RightAwsBase
# File lib/s3/right_s3_interface.rb, line 64 def self.bench_s3 @@bench.service end
# File lib/s3/right_s3_interface.rb, line 61 def self.bench_xml @@bench.xml end
Creates new RightS3 instance.
s3 = RightAws::S3Interface.new('1E3GDYEOGFJPIT7XXXXXX','hgTHt68JY07JKUY08ftHYtERkjgtfERn57XXXXXX', {:logger => Logger.new('/tmp/x.log')}) #=> #<RightAws::S3Interface:0xb7b3c27c>
Params is a hash:
{:server => 's3.amazonaws.com' # Amazon service host: 's3.amazonaws.com'(default) :port => 443 # Amazon service port: 80 or 443(default) :protocol => 'https' # Amazon service protocol: 'http' or 'https'(default) :logger => Logger Object # Logger instance: logs to STDOUT if omitted :no_subdomains => true} # Force placing bucket name into path instead of domain name
# File lib/s3/right_s3_interface.rb, line 94 def initialize(aws_access_key_id=nil, aws_secret_access_key=nil, params={}) init({ :name => 'S3', :default_host => ENV['S3_URL'] ? URI.parse(ENV['S3_URL']).host : DEFAULT_HOST, :default_port => ENV['S3_URL'] ? URI.parse(ENV['S3_URL']).port : DEFAULT_PORT, :default_service => ENV['S3_URL'] ? URI.parse(ENV['S3_URL']).path : DEFAULT_SERVICE, :default_protocol => ENV['S3_URL'] ? URI.parse(ENV['S3_URL']).scheme : DEFAULT_PROTOCOL }, aws_access_key_id || ENV['AWS_ACCESS_KEY_ID'], aws_secret_access_key || ENV['AWS_SECRET_ACCESS_KEY'], params) end
Retrieve bucket location
s3.create_bucket('my-awesome-bucket-us') #=> true puts s3.bucket_location('my-awesome-bucket-us') #=> '' (Amazon's default value assumed) s3.create_bucket('my-awesome-bucket-eu', :location => :eu) #=> true puts s3.bucket_location('my-awesome-bucket-eu') #=> 'EU'
# File lib/s3/right_s3_interface.rb, line 261 def bucket_location(bucket, headers={}) req_hash = generate_rest_request('GET', headers.merge(:url=>"#{bucket}?location")) request_info(req_hash, S3BucketLocationParser.new) rescue on_exception end
Removes all keys from bucket. Returns true or an exception.
s3.clear_bucket('my_awesome_bucket') #=> true
# File lib/s3/right_s3_interface.rb, line 966 def clear_bucket(bucket) incrementally_list_bucket(bucket) do |results| results[:contents].each { |key| delete(bucket, key[:key]) } end true rescue on_exception end
Copy an object.
directive: :copy - copy meta-headers from source (default value) :replace - replace meta-headers by passed ones # copy a key with meta-headers s3.copy('b1', 'key1', 'b1', 'key1_copy') #=> {:e_tag=>"\"e8b...8d\"", :last_modified=>"2008-05-11T10:25:22.000Z"} # copy a key, overwrite meta-headers s3.copy('b1', 'key2', 'b1', 'key2_copy', :replace, 'x-amz-meta-family'=>'Woho555!') #=> {:e_tag=>"\"e8b...8d\"", :last_modified=>"2008-05-11T10:26:22.000Z"}
see: docs.amazonwebservices.com/AmazonS3/2006-03-01/UsingCopyingObjects.html
http://docs.amazonwebservices.com/AmazonS3/2006-03-01/RESTObjectCOPY.html
# File lib/s3/right_s3_interface.rb, line 834 def copy(src_bucket, src_key, dest_bucket, dest_key=nil, directive=:copy, headers={}) dest_key ||= src_key headers['x-amz-metadata-directive'] = directive.to_s.upcase headers['x-amz-copy-source'] = "#{src_bucket}/#{CGI::escape src_key}" req_hash = generate_rest_request('PUT', headers.merge(:url=>"#{dest_bucket}/#{CGI::escape dest_key}")) request_info(req_hash, S3CopyParser.new) rescue on_exception end
Creates new bucket. Returns true or an exception.
# create a bucket at American server s3.create_bucket('my-awesome-bucket-us') #=> true # create a bucket at European server s3.create_bucket('my-awesome-bucket-eu', :location => :eu) #=> true
# File lib/s3/right_s3_interface.rb, line 235 def create_bucket(bucket, headers={}) data = nil location = case headers[:location].to_s when 'us','US' then '' when 'eu' then 'EU' else headers[:location].to_s end unless location.right_blank? data = "<CreateBucketConfiguration><LocationConstraint>#{location}</LocationConstraint></CreateBucketConfiguration>" end req_hash = generate_rest_request('PUT', headers.merge(:url=>bucket, :data => data)) request_info(req_hash, RightHttp2xxParser.new) rescue Exception => e # if the bucket exists AWS returns an error for the location constraint interface. Drop it e.is_a?(RightAws::AwsError) && e.message.include?('BucketAlreadyOwnedByYou') ? true : on_exception end
Generates link for 'CreateBucket'.
s3.create_bucket_link('my_awesome_bucket') #=> url string
# File lib/s3/right_s3_interface.rb, line 1059 def create_bucket_link(bucket, expires=nil, headers={}) generate_link('PUT', headers.merge(:url=>bucket), expires) rescue on_exception end
Deletes key. Returns true or an exception.
