)	Atomically create or replace a single large object.
InputStream	decrypt(URLConnection u)	Decrypt an input stream from get(String, String).
void	delete(String bucket, String key)	Delete a single object.
URLConnection	get(String bucket, String key)	Get the content of a bucket object.
List<String>	list(String bucket, String prefix)	List the names of keys available within a bucket.
void	put(String bucket, String key, byte[] data)	Atomically create or replace a single small object.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

AmazonS3

public AmazonS3(Properties props)

Create a new S3 client for the supplied user information.

The connection properties are a subset of those supported by the popular jets3t library. For example:

 # AWS API signature version, must be one of:
 #   2 - deprecated (not supported in all AWS regions)
 #   4 - latest (supported in all AWS regions)
 # Defaults to 2.
 aws.api.signature.version: 4

 # AWS Access and Secret Keys (required)
 accesskey: <YourAWSAccessKey>
 secretkey: <YourAWSSecretKey>

 # Access Control List setting to apply to uploads, must be one of:
 # PRIVATE, PUBLIC_READ (defaults to PRIVATE).
 acl: PRIVATE

 # S3 Domain
 # AWS S3 Region Domain (defaults to s3.amazonaws.com)
 domain: s3.amazonaws.com

 # AWS S3 Region (required if aws.api.signature.version = 4)
 region: us-west-2

 # Number of times to retry after internal error from S3.
 httpclient.retry-max: 3

 # End-to-end encryption (hides content from S3 owners)
 password: <encryption pass-phrase>
 crypto.algorithm: PBEWithMD5AndDES
 

Parameters:

props - connection properties.

Method Detail

get

public URLConnection get(String bucket,
                         String key)
                  throws IOException

Get the content of a bucket object.

Parameters:

bucket - name of the bucket storing the object.

key - key of the object within its bucket.

Returns:

connection to stream the content of the object. The request properties of the connection may not be modified by the caller as the request parameters have already been signed.

Throws:

IOException - sending the request was not possible.

decrypt

public InputStream decrypt(URLConnection u)
                    throws IOException

Decrypt an input stream from get(String, String).

Parameters:

u - connection previously created by get(String, String)}.

Returns:

stream to read plain text from.

Throws:

IOException - decryption could not be configured.

list

public List<String> list(String bucket,
                         String prefix)
                  throws IOException

List the names of keys available within a bucket.

This method is primarily meant for obtaining a "recursive directory listing" rooted under the specified bucket and prefix location. It returns the keys sorted in reverse order of LastModified time (freshest keys first).

Parameters:

bucket - name of the bucket whose objects should be listed.

prefix - common prefix to filter the results by. Must not be null. Supplying the empty string will list all keys in the bucket. Supplying a non-empty string will act as though a trailing '/' appears in prefix, even if it does not.

Returns:

list of keys starting with prefix, after removing prefix (or prefix + "/")from all of them.

Throws:

IOException - sending the request was not possible, or the response XML document could not be parsed properly.

delete

public void delete(String bucket,
                   String key)
            throws IOException

Delete a single object.

Deletion always succeeds, even if the object does not exist.

Parameters:

bucket - name of the bucket storing the object.

key - key of the object within its bucket.

Throws:

IOException - deletion failed due to communications error.

put

public void put(String bucket,
                String key,
                byte[] data)
         throws IOException

Atomically create or replace a single small object.

This form is only suitable for smaller contents, where the caller can reasonable fit the entire thing into memory.

End-to-end data integrity is assured by internally computing the MD5 checksum of the supplied data and transmitting the checksum along with the data itself.

Parameters:

bucket - name of the bucket storing the object.

key - key of the object within its bucket.

data - new data content for the object. Must not be null. Zero length array will create a zero length object.

Throws:

IOException - creation/updating failed due to communications error.

beginPut

public OutputStream beginPut(String bucket,
                             String key,
                             ProgressMonitor monitor,
                             String monitorTask)
                      throws IOException

Atomically create or replace a single large object.

Initially the returned output stream buffers data into memory, but if the total number of written bytes starts to exceed an internal limit the data is spooled to a temporary file on the local drive.

Network transmission is attempted only when close() gets called at the end of output. Closing the returned stream can therefore take significant time, especially if the written content is very large.

End-to-end data integrity is assured by internally computing the MD5 checksum of the supplied data and transmitting the checksum along with the data itself.

Parameters:

bucket - name of the bucket storing the object.

key - key of the object within its bucket.

monitor - (optional) progress monitor to post upload completion to during the stream's close method.

monitorTask - (optional) task name to display during the close method.

Returns:

a stream which accepts the new data, and transmits once closed.

Throws:

IOException - if encryption was enabled it could not be configured.