Example usage for org.apache.poi.hpsf SummaryInformation getWordCount

List of usage examples for org.apache.poi.hpsf SummaryInformation getWordCount

  1. HOME
  2. Java
  3. org.apache.poi
  4. org.apache.poi.hpsf.*
  5. SummaryInformation
  6. getWordCount

Introduction

In this page you can find the example usage for org.apache.poi.hpsf SummaryInformation getWordCount.

Prototype

public int getWordCount()

Source Link

Document

Returns the word count or 0 if the SummaryInformation does not contain a word count.

Usage

From source file:com.flexive.extractor.FxSummaryInformation.java

License:Open Source License

/**
 * Constructor.//from   w  w  w  .  j  a  v a  2  s.  c o m
 *
 * @param si the summary information
 */
public FxSummaryInformation(SummaryInformation si) {
    author = si.getAuthor();
    applicationName = si.getApplicationName();
    charCount = si.getCharCount();
    comments = si.getComments();
    createdAt = si.getCreateDateTime();
    editTime = new Date(si.getEditTime());
    keywords = si.getKeywords();
    lastModifiedBy = si.getLastAuthor();
    lastPrintedAt = si.getLastPrinted();
    title = si.getTitle();
    lastModifiedAt = si.getLastSaveDateTime();
    pageCount = si.getPageCount();
    revNumber = si.getRevNumber();
    wordCount = si.getWordCount();
    encrypted = false;
}

From source file:com.openkm.util.metadata.MetadataExtractor.java

License:Open Source License

/**
 * Extract metadata from Office Word//from   ww  w.  j a  v a 2  s. co m
 */
public static OfficeMetadata officeExtractor(InputStream is, String mimeType) throws IOException {
    POIFSFileSystem fs = new POIFSFileSystem(is);
    OfficeMetadata md = new OfficeMetadata();
    SummaryInformation si = null;

    if (MimeTypeConfig.MIME_MS_WORD.equals(mimeType)) {
        si = new WordExtractor(fs).getSummaryInformation();
    } else if (MimeTypeConfig.MIME_MS_EXCEL.equals(mimeType)) {
        si = new ExcelExtractor(fs).getSummaryInformation();
    } else if (MimeTypeConfig.MIME_MS_POWERPOINT.equals(mimeType)) {
        si = new PowerPointExtractor(fs).getSummaryInformation();
    }

    if (si != null) {
        md.setTitle(si.getTitle());
        md.setSubject(si.getSubject());
        md.setAuthor(si.getAuthor());
        md.setLastAuthor(si.getLastAuthor());
        md.setKeywords(si.getKeywords());
        md.setComments(si.getComments());
        md.setTemplate(si.getTemplate());
        md.setRevNumber(si.getRevNumber());
        md.setApplicationName(si.getApplicationName());
        md.setEditTime(si.getEditTime());
        md.setPageCount(si.getPageCount());
        md.setWordCount(si.getWordCount());
        md.setCharCount(si.getCharCount());
        md.setSecurity(si.getSecurity());

        Calendar createDateTime = Calendar.getInstance();
        createDateTime.setTime(si.getCreateDateTime());
        md.setCreateDateTime(createDateTime);

        Calendar lastSaveDateTime = Calendar.getInstance();
        lastSaveDateTime.setTime(si.getLastSaveDateTime());
        md.setLastSaveDateTime(lastSaveDateTime);

        Calendar lastPrinted = Calendar.getInstance();
        lastPrinted.setTime(si.getLastPrinted());
        md.setLastPrinted(lastPrinted);
    }

    log.info("officeExtractor: {}", md);
    return md;
}

From source file:com.pnf.plugin.ole.parser.StreamReader.java

License:Apache License

private List<INode> readSummaryInfoStream(ByteBuffer stream) {
    List<INode> roots = new LinkedList<>();
    String propType = "Property";

    try {//from   ww  w . ja  v  a  2 s. com
        SummaryInformation sInfo = new SummaryInformation(new PropertySet(stream.array()));