s3.delete('my_awesome_bucket', 'log/curent/1.log') #=> true
# File lib/s3/right_s3_interface.rb, line 786 def delete(bucket, key='', headers={}) req_hash = generate_rest_request('DELETE', headers.merge(:url=>"#{bucket}/#{CGI::escape key}")) request_info(req_hash, RightHttp2xxParser.new) rescue on_exception end
Deletes new bucket. Bucket must be empty! Returns true or an exception.
s3.delete_bucket('my_awesome_bucket') #=> true
See also: force_delete_bucket method
# File lib/s3/right_s3_interface.rb, line 305 def delete_bucket(bucket, headers={}) req_hash = generate_rest_request('DELETE', headers.merge(:url=>bucket)) request_info(req_hash, RightHttp2xxParser.new) rescue on_exception end
Generates link for 'DeleteBucket'.
s3.delete_bucket_link('my_awesome_bucket') #=> url string
# File lib/s3/right_s3_interface.rb, line 1069 def delete_bucket_link(bucket, expires=nil, headers={}) generate_link('DELETE', headers.merge(:url=>bucket), expires) rescue on_exception end
Deletes all keys where the 'folder_key' may be assumed as 'folder' name. Returns an array of string keys that have been deleted.
s3.list_bucket('my_awesome_bucket').map{|key_data| key_data[:key]} #=> ['test','test/2/34','test/3','test1','test1/logs'] s3.delete_folder('my_awesome_bucket','test') #=> ['test','test/2/34','test/3']
# File lib/s3/right_s3_interface.rb, line 991 def delete_folder(bucket, folder_key, separator='/') folder_key.chomp!(separator) allkeys = [] incrementally_list_bucket(bucket, { 'prefix' => folder_key }) do |results| keys = results[:contents].map{ |s3_key| s3_key[:key][/^#{folder_key}($|#{separator}.*)/] ? s3_key[:key] : nil}.compact keys.each{ |key| delete(bucket, key) } allkeys << keys end allkeys rescue on_exception end
Generates link for 'DeleteObject'.
s3.delete_link('my_awesome_bucket',key) #=> url string
# File lib/s3/right_s3_interface.rb, line 1140 def delete_link(bucket, key, expires=nil, headers={}) generate_link('DELETE', headers.merge(:url=>"#{bucket}/#{s3_link_escape(key)}"), expires) rescue on_exception end
Deletes multiple keys. Returns an array with errors, if any.
s3.delete_multiple('my_awesome_bucket', ['key1', 'key2', ...) #=> [ { :key => 'key2', :code => 'AccessDenied', :message => "Access Denied" } ]
# File lib/s3/right_s3_interface.rb, line 798 def delete_multiple(bucket, keys=[], headers={}) errors = [] keys = Array.new(keys) while keys.length > 0 data = "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n" data += "<Delete>\n<Quiet>true</Quiet>\n" keys.take(MULTI_OBJECT_DELETE_MAX_KEYS).each do |key| data += "<Object><Key>#{AwsUtils::xml_escape(key)}</Key></Object>\n" end data += "</Delete>" req_hash = generate_rest_request('POST', headers.merge( :url => "#{bucket}?delete", :data => data, 'content-md5' => AwsUtils::content_md5(data) )) errors += request_info(req_hash, S3DeleteMultipleParser.new) keys = keys.drop(MULTI_OBJECT_DELETE_MAX_KEYS) end errors rescue on_exception end
Deletes all keys in bucket then deletes bucket. Returns true or an exception.
s3.force_delete_bucket('my_awesome_bucket')
# File lib/s3/right_s3_interface.rb, line 979 def force_delete_bucket(bucket) clear_bucket(bucket) delete_bucket(bucket) rescue on_exception end
Retrieves object data from Amazon. Returns a hash or an exception.
s3.get('my_awesome_bucket', 'log/curent/1.log') #=> {:object => "Ola-la!", :headers => {"last-modified" => "Wed, 23 May 2007 09:08:04 GMT", "content-type" => "", "etag" => "\"000000000096f4ee74bc4596443ef2a4\"", "date" => "Wed, 23 May 2007 09:08:03 GMT", "x-amz-id-2" => "ZZZZZZZZZZZZZZZZZZZZ1HJXZoehfrS4QxcxTdNGldR7w/FVqblP50fU8cuIMLiu", "x-amz-meta-family" => "Woho556!", "x-amz-request-id" => "0000000C246D770C", "server" => "AmazonS3", "content-length" => "7"}}
If a block is provided, yields incrementally to the block as the response is read. For large responses, this function is ideal as the response can be 'streamed'. The hash containing header fields is still returned. Example: foo = File.new('./chunder.txt', File::CREAT|File::RDWR) rhdr = s3.get('aws-test', 'Cent5V1_7_1.img.part.00') do |chunk|
foo.write(chunk)
end foo.close
# File lib/s3/right_s3_interface.rb, line 689 def get(bucket, key, headers={}, &block) req_hash = generate_rest_request('GET', headers.merge(:url=>"#{bucket}/#{CGI::escape key}")) request_info(req_hash, S3HttpResponseBodyParser.new, &block) rescue on_exception end
Retieves the ACL (access control policy) for a bucket or object. Returns a hash of headers and xml doc with ACL data. See: docs.amazonwebservices.com/AmazonS3/2006-03-01/RESTAccessPolicy.html.
