The idea is to build a proprietary Java back end document system using Office Web Apps.
We have created the WOPI client which allows us to view/edit PowerPoint and Excel web app documents but we can only view Word Documents.
In order to edit Word Web App documents you need to implement MS-FSSHTTP.
It appears there is no information about how to actually do this in code. Has anyone performed this or would know how?
recently my team and I have implemented a WOPI-Host that supports viewing and editing of Word, PPT and Excel documents. You can take a look at https://github.com/marx-yu/WopiHost which is a command prompt project that listens on the 8080 port and enables editing and viewing of word documents though the Microsoft Office Web Apps.
We have implemented this solution in a webApi and it works great. Hope this sample project will help you out.
After requested, I will try and add code samples to clarify the way to implement it based on my webApi implementation, but their is a lot of code to implement to actually make it work properly.
First things first, to enabled editing you will need to capture Http Posts in a FilesController. Each posts that concern the actual editing will have the header X-WOPI-Override equal to COBALT. In these post you will find out that the InputStream is and Atom type. Based on the MS-WOPI documentation, in your response you will need to include the following headers X-WOPI-CorrelationID and request-id.
Here is the code of my webApi post method (it is not complete since I'm still implementing that WOPI protocol).
string wopiOverride = Request.Headers.GetValues("X-WOPI-Override").First();
if (wopiOverride.Equals("COBALT"))
{
string filename = name;
EditSession editSession = CobaltSessionManager.Instance.GetSession(filename);
var filePath = HostingEnvironment.MapPath("~/App_Data/");
if (editSession == null){
var fileExt = filename.Substring(filename.LastIndexOf('.') + 1);
if (fileExt.ToLower().Equals(#"xlsx"))
editSession = new FileSession(filename, filePath + "/" + filename, #"yonggui.yu", #"yuyg", #"yonggui.yu#emacle.com", false);
else
editSession = new CobaltSession(filename, filePath + "/" + filename, #"patrick.racicot", #"Patrick Racicot", #"patrick.racicot#hospitalis.com", false);
CobaltSessionManager.Instance.AddSession(editSession);
}
//cobalt, for docx and pptx
var ms = new MemoryStream();
HttpContext.Current.Request.InputStream.CopyTo(ms);
AtomFromByteArray atomRequest = new AtomFromByteArray(ms.ToArray());
RequestBatch requestBatch = new RequestBatch();
Object ctx;
ProtocolVersion protocolVersion;
requestBatch.DeserializeInputFromProtocol(atomRequest, out ctx, out protocolVersion);
editSession.ExecuteRequestBatch(requestBatch);
foreach (Request request in requestBatch.Requests)
{
if (request.GetType() == typeof(PutChangesRequest) && request.PartitionId == FilePartitionId.Content)
{
//upload file to hdfs
editSession.Save();
}
}
var responseContent = requestBatch.SerializeOutputToProtocol(protocolVersion);
var host = Request.Headers.GetValues("Host");
var correlationID = Request.Headers.GetValues("X-WOPI-CorrelationID").First();
response.Headers.Add("X-WOPI-CorrelationID", correlationID);
response.Headers.Add("request-id", correlationID);
MemoryStream memoryStream = new MemoryStream();
var streamContent = new PushStreamContent((outputStream, httpContext, transportContent) =>
{
responseContent.CopyTo(outputStream);
outputStream.Close();
});
response.Content = streamContent;
response.Content.Headers.ContentType = new MediaTypeHeaderValue("application/octet-stream");
response.Content.Headers.ContentLength = responseContent.Length;
}
As you can see in this method I make use of CobaltSessionManager and CobaltSession which are used to create and manage editing sessions on the Cobalt protocol. You will also need a what I call CobaltHostLockingStore which is used to handle the different requests when communicating with the Office Web App server in the edition initialization.
I won't be posting the code for these 3 classes since they are already coded in the sample github project I posted and that they are fairly simple to understand even though they are big.
If you have more questions or if it's not clear enough don't hesitate to comment and I will update my post accordingly.
Patrick Racicot, provided great answer. But i had problem saving docx(exception in CobaltCore.dll), and i even started using dotPeak reflector trying to figure it out.
But after i locked editSession variable in my WebApi method everything started working like magic. It seems that OWA is sending requests that should be handled as a chain, not in parallel as usually controller method acts.
Related
I want to set up the REST API to support file downloads via Java (The java part is not needed at the moment -- I am saying it in here so you can make your answer more specific for my problem).
How would I do that?
For example, I have this file in a folder (./java.jar), how can I stream it in such a way for it to be downloadable by a Java client?
I forgot to say that this, is for some paid-content.
My app should be able to do this
Client: Post to server with username,pass.
Rest: Respond accordingly to what user has bought (so if it has bought that file, download it)
Client: Download file and put it in x folder.
I thought of encoding a file in base64 and then posting the encoded result into the usual .json (maybe with a nice name -- useful for the java application, and with the code inside -- though I would not know how I should rebuild the file at this point). <- Is this plausible? Or is there an easier way?
Also, please do not downvote if unnecessary, although there is no code in the question, that doesn't mean I haven't researched it, it just means that I found nothing suitable for my situation.