        StreamEntry cInfo = new StreamEntry("Creation Information");
        cInfo.addChild(new StreamEntry("Application Name", propType, sInfo.getApplicationName()));
        cInfo.addChild(new StreamEntry("Creation", "Time",
                sInfo.getCreateDateTime() != null ? sInfo.getCreateDateTime().toString() : null));
        cInfo.addChild(new StreamEntry("Author", propType, sInfo.getAuthor()));
        cInfo.addChild(new StreamEntry("Last Author", propType, sInfo.getLastAuthor()));
        cInfo.addChild(new StreamEntry("Template", propType, sInfo.getTemplate()));
        roots.add(cInfo);

        propType = "Time";
        StreamEntry timeInfo = new StreamEntry("Times");
        timeInfo.addChild(new StreamEntry("Total Edit Time", propType, String.valueOf(sInfo.getEditTime())));
        timeInfo.addChild(new StreamEntry("Last Saved", propType,
                sInfo.getLastSaveDateTime() != null ? sInfo.getLastSaveDateTime().toString() : null));
        timeInfo.addChild(new StreamEntry("Last Printed", propType,
                sInfo.getLastPrinted() != null ? sInfo.getLastPrinted().toString() : null));
        roots.add(timeInfo);

        propType = "Misc";
        StreamEntry misc = new StreamEntry("Miscellaneous");
        misc.addChild(new StreamEntry("OS Version", "int", String.valueOf(sInfo.getOSVersion())));
        misc.addChild(new StreamEntry("Revision Number", "int", sInfo.getRevNumber()));
        misc.addChild(new StreamEntry("Page Count", "int", String.valueOf(sInfo.getPageCount())));
        misc.addChild(new StreamEntry("Word Count", "int", String.valueOf(sInfo.getWordCount())));

        int secVal = sInfo.getSecurity();
        String security = null;

        if (!sInfo.wasNull()) { // Set description according to POI documentation
            switch (secVal) {
            case 0:
                security = "No security";
                break;
            case 1:
                security = "Password protected";
                break;
            case 2:
                security = "Read-only recommended";
                break;
            case 4:
                security = "Read-only enforced";
                break;
            case 8:
                security = "Locked for annotations";
                break;
            default:
                break;
            }

            security += " (code " + secVal + ")";
        } else {
            security = "Field not set";
        }

        misc.addChild(new StreamEntry("Document Security", "int", security));
        misc.addChild(new StreamEntry("Subject", propType, sInfo.getSubject()));
        misc.addChild(new StreamEntry("Keywords", propType, sInfo.getKeywords()));
        roots.add(misc);
    } catch (Throwable t) {
        addMessage("Attempted to read " + SUMM_INFO + " stream but no property sets were found.", null,
                Message.CORRUPT);
    }

    return roots;
}

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

License:Apache License

private void parse(SummaryInformation summary) {
    set(TikaCoreProperties.TITLE, summary.getTitle());
    set(TikaCoreProperties.CREATOR, summary.getAuthor());
    set(TikaCoreProperties.KEYWORDS, summary.getKeywords());
    // TODO Move to OO subject in Tika 2.0
    set(TikaCoreProperties.TRANSITION_SUBJECT_TO_OO_SUBJECT, summary.getSubject());
    set(TikaCoreProperties.MODIFIER, summary.getLastAuthor());
    set(TikaCoreProperties.COMMENTS, summary.getComments());
    set(OfficeOpenXMLExtended.TEMPLATE, summary.getTemplate());
    set(OfficeOpenXMLExtended.APPLICATION, summary.getApplicationName());
    set(OfficeOpenXMLCore.REVISION, summary.getRevNumber());
    set(TikaCoreProperties.CREATED, summary.getCreateDateTime());
    set(TikaCoreProperties.MODIFIED, summary.getLastSaveDateTime());
    set(TikaCoreProperties.PRINT_DATE, summary.getLastPrinted());
    set(Metadata.EDIT_TIME, summary.getEditTime());
    set(OfficeOpenXMLExtended.DOC_SECURITY, summary.getSecurity());