s3.get_acl('my_awesome_bucket', 'log/curent/1.log') #=> {:headers => {"x-amz-id-2"=>"B3BdDMDUz+phFF2mGBH04E46ZD4Qb9HF5PoPHqDRWBv+NVGeA3TOQ3BkVvPBjgxX", "content-type"=>"application/xml;charset=ISO-8859-1", "date"=>"Wed, 23 May 2007 09:40:16 GMT", "x-amz-request-id"=>"B183FA7AB5FBB4DD", "server"=>"AmazonS3", "transfer-encoding"=>"chunked"}, :object => "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<AccessControlPolicy xmlns=\"http://s3.amazonaws.com/doc/2006-03-01/\"><Owner> <ID>16144ab2929314cc309ffe736daa2b264357476c7fea6efb2c3347ac3ab2792a</ID><DisplayName>root</DisplayName></Owner> <AccessControlList><Grant><Grantee xmlns:xsi=\"http://www.w3.org/2001/XMLSchema-instance\" xsi:type=\"CanonicalUser\"><ID> 16144ab2929314cc309ffe736daa2b264357476c7fea6efb2c3347ac3ab2792a</ID><DisplayName>root</DisplayName></Grantee> <Permission>FULL_CONTROL</Permission></Grant></AccessControlList></AccessControlPolicy>" }
# File lib/s3/right_s3_interface.rb, line 885 def get_acl(bucket, key='', headers={}) key = key.right_blank? ? '' : "/#{CGI::escape key}" req_hash = generate_rest_request('GET', headers.merge(:url=>"#{bucket}#{key}?acl")) request_info(req_hash, S3HttpResponseBodyParser.new) rescue on_exception end
Generates link for 'GetACL'.
s3.get_acl_link('my_awesome_bucket',key) #=> url string
# File lib/s3/right_s3_interface.rb, line 1151 def get_acl_link(bucket, key='', headers={}) return generate_link('GET', headers.merge(:url=>"#{bucket}/#{s3_link_escape(key)}?acl")) rescue on_exception end
Retieves the ACL (access control policy) for a bucket or object. Returns a hash of {:owner, :grantees}
s3.get_acl_parse('my_awesome_bucket', 'log/curent/1.log') #=> { :grantees=> { "16...2a"=> { :display_name=>"root", :permissions=>["FULL_CONTROL"], :attributes=> { "xsi:type"=>"CanonicalUser", "xmlns:xsi"=>"http://www.w3.org/2001/XMLSchema-instance"}}, "http://acs.amazonaws.com/groups/global/AllUsers"=> { :display_name=>"AllUsers", :permissions=>["READ"], :attributes=> { "xsi:type"=>"Group", "xmlns:xsi"=>"http://www.w3.org/2001/XMLSchema-instance"}}}, :owner=> { :id=>"16..2a", :display_name=>"root"}}
# File lib/s3/right_s3_interface.rb, line 915 def get_acl_parse(bucket, key='', headers={}) key = key.right_blank? ? '' : "/#{CGI::escape key}" req_hash = generate_rest_request('GET', headers.merge(:url=>"#{bucket}#{key}?acl")) acl = request_info(req_hash, S3AclParser.new(:logger => @logger)) result = {} result[:owner] = acl[:owner] result[:grantees] = {} acl[:grantees].each do |grantee| key = grantee[:id] || grantee[:uri] if result[:grantees].key?(key) result[:grantees][key][:permissions] << grantee[:permissions] else result[:grantees][key] = { :display_name => grantee[:display_name] || grantee[:uri].to_s[/[^\/]*$/], :permissions => Array(grantee[:permissions]), :attributes => grantee[:attributes] } end end result rescue on_exception end
Retieves the ACL (access control policy) for a bucket. Returns a hash of headers and xml doc with ACL data.
# File lib/s3/right_s3_interface.rb, line 948 def get_bucket_acl(bucket, headers={}) return get_acl(bucket, '', headers) rescue on_exception end
Generates link for 'GetBucketACL'.
s3.get_acl_link('my_awesome_bucket',key) #=> url string
# File lib/s3/right_s3_interface.rb, line 1171 def get_bucket_acl_link(bucket, headers={}) return get_acl_link(bucket, '', headers) rescue on_exception end
Generates link for 'GetObject'.
if a bucket comply with virtual hosting naming then retuns a link with the bucket as a part of host name:
s3.get_link('my-awesome-bucket',key) #=> https://my-awesome-bucket.s3.amazonaws.com:443/asia%2Fcustomers?Signature=nh7...
otherwise returns an old style link (the bucket is a part of path):
s3.get_link('my_awesome_bucket',key) #=> https://s3.amazonaws.com:443/my_awesome_bucket/asia%2Fcustomers?Signature=QAO...
see docs.amazonwebservices.com/AmazonS3/2006-03-01/VirtualHosting.html
To specify response-* parameters, define them in the response_params hash:
s3.get_link('my_awesome_bucket',key,nil,{},{ "response-content-disposition" => "attachment; filename=caf�.png", "response-content-type" => "image/png"}) #=> https://s3.amazonaws.com:443/my_awesome_bucket/asia%2Fcustomers?response-content-disposition=attachment%3B%20filename%3Dcaf%25C3%25A9.png&response-content-type=image%2Fpng&Signature=wio...