Thanks.
What you need is a regular file streaming, using a valid URL.
Below code is an excerpt from here
import java.net.*;
import java.io.*;
public class URLReader {
public static void main(String[] args) throws Exception {
URL oracle = new URL("http://www.oracle.com/");
BufferedReader in = new BufferedReader(
new InputStreamReader(oracle.openStream()));
String inputLine;
while ((inputLine = in.readLine()) != null)
System.out.println(inputLine);
in.close();
}
}
For your needs, based on your updated comments on the above answer, you could call your REST endpoint after user logs in(with Auth and other headers/body you wish to receive) and proceed to the download.
Convert your jar/downloadable content to bytes. More on this
Java Convert File to Byte Array and visa versa
Later, in case if you dont want regular streaming as aforementioned in previous answers, you can put the byte content in the body as Base64 String. You can encode to Base64 from your byte array using something like below.
Base64.encodeToString(byte[], Base64.NO_WRAP + Base64.URL_SAFE);
Reference from here: How to send byte[] and strings to a restful webservice and retrieve this information in the web method implementation
Again, there are many ways to do this, this is one of the ways you can probably do using REST.
Final update (please see comments)
I have given up on implementing my own QR generator due to GAE limitations. The former Google Charts API services still can be used, and also replicated using ZXing's servlet. For more details, see this link. Thank you for the answers.
Update 2 (see original question below)
So I dug into the source of ZXing and they seem to use BufferedImage in all of their processes. My question now could be phrased as:
Is there any way to use ZXing with Google App Engine?
Is there any way to generate a QR code in a servlet that could be
deployed to Google App Engine?
Update (see original question below)
The following line causes the error apparently:
MatrixToImageWriter.writeToStream(encoded, "png", outs);
It seems to be using BufferedImage. Is it possible to convert a BitMatrix to a byte[] without doing this step?
Original question
I am trying to create a servlet on Google App Engine to return a QR code with given parameters. So far I have created solutions both with QRGen and ZXing which work perfectly when testing in App Engine local development mode. My problem is that both of these implementations fail after I deploy my servlet to App Engine, saying either
Could not initialize class
com.google.apphosting.runtime.security.shared.stub.java.awt.image.BufferedImage
or
java.awt.Image is a restricted class. Please see the Google App Engine
developer's guide for more details.
I don't understand completely how these tools work internally, what I know that java.awt.image classes are not on the Class whitelist. My question is that is there any way to get around this security feature, and actually return a QR code from a servlet. Here is my latest implementation (which works fine in development mode, but also fails on a deployed servlet) using ZXing (it seems to offer more options than QRGen):
#Override
protected void doGet(HttpServletRequest req, HttpServletResponse resp) throws ServletException, IOException
{
String returnString = "myString";
Writer writer = new QRCodeWriter();
BitMatrix encoded = null;
try
{
encoded = writer.encode(returnString, BarcodeFormat.QR_CODE, 300, 300);
}
catch (WriterException ex)
{
Logger.getLogger(QRService.class.getName()).log(Level.SEVERE, null, ex);
}
ByteArrayOutputStream outs = new ByteArrayOutputStream();
MatrixToImageWriter.writeToStream(encoded, "png", outs);
byte[] out = outs.toByteArray();
Blob qrImage = new Blob(out);
resp.setContentType("image/jpeg");
resp.getOutputStream().write(qrImage.getBytes());
}
Here, look into the Google Charts API
https://developers.google.com/chart/infographics/docs/qr_codes
It seems it allows you to create a QR code with simple HTTP GET request. To create an HTTP GET request in App Engine, try this:
https://developers.google.com/appengine/docs/java/urlfetch/
Im really struggling with the documentation for PDFBox. For such a popular library info seems to be a little thin on the ground (for me!).
Anyway the problem Im having relates to protecting the PDF. At the moment all I want is to control the access permissions of the users. specifically I want to prevent the user from being able to modify the PDF.
If I omit the access permission code everything works perfectly. I am reading in a PDF from an external resource. I am then reading and populating the fields, adding some images before saving the new PDF. That all works perfectly.
The problem comes when I add the following code to manage the access:
/* Secure the PDF so that it cannot be edited */
try {
String ownerPassword = "DSTE$gewRges43";
String userPassword = "";
AccessPermission ap = new AccessPermission();
ap.setCanModify(false);
StandardProtectionPolicy spp = new StandardProtectionPolicy(ownerPassword, userPassword, ap);
pdf.protect(spp);
} catch (BadSecurityHandlerException ex) {
Logger.getLogger(PDFManager.class.getName()).log(Level.SEVERE, null, ex);
}
When I add this code, all the text and images are striped from the outgoing pdf. The fields are still present in the document but they are all empty and all the text and images that where part of the original PDF and that were added dynamically in the code are gone.
UPDATE:
Ok, as best as I can tell the problem is coming from a bug relating to the form fields. I'm going to try a different approach without the form fields and see what it gives.
I found the solution to this problem. It would appear that if the PDF comes from an external source, sometimes the PDF is protected or encrypted.