    // New style counts
    set(Office.WORD_COUNT, summary.getWordCount());
    set(Office.CHARACTER_COUNT, summary.getCharCount());
    set(Office.PAGE_COUNT, summary.getPageCount());
    if (summary.getPageCount() > 0) {
        metadata.set(PagedText.N_PAGES, summary.getPageCount());
    }/*w  w  w .  java 2s . c  o  m*/

    // Old style, Tika 1.0 properties
    // TODO Remove these in Tika 2.0
    set(Metadata.TEMPLATE, summary.getTemplate());
    set(Metadata.APPLICATION_NAME, summary.getApplicationName());
    set(Metadata.REVISION_NUMBER, summary.getRevNumber());
    set(Metadata.SECURITY, summary.getSecurity());
    set(MSOffice.WORD_COUNT, summary.getWordCount());
    set(MSOffice.CHARACTER_COUNT, summary.getCharCount());
    set(MSOffice.PAGE_COUNT, summary.getPageCount());
}

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

License:Apache License

private void parse(SummaryInformation summary) {
    set(TikaCoreProperties.TITLE, summary.getTitle());
    addMulti(metadata, TikaCoreProperties.CREATOR, summary.getAuthor());
    set(TikaCoreProperties.KEYWORDS, summary.getKeywords());
    // TODO Move to OO subject in Tika 2.0
    set(TikaCoreProperties.TRANSITION_SUBJECT_TO_OO_SUBJECT, summary.getSubject());
    set(TikaCoreProperties.MODIFIER, summary.getLastAuthor());
    set(TikaCoreProperties.COMMENTS, summary.getComments());
    set(OfficeOpenXMLExtended.TEMPLATE, summary.getTemplate());
    set(OfficeOpenXMLExtended.APPLICATION, summary.getApplicationName());
    set(OfficeOpenXMLCore.REVISION, summary.getRevNumber());
    set(TikaCoreProperties.CREATED, summary.getCreateDateTime());
    set(TikaCoreProperties.MODIFIED, summary.getLastSaveDateTime());
    set(TikaCoreProperties.PRINT_DATE, summary.getLastPrinted());
    set(Metadata.EDIT_TIME, summary.getEditTime());
    set(OfficeOpenXMLExtended.DOC_SECURITY, summary.getSecurity());

    // New style counts
    set(Office.WORD_COUNT, summary.getWordCount());
    set(Office.CHARACTER_COUNT, summary.getCharCount());
    set(Office.PAGE_COUNT, summary.getPageCount());
    if (summary.getPageCount() > 0) {
        metadata.set(PagedText.N_PAGES, summary.getPageCount());
    }/*  w ww . java  2 s .c  o  m*/

    // Old style, Tika 1.0 properties
    // TODO Remove these in Tika 2.0
    set(Metadata.TEMPLATE, summary.getTemplate());
    set(Metadata.APPLICATION_NAME, summary.getApplicationName());
    set(Metadata.REVISION_NUMBER, summary.getRevNumber());
    set(Metadata.SECURITY, summary.getSecurity());
    set(MSOffice.WORD_COUNT, summary.getWordCount());
    set(MSOffice.CHARACTER_COUNT, summary.getCharCount());
    set(MSOffice.PAGE_COUNT, summary.getPageCount());
}

From source file:org.modeshape.sequencer.msoffice.MSOfficeMetadata.java

License:Apache License

public void setSummaryInformation(SummaryInformation si) {
    title = si.getTitle();//ww  w  . j a  va 2  s  . c o  m
    subject = si.getSubject();
    author = si.getAuthor();
    keywords = si.getKeywords();
    comment = si.getComments();
    template = si.getTemplate();
    lastSaved = si.getLastSaveDateTime();
    revision = si.getRevNumber();
    totalEditingTime = si.getEditTime();
    lastPrinted = si.getLastPrinted();
    created = si.getCreateDateTime();
    pages = si.getPageCount();
    words = si.getWordCount();
    characters = si.getCharCount();
    creatingApplication = si.getApplicationName();
    thumbnail = si.getThumbnail();
}