# File lib/s3/right_s3_interface.rb, line 1115 def get_link(bucket, key, expires=nil, headers={}, response_params={}) if response_params.size > 0 response_params = '?' + response_params.map { |k, v| "#{k}=#{s3_link_escape(v)}" }.join('&') else response_params = '' end generate_link('GET', headers.merge(:url=>"#{bucket}/#{s3_link_escape(key)}#{response_params}"), expires) rescue on_exception end
Retrieves the logging configuration for a bucket. Returns a hash of {:enabled, :targetbucket, :targetprefix}
s3.interface.get_logging_parse(:bucket => "asset_bucket")
=> {:enabled=>true, :targetbucket=>"mylogbucket", :targetprefix=>"loggylogs/"}
# File lib/s3/right_s3_interface.rb, line 275 def get_logging_parse(params) AwsUtils.mandatory_arguments([:bucket], params) AwsUtils.allow_only([:bucket, :headers], params) params[:headers] = {} unless params[:headers] req_hash = generate_rest_request('GET', params[:headers].merge(:url=>"#{params[:bucket]}?logging")) request_info(req_hash, S3LoggingParser.new) rescue on_exception end
Retrieves object data only (headers are omitted). Returns string or an exception.
s3.get('my_awesome_bucket', 'log/curent/1.log') #=> 'Ola-la!'
# File lib/s3/right_s3_interface.rb, line 1008 def get_object(bucket, key, headers={}) get(bucket, key, headers)[:object] rescue on_exception end
Retrieves object metadata. Returns a hash of http_response_headers.
s3.head('my_awesome_bucket', 'log/curent/1.log') #=> {"last-modified" => "Wed, 23 May 2007 09:08:04 GMT", "content-type" => "", "etag" => "\"000000000096f4ee74bc4596443ef2a4\"", "date" => "Wed, 23 May 2007 09:08:03 GMT", "x-amz-id-2" => "ZZZZZZZZZZZZZZZZZZZZ1HJXZoehfrS4QxcxTdNGldR7w/FVqblP50fU8cuIMLiu", "x-amz-meta-family" => "Woho556!", "x-amz-request-id" => "0000000C246D770C", "server" => "AmazonS3", "content-length" => "7"}
# File lib/s3/right_s3_interface.rb, line 775 def head(bucket, key, headers={}) req_hash = generate_rest_request('HEAD', headers.merge(:url=>"#{bucket}/#{CGI::escape key}")) request_info(req_hash, S3HttpResponseHeadParser.new) rescue on_exception end
Generates link for 'HeadObject'.
s3.head_link('my_awesome_bucket',key) #=> url string
# File lib/s3/right_s3_interface.rb, line 1130 def head_link(bucket, key, expires=nil, headers={}) generate_link('HEAD', headers.merge(:url=>"#{bucket}/#{s3_link_escape(key)}"), expires) rescue on_exception end
Incrementally list the contents of a bucket. Yields the following hash to a block:
s3.incrementally_list_bucket('my_awesome_bucket', { 'prefix'=>'t', 'marker'=>'', 'max-keys'=>5, delimiter=>'' }) yields { :name => 'bucketname', :prefix => 'subfolder/', :marker => 'fileN.jpg', :max_keys => 234, :delimiter => '/', :is_truncated => true, :next_marker => 'fileX.jpg', :contents => [ { :key => "file1", :last_modified => "2007-05-18T07:00:59.000Z", :e_tag => "000000000059075b964b07152d234b70", :size => 3, :storage_class => "STANDARD", :owner_id => "00000000009314cc309ffe736daa2b264357476c7fea6efb2c3347ac3ab2792a", :owner_display_name => "root" }, { :key, ...}, ... {:key, ...} ] :common_prefixes => [ "prefix1", "prefix2", ..., "prefixN" ] }
# File lib/s3/right_s3_interface.rb, line 363 def incrementally_list_bucket(bucket, options={}, headers={}, &block) internal_options = options.right_symbolize_keys begin internal_bucket = bucket.dup internal_bucket += '?'+internal_options.map{|k, v| "#{k.to_s}=#{CGI::escape v.to_s}"}.join('&') unless internal_options.right_blank? req_hash = generate_rest_request('GET', headers.merge(:url=>internal_bucket)) response = request_info(req_hash, S3ImprovedListBucketParser.new(:logger => @logger)) there_are_more_keys = response[:is_truncated] if(there_are_more_keys) internal_options[:marker] = decide_marker(response) total_results = response[:contents].length + response[:common_prefixes].length internal_options[:'max-keys'] ? (internal_options[:'max-keys'] -= total_results) : nil end yield response end while there_are_more_keys && under_max_keys(internal_options) true rescue on_exception end
docs.amazonwebservices.com/AmazonS3/2006-03-01/index.html?BucketRestrictions.html
# File lib/s3/right_s3_interface.rb, line 147 def is_dns_bucket?(bucket_name) bucket_name = bucket_name.to_s return nil unless (3..63) === bucket_name.size bucket_name.split('.').each do |component| return nil unless component[/^[a-z0-9]([a-z0-9-]*[a-z0-9])?$/] end true end
Returns an array of customer's buckets. Each item is a hash.
s3.list_all_my_buckets #=> [{:owner_id => "00000000009314cc309ffe736daa2b264357476c7fea6efb2c3347ac3ab2792a", :owner_display_name => "root", :name => "bucket_name", :creation_date => "2007-04-19T18:47:43.000Z"}, ..., {...}]
# File lib/s3/right_s3_interface.rb, line 221 def list_all_my_buckets(headers={}) req_hash = generate_rest_request('GET', headers.merge(:url=>'')) request_info(req_hash, S3ListAllMyBucketsParser.new(:logger => @logger)) rescue on_exception end
Generates link for 'ListAllMyBuckets'.
