Java Client SDK 3.10

Blob

Updated: July 18, 2024

This section is about how handling blobs with Nuxeo Java Client.

There are several ways in Nuxeo and so in Java Client to handle blobs.

Blob Interface

In Java Client, we use Blob as the main interface to deal with blobs with server. In this page we mainly use FileBlob for clarity nevertheless every example are also working with StreamBlob.

For both implementation, we highly recommend to give mimeType if possible. This allows less processing for Nuxeo Server.

Batch Upload

The batch upload manager leverage batch upload endpoint server side. It is designed to upload a batch of files that can then be used in Automation or as a property of a document through the REST API. It also supports the chunk mode if manager is configured to use it.

Let's first retrieve batch upload manager to handle our batch:

BatchUploadManager batchUploadManager = client.batchUploadManager();

Attach a Blob to a Document Using Automation

We're assuming /file001 already exists. We're going to upload a blob with batch upload manager and then attach it to file001.

// create a new batch
BatchUpload batchUpload = batchUploadManager.createBatch();
Blob fileBlob = new FileBlob(file);
batchUpload = batchUpload.upload("0", fileBlob);

// attach the blob
batchUpload.operation(Operations.BLOB_ATTACH_ON_DOCUMENT)
           .param("document", "/file001") // document to attach
           .execute();

Attach a Blob to a Document Using REST API

We're assuming /file001 already exists. We're gonna upload a blob with batch upload manager and then attach it to file001.

// create a new batch
BatchUpload batchUpload = batchUploadManager.createBatch();
Blob fileBlob = new FileBlob(file);
batchUpload = batchUpload.upload("0", fileBlob);

// attach the blob
Document doc = nuxeoClient.repository().fetchDocumentByPath("/file001");
doc.setPropertyValue("file:content", batchUpload.getBatchBlob());
doc = doc.updateDocument();

Create a Document from a Blob

We're gonna upload a blob with batch upload manager and then create a document referencing this blob.

// create a new batch
BatchUpload batchUpload = batchUploadManager.createBatch();
Blob fileBlob = new FileBlob(file);
batchUpload = batchUpload.upload("0", fileBlob);

// create document with the blob
Document doc = Document.createWithName("file002", "File");
doc.setPropertyValue("file:content", batchUpload.getBatchBlob());
doc = nuxeoClient.repository().createDocumentByPath("/", doc);

Automation (Multipart Upload)

You can upload blob when using automation. This will upload the blob using HTTP multipart.

Attach a Blob to a Document

We're assuming /file001 already exists. We're gonna upload and attach several blobs to file001 using Blob.AttachOnDocument.

nuxeoClient.operation(Operations.BLOB_ATTACH_ON_DOCUMENT)
           .voidOperation(true) // allows to not download blob in response
           .param("document", "/file001") // document to attach
           .param("xpath", "files:files") // xpath to store blobs
           .input(new Blobs(Arrays.asList(new FileBlob(file1), new FileBlob(file2))))
           .execute();

Create a Document from a Blob

We're gonna upload, create file002 and attach a blob using FileManager.Import:

Document file = nuxeoClient.operation("FileManager.Import")
                           .context("currentDocument", "/") // parent document
                           .input(new FileBlob(file1, "file002"))
                           .execute();

Download Blob

APIs below are available on repository manager to fetch blob:

  • streamBlobById which takes a document id and a xpath
  • streamBlobByPath which takes a document path and a xpath

You also have streamBlob on Document when object is connected.

In both case, you get a StreamBlob exposing an InputStream which must be closed because it is directly linked to the HTTP connection.

StreamBlob blob = nuxeoClient.repository().streamBlobByPath("/file001", "file:content");
try (InputStream is = blob.getStream()) {
    String content = org.apache.commons.io.IOUtils.toString(is, StandardCharsets.UTF_8);
}