From source file:org.sakaiproject.contentreview.impl.compilatio.CompilatioContentValidator.java

License:Educational Community License

private int wordDocLength(ContentResource resource) {
    if (!serverConfigurationService.getBoolean("tii.checkWordLength", false))
        return 100;

    try {/*ww w.j  a v  a2  s . com*/
        POIFSFileSystem pfs = new POIFSFileSystem(resource.streamContent());
        HWPFDocument doc = new HWPFDocument(pfs);
        SummaryInformation dsi = doc.getSummaryInformation();
        int count = dsi.getWordCount();
        log.debug("got a count of " + count);
        //if this == 0 then its likely that something went wrong -poi couldn't read it
        if (count == 0)
            return 100;
        return count;
    } catch (IOException e) {
        // TODO Auto-generated catch block
        e.printStackTrace();
    } catch (ServerOverloadException e) {
        // TODO Auto-generated catch block
        e.printStackTrace();
    }
    //in case we can't read this lets err on the side of caution
    return 100;
}

From source file:poi.poifs.poibrowser.PropertySetDescriptorRenderer.java

License:Apache License

public Component getTreeCellRendererComponent(final JTree tree, final Object value, final boolean selected,
        final boolean expanded, final boolean leaf, final int row, final boolean hasFocus) {
    final PropertySetDescriptor d = (PropertySetDescriptor) ((DefaultMutableTreeNode) value).getUserObject();
    final PropertySet ps = d.getPropertySet();
    final JPanel p = new JPanel();
    final JTextArea text = new JTextArea();
    text.setBackground(new Color(200, 255, 200));
    text.setFont(new Font("Monospaced", Font.PLAIN, 10));
    text.append(renderAsString(d));//from w  w w  .ja  va2 s  .  co m
    text.append("\nByte order: " + Codec.hexEncode((short) ps.getByteOrder()));
    text.append("\nFormat: " + Codec.hexEncode((short) ps.getFormat()));
    text.append("\nOS version: " + Codec.hexEncode(ps.getOSVersion()));
    text.append("\nClass ID: " + Codec.hexEncode(ps.getClassID()));
    text.append("\nSection count: " + ps.getSectionCount());
    text.append(sectionsToString(ps.getSections()));
    p.add(text);

    if (ps instanceof SummaryInformation) {
        /* Use the convenience methods. */
        final SummaryInformation si = (SummaryInformation) ps;
        text.append("\n");
        text.append("\nTitle:               " + si.getTitle());
        text.append("\nSubject:             " + si.getSubject());
        text.append("\nAuthor:              " + si.getAuthor());
        text.append("\nKeywords:            " + si.getKeywords());
        text.append("\nComments:            " + si.getComments());
        text.append("\nTemplate:            " + si.getTemplate());
        text.append("\nLast Author:         " + si.getLastAuthor());
        text.append("\nRev. Number:         " + si.getRevNumber());
        text.append("\nEdit Time:           " + si.getEditTime());
        text.append("\nLast Printed:        " + si.getLastPrinted());
        text.append("\nCreate Date/Time:    " + si.getCreateDateTime());
        text.append("\nLast Save Date/Time: " + si.getLastSaveDateTime());
        text.append("\nPage Count:          " + si.getPageCount());
        text.append("\nWord Count:          " + si.getWordCount());
        text.append("\nChar Count:          " + si.getCharCount());
        // text.append("\nThumbnail:           " + si.getThumbnail());
        text.append("\nApplication Name:    " + si.getApplicationName());
        text.append("\nSecurity:            " + si.getSecurity());
    }

    if (selected)
        Util.invert(text);
    return p;
}