s3.list_all_my_buckets_link #=> url string
# File lib/s3/right_s3_interface.rb, line 1049 def list_all_my_buckets_link(expires=nil, headers={}) generate_link('GET', headers.merge(:url=>''), expires) rescue on_exception end
Returns an array of bucket's keys. Each array item (key data) is a hash.
s3.list_bucket('my_awesome_bucket', { 'prefix'=>'t', 'marker'=>'', 'max-keys'=>5, delimiter=>'' }) #=> [{:key => "test1", :last_modified => "2007-05-18T07:00:59.000Z", :owner_id => "00000000009314cc309ffe736daa2b264357476c7fea6efb2c3347ac3ab2792a", :owner_display_name => "root", :e_tag => "000000000059075b964b07152d234b70", :storage_class => "STANDARD", :size => 3, :service=> {'is_truncated' => false, 'prefix' => "t", 'marker' => "", 'name' => "my_awesome_bucket", 'max-keys' => "5"}, ..., {...}]
# File lib/s3/right_s3_interface.rb, line 328 def list_bucket(bucket, options={}, headers={}) bucket += '?'+options.map{|k, v| "#{k.to_s}=#{CGI::escape v.to_s}"}.join('&') unless options.right_blank? req_hash = generate_rest_request('GET', headers.merge(:url=>bucket)) request_info(req_hash, S3ListBucketParser.new(:logger => @logger)) rescue on_exception end
Generates link for 'ListBucket'.
s3.list_bucket_link('my_awesome_bucket') #=> url string
# File lib/s3/right_s3_interface.rb, line 1079 def list_bucket_link(bucket, options=nil, expires=nil, headers={}) bucket += '?' + options.map{|k, v| "#{k.to_s}=#{s3_link_escape(v)}"}.join('&') unless options.right_blank? generate_link('GET', headers.merge(:url=>bucket), expires) rescue on_exception end
Move an object.
directive: :copy - copy meta-headers from source (default value) :replace - replace meta-headers by passed ones # move bucket1/key1 to bucket1/key2 s3.move('bucket1', 'key1', 'bucket1', 'key2') #=> {:e_tag=>"\"e8b...8d\"", :last_modified=>"2008-05-11T10:27:22.000Z"} # move bucket1/key1 to bucket2/key2 with new meta-headers assignment s3.copy('bucket1', 'key1', 'bucket2', 'key2', :replace, 'x-amz-meta-family'=>'Woho555!') #=> {:e_tag=>"\"e8b...8d\"", :last_modified=>"2008-05-11T10:28:22.000Z"}
# File lib/s3/right_s3_interface.rb, line 854 def move(src_bucket, src_key, dest_bucket, dest_key=nil, directive=:copy, headers={}) copy_result = copy(src_bucket, src_key, dest_bucket, dest_key, directive, headers) # delete an original key if it differs from a destination one delete(src_bucket, src_key) unless src_bucket == dest_bucket && src_key == dest_key copy_result end
get custom option
# File lib/s3/right_s3_interface.rb, line 76 def param(name) # - check explicitly defined param (@params) # - otherwise check implicitly defined one (@@params) @params.has_key?(name) ? @params[name] : @@params[name] end
Saves object to Amazon. Returns true or an exception. Any header starting with AMAZON_METADATA_PREFIX is considered user metadata. It will be stored with the object and returned when you retrieve the object. The total size of the HTTP request, not including the body, must be less than 4 KB.
s3.put('my_awesome_bucket', 'log/current/1.log', 'Ola-la!', 'x-amz-meta-family'=>'Woho556!') #=> true
This method is capable of 'streaming' uploads; that is, it can upload data from a file or other IO object without first reading all the data into memory. This is most useful for large PUTs - it is difficult to read a 2 GB file entirely into memory before sending it to S3. To stream an upload, pass an object that responds to 'read' (like the read method of IO) and to either 'lstat' or 'size'. For files, this means streaming is enabled by simply making the call:
s3.put(bucket_name, 'S3keyname.forthisfile', File.open('localfilename.dat'))
If the IO object you wish to stream from responds to the read method but doesn't implement lstat or size, you can extend the object dynamically to implement these methods, or define your own class which defines these methods. Be sure that your class returns 'nil' from read() after having read 'size' bytes. Otherwise S3 will drop the socket after 'Content-Length' bytes have been uploaded, and HttpConnection will interpret this as an error.
This method now supports very large PUTs, where very large is > 2 GB.
For Win32 users: Files and IO objects should be opened in binary mode. If a text mode IO object is passed to PUT, it will be converted to binary mode.
# File lib/s3/right_s3_interface.rb, line 438 def put(bucket, key, data=nil, headers={}, &blck) # On Windows, if someone opens a file in text mode, we must reset it so # to binary mode for streaming to work properly if(data.respond_to?(:binmode)) data.binmode end if (data.respond_to?(:lstat) && data.lstat.size >= USE_100_CONTINUE_PUT_SIZE) || (data.respond_to?(:size) && data.size >= USE_100_CONTINUE_PUT_SIZE) headers['expect'] = '100-continue' end req_hash = generate_rest_request('PUT', headers.merge(:url=>"#{bucket}/#{CGI::escape key}", :data=>data)) request_info(req_hash, RightHttp2xxParser.new, &blck) rescue on_exception end
Sets the ACL on a bucket or object.
