(Ruby) S3 Upload the Parts for a Multipart Upload

This example uploads a large file in parts. The multipart upload needs to have been first initiated prior to uploading the parts.

See http://docs.aws.amazon.com/AmazonS3/latest/API/mpUploadUploadPart.html for more information about uploading parts.

Chilkat Ruby Downloads install from rubygems.org gem install chilkat or download... Ruby Library for Windows, MacOS, Linux, Alpine Linux

require 'chilkat'

# In the 1st step for uploading a large file, the multipart upload was initiated
# as shown here: Initiate Multipart Upload

# Other S3 Multipart Upload Examples:
# Complete Multipart Upload
# Abort Multipart Upload
# List Parts

# When we initiated the multipart upload, we saved the XML response to a file.  This
# XML response contains the UploadId.  We'll begin by loading that XML and getting
# the Upload ID.

xmlInit = Chilkat::CkXml.new()
success = xmlInit.LoadXmlFile("s3_multipart_uploads/initiate.xml")
if (success != true)
    print "Did not find the initiate.xml XML file." + "\n";
    exit
end

uploadId = xmlInit.getChildContent("UploadId")
print "UploadId = " + uploadId + "\n";

# When uploading parts, we need to keep an XML record of each part number
# and its corresponding ETag, which is received in the response for each part.
# There can be up to 10000 parts, numbered 1 to 10000.  
# After all parts have been uploaded, the final step will be to complete
# the multipart upload (see Complete Multipart Upload)

# In this example, the large file we want to upload is somethingBig.zip
fileToUploadPath = "s3_multipart_uploads/somethingBig.zip"

# The minimum allowed part size is 5MB (5242880 bytes).  The last part can be smaller because
# it will contain the remainder of the file.  (This minimum is enforced by the AWS service.)
# We'll use the minimum allowed part size for this example.
partSize = 5242880

# Let's use Chilkat's FileAccess API to examine the file to be uploaded.  We'll get the size
# of the file and find out how many parts will be needed, including the final "partial" part.
fac = Chilkat::CkFileAccess.new()
fac.OpenForRead(fileToUploadPath)

# How many parts will there be if each part is 5242880 bytes?
numParts = fac.GetNumBlocks(partSize)
print "numParts = " + numParts.to_s() + "\n";
fac.FileClose()

# Imagine that we may be running this for the 1st time, or maybe we already
# attempted to upload parts, and something failed. Maybe there was a network problem
# the resulted in not all parts getting uploaded.  We'll write this code so that if run again,
# it will upload whatever parts haven't yet been uploaded.

# We'll keep a partsList.xml file to record the parts that have already been successfully
# uploaded.  If this file does not yet exist, we'll create it..
partsListFile = "s3_multipart_uploads/partsList.xml"
partsListXml = Chilkat::CkXml.new()
if (fac.FileExists(partsListFile) == true)
    partsListXml.LoadXmlFile(partsListFile)
end

# Make sure the top-level tag is "CompleteMultipartUpload"
partsListXml.put_Tag("CompleteMultipartUpload")

# --------------------------------------
# Before entering the loop to upload parts,
# setup the REST object with AWS authentication,
# and make the initial connection.
rest = Chilkat::CkRest.new()

# Connect to the Amazon AWS REST server.
bTls = true
port = 443
bAutoReconnect = true
success = rest.Connect("s3.amazonaws.com",port,bTls,bAutoReconnect)

# ----------------------------------------------------------------------------
# Important: For buckets created in regions outside us-east-1,
# there are three important changes that need to be made.
# See Working with S3 Buckets in Non-us-east-1 Regions for the details.
# ----------------------------------------------------------------------------

# Provide AWS credentials for the REST call.
authAws = Chilkat::CkAuthAws.new()
authAws.put_AccessKey("AWS_ACCESS_KEY")
authAws.put_SecretKey("AWS_SECRET_KEY")
authAws.put_ServiceName("s3")
success = rest.SetAuthAws(authAws)

# Set the bucket name via the HOST header.
# In this case, the bucket name is "chilkat100".
rest.put_Host("chilkat100.s3.amazonaws.com")
# --------------------------------------

partNumber = 1
sbPartNumber = Chilkat::CkStringBuilder.new()

while (partNumber <= numParts)
    print "---- " + partNumber.to_s() + " ----" + "\n";

    # This cumbersome way of converting an integer to a string is because
    # Chilkat examples are written in a script that is converted to many programming languages.
    # At this time, the translator does not have integer-to-string code generation capability..
    sbPartNumber.Clear()
    sbPartNumber.AppendInt(partNumber)

    bPartAlreadyUploaded = false

    # If there are no children, then the XML is empty and no parts have yet been uploaded.
    numUploadedParts = partsListXml.get_NumChildren()
    if (numUploadedParts > 0)
        # If some parts have been uploaded, check to see if this particular part was already upload.
        # If so, then it can be skipped.