If you get a blank output when loading up a PDF document from an external source and add protections you are probably working with an encrypted document. I have a stream processing system working on PDF documents. So the following code works for me. If you are just working with PDF inputs then you could integrate the below code with your flow.
public InputStream convertDocument(InputStream dataStream) throws Exception {
// just acts as a pass through since already in pdf format
PipedOutputStream os = new PipedOutputStream();
PipedInputStream is = new PipedInputStream(os);
System.setProperty("org.apache.pdfbox.baseParser.pushBackSize", "2024768"); //for large files
PDDocument doc = PDDocument.load(dataStream, true);
if (doc.isEncrypted()) { //remove the security before adding protections
doc.decrypt("");
doc.setAllSecurityToBeRemoved(true);
}
doc.save(os);
doc.close();
dataStream.close();
os.close();
return is;
}
Now take that returned InputStream and use it for your security application;
PipedOutputStream os = new PipedOutputStream();
PipedInputStream is = new PipedInputStream(os);
System.setProperty("org.apache.pdfbox.baseParser.pushBackSize", "2024768");
InputStream dataStream = secureData.data();
PDDocument doc = PDDocument.load(dataStream, true);
AccessPermission ap = new AccessPermission();
//add what ever perms you need blah blah...
ap.setCanModify(false);
ap.setCanExtractContent(false);
ap.setCanPrint(false);
ap.setCanPrintDegraded(false);
ap.setReadOnly();
StandardProtectionPolicy spp = new StandardProtectionPolicy(UUID.randomUUID().toString(), "", ap);
doc.protect(spp);
doc.save(os);
doc.close();
dataStream.close();
os.close();
Now this should return a proper document with no blank output!
Trick is to remove encryption first!
I'm working on a way of programatically accessing a Lotus Notes database to gather information on embedded attachments of records over a given period.
My goal is to find records over a given period, then use Apache-POI to get metadata about document size, character count, etc.
The POI part works fine, and so far, I've been able to access the Lotus Notes records thanks to this help:
lotus notes search by date with Java api
and this answer also shows me how to download/copy the attachments:
How do I get all the attachments from a .nsf(lotus notes) file using java
from there I could use my POI code do my job and at the end, just delete the copied attachments. This approach, basically works, but I want to avoid the overhead of copying, saving and then at the end deleting my copy of these attached documents from the database.
I tried passing the result of the EmbeddedObject getSource() method as an input to my POI code and got a FileNotFoundException in the POI code that was expecting a String to make a File.
Is there a way of getting a File reference I can pass to POI, without copying and saving the attachment? Or, what I mean is, is it as simple as getting a File (+path) for the Lotus Notes EmbeddedObject attachment, and how do I do this?
I found the answer and posted it below.
Answering my own question...
...here's the solution I found a little while after posting the question above:
EmbeddedObject's getInputStream to the rescue...
//from the answer in the link in the question above　
Database db = agentContext.getCurrentDatabase();
DocumentCollection dc = db.getAllDocuments();
Document doc = dc.getFirstDocument();
boolean saveFlag = false;
while (doc != null) {
RichTextItem body =
(RichTextItem)doc.getFirstItem("Body");
System.out.println(doc.getItemValueString("Subject"));
Vector v = body.getEmbeddedObjects();
Enumeration e = embeddedObjs.elements();
while(e.hasMoreElements()){
EmbeddedObject eo = (EmbeddedObject)e.nextElement();
if(eo.getType() == EmbeddedObject.EMBED_ATTACHMENT){
//this next line gives Apache-POI access to the InputStream
InputStream is = eo.getInputStream();
POIFSFileSystem POIfs =
HWPFDocument.verifyAndBuildPOIFS(is);
POIOLE2TextExtractor extractor =
ExtractorFactory.createExtractor(POIfs);
System.out.println("extracted text: " + extractor.getText());
is.close(); //closing InputStream
}
eo.recycle(); //recycling EmbeddedObject
//thanks to rhsatrhs for the close() and recycle() tip!
[edit]
I've removed my convoluted and badly malformed question so that it doesn't detract from the very neat and correct answer beneath. Given the (surprising) difficulty of finding an on-line example for doing this incredibly common task, I hope Yoni gets a few more up-ticks for his response.
So... the question in a nutshell...
How do I use Apache.Commons to upload a file to some destination. I'm using it in Android and uploading to a PHP script, but obviously it can work from any Java program and to any HTTP based listener.
From the api of MultipartRequestEntity:
File f = new File("/path/fileToUpload.txt");
PostMethod filePost = new PostMethod("http://host/some_path");
Part[] parts = {
new StringPart("param_name", "value"),
new FilePart(f.getName(), f)
};
filePost.setRequestEntity(
new MultipartRequestEntity(parts, filePost.getParams())
);
HttpClient client = new HttpClient();
int status = client.executeMethod(filePost);
I don't think you need the content-disposition part, that is used for the other direction (when the browser downloads a file and needs to know what to do with it).
getParams.setParameter is optional. You can also set it directly on the HttpClient instance.
AFAIK, the order of setting request headers is irrelevant, as long as they are all set before you set the request body.