# File lib/s3/right_s3_interface.rb, line 939 def put_acl(bucket, key, acl_xml_doc, headers={}) key = key.right_blank? ? '' : "/#{CGI::escape key}" req_hash = generate_rest_request('PUT', headers.merge(:url=>"#{bucket}#{key}?acl", :data=>acl_xml_doc)) request_info(req_hash, S3HttpResponseBodyParser.new) rescue on_exception end
Generates link for 'PutACL'.
s3.put_acl_link('my_awesome_bucket',key) #=> url string
# File lib/s3/right_s3_interface.rb, line 1161 def put_acl_link(bucket, key='', headers={}) return generate_link('PUT', headers.merge(:url=>"#{bucket}/#{s3_link_escape(key)}?acl")) rescue on_exception end
Sets the ACL on a bucket only.
# File lib/s3/right_s3_interface.rb, line 955 def put_bucket_acl(bucket, acl_xml_doc, headers={}) return put_acl(bucket, '', acl_xml_doc, headers) rescue on_exception end
Generates link for 'PutBucketACL'.
s3.put_acl_link('my_awesome_bucket',key) #=> url string
# File lib/s3/right_s3_interface.rb, line 1181 def put_bucket_acl_link(bucket, acl_xml_doc, headers={}) return put_acl_link(bucket, '', acl_xml_doc, headers) rescue on_exception end
Generates link for 'PutObject'.
s3.put_link('my_awesome_bucket',key, object) #=> url string
# File lib/s3/right_s3_interface.rb, line 1090 def put_link(bucket, key, data=nil, expires=nil, headers={}) generate_link('PUT', headers.merge(:url=>"#{bucket}/#{s3_link_escape(key)}", :data=>data), expires) rescue on_exception end
Sets logging configuration for a bucket from the XML configuration document.
params: :bucket :xmldoc
# File lib/s3/right_s3_interface.rb, line 289 def put_logging(params) AwsUtils.mandatory_arguments([:bucket,:xmldoc], params) AwsUtils.allow_only([:bucket,:xmldoc, :headers], params) params[:headers] = {} unless params[:headers] req_hash = generate_rest_request('PUT', params[:headers].merge(:url=>"#{params[:bucket]}?logging", :data => params[:xmldoc])) request_info(req_hash, RightHttp2xxParser.new) rescue on_exception end
Rename an object.
# rename bucket1/key1 to bucket1/key2 s3.rename('bucket1', 'key1', 'key2') #=> {:e_tag=>"\"e8b...8d\"", :last_modified=>"2008-05-11T10:29:22.000Z"}
# File lib/s3/right_s3_interface.rb, line 866 def rename(src_bucket, src_key, dest_key, headers={}) move(src_bucket, src_key, src_bucket, dest_key, :copy, headers) end
New experimental API for retrieving objects, introduced in RightAws 1.8.1. retrieve_object is similar in function to the older function get. It allows for optional verification of object md5 checksums on retrieval. Parameters are passed as hash entries and are checked for completeness as well as for spurious arguments.
If the optional :md5 argument is provided, retrieve_object verifies that the given md5 matches the md5 returned by S3. The :verified_md5 field in the response hash is set true or false depending on the outcome of this check. If no :md5 argument is given, :verified_md5 will be false in the response.
The optional argument of :headers allows the caller to specify arbitrary request header values. Mandatory arguments:
:bucket - the bucket in which the object is stored :key - the object address (or path) within the bucket
Optional arguments:
:headers - hash of additional HTTP headers to include with the request :md5 - MD5 checksum against which to verify the retrieved object s3.retrieve_object(:bucket => "foobucket", :key => "foo") => {:verified_md5=>false, :headers=>{"last-modified"=>"Mon, 29 Sep 2008 18:58:56 GMT", "x-amz-id-2"=>"2Aj3TDz6HP5109qly//18uHZ2a1TNHGLns9hyAtq2ved7wmzEXDOPGRHOYEa3Qnp", "content-type"=>"", "etag"=>"\"a507841b1bc8115094b00bbe8c1b2954\"", "date"=>"Tue, 30 Sep 2008 00:52:44 GMT", "x-amz-request-id"=>"EE4855DE27A2688C", "server"=>"AmazonS3", "content-length"=>"10"}, :object=>"polemonium"} s3.retrieve_object(:bucket => "foobucket", :key => "foo", :md5=>'a507841b1bc8115094b00bbe8c1b2954') => {:verified_md5=>true, :headers=>{"last-modified"=>"Mon, 29 Sep 2008 18:58:56 GMT", "x-amz-id-2"=>"mLWQcI+VuKVIdpTaPXEo84g0cz+vzmRLbj79TS8eFPfw19cGFOPxuLy4uGYVCvdH", "content-type"=>"", "etag"=>"\"a507841b1bc8115094b00bbe8c1b2954\"", "date"=>"Tue, 30 Sep 2008 00:53:08 GMT", "x-amz-request-id"=>"6E7F317356580599", "server"=>"AmazonS3", "content-length"=>"10"}, :object=>"polemonium"}
If a block is provided, yields incrementally to the block as the response is read. For large responses, this function is ideal as the response can be 'streamed'. The hash containing header fields is still returned.
