Example usage for org.apache.poi.hmef Attachment getExtension

Introduction

In this page you can find the example usage for org.apache.poi.hmef Attachment getExtension.

Prototype

public String getExtension() 

Source Link

Document

Returns the file extension

Usage

From source file:mj.ocraptor.extraction.tika.parser.microsoft.TNEFParser.java

License:Apache License

/**
 * Extracts properties and text from an MS Document input stream
 *//*ww w  .  j ava 2s. c  o  m*/
public void parse(InputStream stream, ContentHandler handler, Metadata metadata, ParseContext context)
        throws IOException, SAXException, TikaException {

    // We work by recursing, so get the appropriate bits 
    EmbeddedDocumentExtractor ex = context.get(EmbeddedDocumentExtractor.class);
    EmbeddedDocumentExtractor embeddedExtractor;
    if (ex == null) {
        embeddedExtractor = new ParsingEmbeddedDocumentExtractor(context);
    } else {
        embeddedExtractor = ex;
    }

    // Ask POI to process the file for us
    HMEFMessage msg = new HMEFMessage(stream);

    // Set the message subject if known
    String subject = msg.getSubject();
    if (subject != null && subject.length() > 0) {
        // TODO: Move to title in Tika 2.0
        metadata.set(TikaCoreProperties.TRANSITION_SUBJECT_TO_DC_TITLE, subject);
    }

    // Recurse into the message body RTF
    MAPIAttribute attr = msg.getMessageMAPIAttribute(MAPIProperty.RTF_COMPRESSED);
    if (attr != null && attr instanceof MAPIRtfAttribute) {
        MAPIRtfAttribute rtf = (MAPIRtfAttribute) attr;
        handleEmbedded("message.rtf", "application/rtf", rtf.getData(), embeddedExtractor, handler);
    }

    // Recurse into each attachment in turn
    for (Attachment attachment : msg.getAttachments()) {
        String name = attachment.getLongFilename();
        if (name == null || name.length() == 0) {
            name = attachment.getFilename();
        }
        if (name == null || name.length() == 0) {
            String ext = attachment.getExtension();
            if (ext != null) {
                name = "unknown" + ext;
            }
        }
        handleEmbedded(name, null, attachment.getContents(), embeddedExtractor, handler);
    }
}

---

From source file:org.alfresco.repo.imap.AttachmentsExtractor.java

License:Open Source License

/**
 * Create an attachment given a mime part
 * //from  www  .  j  a v a2  s  .  c om
 * @param messageFile the file containing the message
 * @param attachmentsFolderRef where to put the attachment
 * @param part the mime part
 * @throws MessagingException
 * @throws IOException
 */
private void createAttachment(NodeRef messageFile, NodeRef attachmentsFolderRef, Part part)
        throws MessagingException, IOException {
    String fileName = part.getFileName();
    if (fileName == null || fileName.isEmpty()) {
        fileName = "unnamed";
    }
    try {
        fileName = MimeUtility.decodeText(fileName);
    } catch (UnsupportedEncodingException e) {
        if (logger.isWarnEnabled()) {
            logger.warn("Cannot decode file name '" + fileName + "'", e);
        }
    }

    ContentType contentType = new ContentType(part.getContentType());

    if (contentType.getBaseType().equalsIgnoreCase("application/ms-tnef")) {
        // The content is TNEF
        HMEFMessage hmef = new HMEFMessage(part.getInputStream());

        // hmef.getBody();
        List<org.apache.poi.hmef.Attachment> attachments = hmef.getAttachments();
        for (org.apache.poi.hmef.Attachment attachment : attachments) {
            String subName = attachment.getLongFilename();

            NodeRef attachmentNode = fileFolderService.searchSimple(attachmentsFolderRef, subName);
            if (attachmentNode == null) {
                /*
                 * If the node with the given name does not already exist Create the content node to contain the attachment
                 */
                FileInfo createdFile = fileFolderService.create(attachmentsFolderRef, subName,
                        ContentModel.TYPE_CONTENT);

                attachmentNode = createdFile.getNodeRef();

                serviceRegistry.getNodeService().createAssociation(messageFile, attachmentNode,
                        ImapModel.ASSOC_IMAP_ATTACHMENT);

                byte[] bytes = attachment.getContents();
                ContentWriter writer = fileFolderService.getWriter(attachmentNode);

