List of usage examples for org.apache.poi.hpsf SummaryInformation getCharCount
public int getCharCount()
From source file:com.flexive.extractor.FxSummaryInformation.java
License:Open Source License
/** * Constructor./*ww w.j a v a 2s . co 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 w ww .ja v a2 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: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 ww . ja v a2s. 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 . ja v a 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.modeshape.sequencer.msoffice.MSOfficeMetadata.java
License:Apache License
public void setSummaryInformation(SummaryInformation si) { title = si.getTitle();// w ww .j av a2 s .c om 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: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));//w ww. j a v a2s . c o 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; }