# File lib/s3/right_s3_interface.rb, line 737 def retrieve_object(params, &block) AwsUtils.mandatory_arguments([:bucket, :key], params) AwsUtils.allow_only([:bucket, :key, :headers, :md5], params) params[:headers] = {} unless params[:headers] req_hash = generate_rest_request('GET', params[:headers].merge(:url=>"#{params[:bucket]}/#{CGI::escape params[:key]}")) resp = request_info(req_hash, S3HttpResponseBodyParser.new, &block) resp[:verified_md5] = false if(params[:md5] && (resp[:headers]['etag'].gsub(/\"/,'') == params[:md5])) resp[:verified_md5] = true end resp rescue on_exception end
Identical in function to retrieve_object, but requires verification that the returned ETag is identical to the checksum passed in by the user as the 'md5' argument. If the check passes, returns the response metadata with the "verified_md5" field set true. Raises an exception if the checksums conflict. This call is implemented as a wrapper around retrieve_object and the user may gain different semantics by creating a custom wrapper.
# File lib/s3/right_s3_interface.rb, line 755 def retrieve_object_and_verify(params, &block) AwsUtils.mandatory_arguments([:md5], params) resp = retrieve_object(params, &block) return resp if resp[:verified_md5] raise AwsError.new("Retrieved object failed MD5 checksum verification: #{resp.inspect}") end
# File lib/s3/right_s3_interface.rb, line 1018 def s3_link_escape(text) #CGI::escape(text.to_s).gsub(/[+]/, '%20') AwsUtils::amz_escape(text.to_s) end
New experimental API for uploading objects, introduced in RightAws 1.8.1. store_object is similar in function to the older function put, but returns the full response metadata. It also allows for optional verification of object md5 checksums on upload. Parameters are passed as hash entries and are checked for completeness as well as for spurious arguments. The hash of the response headers contains useful information like the Amazon request ID and the object ETag (MD5 checksum).
If the optional :md5 argument is provided, store_object verifies that the given md5 matches the md5 returned by S3. The :verified_md5 field in the response hash is set true or false depending on the outcome of this check. If no :md5 argument is given, :verified_md5 will be false in the response.
The optional argument of :headers allows the caller to specify arbitrary request header values.
s3.store_object(:bucket => "foobucket", :key => "foo", :md5 => "a507841b1bc8115094b00bbe8c1b2954", :data => "polemonium" )
=> {"x-amz-id-2"=>"SVsnS2nfDaR+ixyJUlRKM8GndRyEMS16+oZRieamuL61pPxPaTuWrWtlYaEhYrI/", "etag"=>"\"a507841b1bc8115094b00bbe8c1b2954\"", "date"=>"Mon, 29 Sep 2008 18:57:46 GMT", :verified_md5=>true, "x-amz-request-id"=>"63916465939995BA", "server"=>"AmazonS3", "content-length"=>"0"}
s3.store_object(:bucket => "foobucket", :key => "foo", :data => "polemonium" )
=> {"x-amz-id-2"=>"MAt9PLjgLX9UYJ5tV2fI/5dBZdpFjlzRVpWgBDpvZpl+V+gJFcBMW2L+LBstYpbR", "etag"=>"\"a507841b1bc8115094b00bbe8c1b2954\"", "date"=>"Mon, 29 Sep 2008 18:58:56 GMT", :verified_md5=>false, "x-amz-request-id"=>"3B25A996BC2CDD3B", "server"=>"AmazonS3", "content-length"=>"0"}
# File lib/s3/right_s3_interface.rb, line 484 def store_object(params) AwsUtils.allow_only([:bucket, :key, :data, :headers, :md5], params) AwsUtils.mandatory_arguments([:bucket, :key, :data], params) params[:headers] = {} unless params[:headers] params[:data].binmode if(params[:data].respond_to?(:binmode)) # On Windows, if someone opens a file in text mode, we must reset it to binary mode for streaming to work properly if (params[:data].respond_to?(:lstat) && params[:data].lstat.size >= USE_100_CONTINUE_PUT_SIZE) || (params[:data].respond_to?(:size) && params[:data].size >= USE_100_CONTINUE_PUT_SIZE) params[:headers]['expect'] = '100-continue' end req_hash = generate_rest_request('PUT', params[:headers].merge(:url=>"#{params[:bucket]}/#{CGI::escape params[:key]}", :data=>params[:data])) resp = request_info(req_hash, S3HttpResponseHeadParser.new) if(params[:md5]) resp[:verified_md5] = (resp['etag'].gsub(/\"/, '') == params[:md5]) ? true : false else resp[:verified_md5] = false end resp rescue on_exception end
Identical in function to store_object, but requires verification that the returned ETag is identical to the checksum passed in by the user as the 'md5' argument. If the check passes, returns the response metadata with the "verified_md5" field set true. Raises an exception if the checksums conflict. This call is implemented as a wrapper around store_object and the user may gain different semantics by creating a custom wrapper.