                // TODO ENCODING - attachment.getAttribute(TNEFProperty.);
                String extension = attachment.getExtension();
                String mimetype = mimetypeService.getMimetype(extension);
                if (mimetype != null) {
                    writer.setMimetype(mimetype);
                }

                OutputStream os = writer.getContentOutputStream();
                ByteArrayInputStream is = new ByteArrayInputStream(bytes);
                FileCopyUtils.copy(is, os);
            }
        }
    } else {
        // not TNEF
        NodeRef attachmentFile = fileFolderService.searchSimple(attachmentsFolderRef, fileName);
        // The one possible behaviour
        /*
         * if (result.size() > 0) { for (FileInfo fi : result) { fileFolderService.delete(fi.getNodeRef()); } }
         */
        // And another one behaviour which will overwrite the content of the existing file. It is performance preferable.
        if (attachmentFile == null) {
            FileInfo createdFile = fileFolderService.create(attachmentsFolderRef, fileName,
                    ContentModel.TYPE_CONTENT);
            nodeService.createAssociation(messageFile, createdFile.getNodeRef(),
                    ImapModel.ASSOC_IMAP_ATTACHMENT);
            attachmentFile = createdFile.getNodeRef();
        } else {

            String newFileName = imapService.generateUniqueFilename(attachmentsFolderRef, fileName);

            FileInfo createdFile = fileFolderService.create(attachmentsFolderRef, newFileName,
                    ContentModel.TYPE_CONTENT);
            nodeService.createAssociation(messageFile, createdFile.getNodeRef(),
                    ImapModel.ASSOC_IMAP_ATTACHMENT);
            attachmentFile = createdFile.getNodeRef();

        }

        nodeService.setProperty(attachmentFile, ContentModel.PROP_DESCRIPTION,
                nodeService.getProperty(messageFile, ContentModel.PROP_NAME));

        ContentWriter writer = fileFolderService.getWriter(attachmentFile);
        writer.setMimetype(contentType.getBaseType());
        OutputStream os = writer.getContentOutputStream();
        FileCopyUtils.copy(part.getInputStream(), os);
    }
}

---

From source file:org.apache.tika.parser.microsoft.TNEFParser.java

License:Apache License

/**
 * Extracts properties and text from an MS Document input stream
 *///www. j a v a2 s.  c  o m
public void parse(InputStream stream, ContentHandler handler, Metadata metadata, ParseContext context)
        throws IOException, SAXException, TikaException {

    // We work by recursing, so get the appropriate bits
    EmbeddedDocumentExtractor ex = context.get(EmbeddedDocumentExtractor.class);
    EmbeddedDocumentExtractor embeddedExtractor;
    if (ex == null) {
        embeddedExtractor = new ParsingEmbeddedDocumentExtractor(context);
    } else {
        embeddedExtractor = ex;
    }

    // Ask POI to process the file for us
    HMEFMessage msg = new HMEFMessage(stream);

    // Set the message subject if known
    String subject = msg.getSubject();
    if (subject != null && subject.length() > 0) {
        // TODO: Move to title in Tika 2.0
        metadata.set(TikaCoreProperties.TRANSITION_SUBJECT_TO_DC_TITLE, subject);
    }

    // Recurse into the message body RTF
    MAPIAttribute attr = msg.getMessageMAPIAttribute(MAPIProperty.RTF_COMPRESSED);
    if (attr != null && attr instanceof MAPIRtfAttribute) {
        MAPIRtfAttribute rtf = (MAPIRtfAttribute) attr;
        handleEmbedded("message.rtf", "application/rtf", rtf.getData(), embeddedExtractor, handler);
    }

    // Recurse into each attachment in turn
    for (Attachment attachment : msg.getAttachments()) {
        String name = attachment.getLongFilename();
        if (name == null || name.length() == 0) {
            name = attachment.getFilename();
        }
        if (name == null || name.length() == 0) {
            String ext = attachment.getExtension();
            if (ext != null) {
                name = "unknown" + ext;
            }
        }
        handleEmbedded(name, null, attachment.getContents(), embeddedExtractor, handler);
    }
}

---