        # Position ourselves at the 1st record.
        # xRec0 is a CkXml
        xRec0 = partsListXml.GetChild(0)
        # foundRec is a CkXml
        foundRec = xRec0.FindNextRecord("PartNumber",sbPartNumber.getAsString())
        if (xRec0.get_LastMethodSuccess() == true)
            bPartAlreadyUploaded = true
            print "Part " + partNumber.to_s() + " was previously uploaded." + "\n";
            print foundRec.getXml() + "\n";

        end

    end

    # If this part was not already uploaded, we need to upload.
    # Also update the partsListXml and save as each part is successfully uploaded.
    if (bPartAlreadyUploaded == false)
        print "Uploading part " + partNumber.to_s() + " ..." + "\n";

        # Setup the stream source for the large file to be uploaded..
        fileStream = Chilkat::CkStream.new()
        fileStream.put_SourceFile(fileToUploadPath)
        # The Chilkat Stream API has features to make uploading a parts
        # of a file easy.  Indicate the part size by setting the SourceFilePartSize
        # property.
        fileStream.put_SourceFilePartSize(partSize)

        # Our HTTP start line to upload a part will look like this:
        # PUT /ObjectName?partNumber=PartNumber&uploadId=UploadId HTTP/1.1

        # Set the query params.  We'll need partNumber and uploadId.
        # Make sure the query params from previous iterations are clear.
        rest.ClearAllQueryParams()
        rest.AddQueryParam("partNumber",sbPartNumber.getAsString())
        rest.AddQueryParam("uploadId",uploadId)

        # Upload this particular file part.
        # Tell the fileStream which part is being uploaded.
        # Our partNumber is 1-based (the 1st part is at index 1), but the fileStream's SourceFilePart
        # property is 0-based.  Therefore we use partNumber-1.
        fileStream.put_SourceFilePart(partNumber - 1)

        # Because the SourceFilePart and SourceFilePartSize properties are set, the stream will 
        # will provide just that part of the file.  
        responseStr = rest.fullRequestStream("PUT","/somethingBig.zip",fileStream)
        if (rest.get_LastMethodSuccess() != true)
            print rest.lastErrorText() + "\n";
            exit
        end

        if (rest.get_ResponseStatusCode() != 200)
            # Examine the request/response to see what happened.
            print "response status code = " + rest.get_ResponseStatusCode().to_s() + "\n";
            print "response status text = " + rest.responseStatusText() + "\n";
            print "response header: " + rest.responseHeader() + "\n";
            print "response body: " + responseStr + "\n";
            print "---" + "\n";
            print "LastRequestStartLine: " + rest.lastRequestStartLine() + "\n";
            print "LastRequestHeader: " + rest.lastRequestHeader() + "\n";
            exit
        end

        # OK, this part was uploaded..
        # The response will have a 0-length body.  The only information we need is the 
        # ETag response header field.
        etag = rest.responseHdrByName("ETag")
        # It should be present, but just in case there was no ETag header...
        if (rest.get_LastMethodSuccess() != true)
            print "No ETag response header found!" + "\n";
            print "response header: " + rest.responseHeader() + "\n";
            exit
        end

        # We need to add record to the partsListXml.
        # The record will look like this:
        # &lt;Part>
        #   &lt;PartNumber>PartNumber&lt;/PartNumber>
        #   &lt;ETag>ETag&lt;/ETag>
        # &lt;/Part>
        # xPart is a CkXml
        xPart = partsListXml.NewChild("Part","")
        xPart.NewChildInt2("PartNumber",partNumber)
        xPart.NewChild2("ETag",etag)

        success = partsListXml.SaveXml(partsListFile)
        if (success != true)
            print partsListXml.lastErrorText() + "\n";
            exit
        end

        print "-- Part " + partNumber.to_s() + " uploaded. ---------------------" + "\n";
    end

    partNumber = partNumber + 1
end

print "Finished.  All parts uploaded." + "\n";