s3.store_object_and_verify(:bucket => "foobucket", :key => "foo", :md5 => "a507841b1bc8115094b00bbe8c1b2954", :data => "polemonium" )
=> {"x-amz-id-2"=>"IZN3XsH4FlBU0+XYkFTfHwaiF1tNzrm6dIW2EM/cthKvl71nldfVC0oVQyydzWpb", "etag"=>"\"a507841b1bc8115094b00bbe8c1b2954\"", "date"=>"Mon, 29 Sep 2008 18:38:32 GMT", :verified_md5=>true, "x-amz-request-id"=>"E8D7EA4FE00F5DF7", "server"=>"AmazonS3", "content-length"=>"0"}
s3.store_object_and_verify(:bucket => "foobucket", :key => "foo", :md5 => "a507841b1bc8115094b00bbe8c1b2953", :data => "polemonium" )
RightAws::AwsError: Uploaded object failed MD5 checksum verification: {"x-amz-id-2"=>"HTxVtd2bf7UHHDn+WzEH43MkEjFZ26xuYvUzbstkV6nrWvECRWQWFSx91z/bl03n", "etag"=>"\"a507841b1bc8115094b00bbe8c1b2954\"", "date"=>"Mon, 29 Sep 2008 18:38:41 GMT", :verified_md5=>false, "x-amz-request-id"=>"0D7ADE09F42606F2", "server"=>"AmazonS3", "content-length"=>"0"}
# File lib/s3/right_s3_interface.rb, line 528 def store_object_and_verify(params) AwsUtils.mandatory_arguments([:md5], params) r = store_object(params) r[:verified_md5] ? (return r) : (raise AwsError.new("Uploaded object failed MD5 checksum verification: #{r.inspect}")) end
New experimental API for uploading objects using the multipart upload API. store_object_multipart is similar in function to the store_object method, but breaks the input into parts and transmits each part separately. The multipart upload API has the benefit of being be able to retransmit a part in isolation without needing to restart the entire upload. This makes it ideal for uploading large files over unreliable networks. It also does not require the file size to be known before starting the upload, making it useful for stream data as it is created (say via reading a pipe or socket). The hash of the response headers contains useful information like the location (the URI for the newly created object), bucket, key, and etag).
The optional argument of :headers allows the caller to specify arbitrary request header values.
s3.store_object_multipart(:bucket => "foobucket", :key => "foo", :data => "polemonium" )
=> {:location=>"https://s3.amazonaws.com/right_s3_awesome_test_bucket_000B1_officedrop/test%2Flarge_multipart_file", :e_tag=>"\"72b81ac08aed4d4d1055c11f56c2a258-1\"", :key=>"test/large_multipart_file", :bucket=>"right_s3_awesome_test_bucket_000B1_officedrop"}
f = File.new("some_file", "r") s3.store_object_multipart(:bucket => "foobucket", :key => "foo", :data => f )
=> {:location=>"https://s3.amazonaws.com/right_s3_awesome_test_bucket_000B1_officedrop/test%2Flarge_multipart_file", :e_tag=>"\"72b81ac08aed4d4d1055c11f56c2a258-1\"", :key=>"test/large_multipart_file", :bucket=>"right_s3_awesome_test_bucket_000B1_officedrop"}
# File lib/s3/right_s3_interface.rb, line 555 def store_object_multipart(params) AwsUtils.allow_only([:bucket, :key, :data, :headers, :part_size, :retry_count], params) AwsUtils.mandatory_arguments([:bucket, :key, :data], params) params[:headers] = {} unless params[:headers] params[:data].binmode if(params[:data].respond_to?(:binmode)) # On Windows, if someone opens a file in text mode, we must reset it to binary mode for streaming to work properly # detect whether we are using straight read or converting to string first unless(params[:data].respond_to?(:read)) params[:data] = StringIO.new(params[:data].to_s) end # make sure part size is > 5 MB minimum params[:part_size] ||= MINIMUM_PART_SIZE if params[:part_size] < MINIMUM_PART_SIZE raise AwsError.new("Part size for a multipart upload must be greater than or equal to #{5 * 1024 * 1024} bytes. #{params[:part_size]} bytes was provided.") end # make sure retry_count is positive params[:retry_count] ||= DEFAULT_RETRY_COUNT if params[:retry_count] < 0 raise AwsError.new("Retry count must be positive. #{params[:retry_count]} bytes was provided.") end # Set 100-continue for large part sizes if (params[:part_size] >= USE_100_CONTINUE_PUT_SIZE) params[:headers]['expect'] = '100-continue' end # initiate upload initiate_hash = generate_rest_request('POST', params[:headers].merge(:url=>"#{params[:bucket]}/#{CGI::escape params[:key]}?uploads")) initiate_resp = request_info(initiate_hash, S3MultipartUploadInitiateResponseParser.new) upload_id = initiate_resp[:upload_id] # split into parts and upload each one, re-trying if necessary # upload occurs serially at this time. part_etags = [] part_data = "" index = 1 until params[:data].eof? part_data = params[:data].read(params[:part_size]) unless part_data.size == 0 retry_attempts = 1 while true begin send_part_hash = generate_rest_request('PUT', params[:headers].merge({ :url=>"#{params[:bucket]}/#{CGI::escape params[:key]}?partNumber=#{index}&uploadId=#{upload_id}", :data=>part_data } )) send_part_resp = request_info(send_part_hash, S3HttpResponseHeadParser.new) part_etags << {:part_num => index, :etag => send_part_resp['etag']} index += 1 break # successful, can move to next part rescue AwsError => e if retry_attempts >= params[:retry_count] raise e else #Hit an error attempting to transmit part, retry until retry_attemts have been exhausted retry_attempts += 1 end end end end end # assemble complete upload message complete_body = "<CompleteMultipartUpload>" part_etags.each do |part_hash| complete_body << "<Part><PartNumber>#{part_hash[:part_num]}</PartNumber><ETag>#{part_hash[:etag]}</ETag></Part>" end complete_body << "</CompleteMultipartUpload>" complete_req_hash = generate_rest_request('POST', {:url=>"#{params[:bucket]}/#{CGI::escape params[:key]}?uploadId=#{upload_id}", :data => complete_body}) return request_info(complete_req_hash, S3CompleteMultipartParser.new) rescue on_exception end
Generated with the Darkfish Rdoc Generator 2.