Example usage for org.apache.poi.poifs.eventfilesystem POIFSReader POIFSReader

Introduction

In this page you can find the example usage for org.apache.poi.poifs.eventfilesystem POIFSReader POIFSReader.

Prototype

POIFSReader

Source Link

Usage

From source file:com.armorize.hackalert.extractor.msword.MSExtractor.java

License:Apache License

/**
 * Extracts properties and text from an MS Document input stream
 *//*from w w w .  j a va2  s.c  o  m*/
protected void extract(InputStream input) throws Exception {
    // First, extract properties
    this.reader = new POIFSReader();
    this.properties = new PropertiesBroker();
    this.reader.registerListener(new PropertiesReaderListener(this.properties),
            SummaryInformation.DEFAULT_STREAM_NAME);
    input.reset();
    if (input.available() > 0) {
        reader.read(input);
    }
    // Then, extract text
    input.reset();
    this.text = extractText(input);
}

---

From source file:com.bluecubs.xinco.index.filetypes.XincoIndexMicrosoftPowerpoint.java

License:Apache License

public String getFileContentString(File f) {
    String text = null;//from   w  ww  . j ava  2 s.c  o  m
    try {
        POIFSReader r = new POIFSReader();
        XincoIndexMicrosoftPowerpointPOIFSReaderListener ximpprl = new XincoIndexMicrosoftPowerpointPOIFSReaderListener();
        r.registerListener(ximpprl);
        r.read(new FileInputStream(f));
        text = ximpprl.getEventText();
    } catch (Exception e) {
        text = null;
    }
    return text;
}

---

From source file:com.duroty.lucene.parser.MSPowerPointParser.java

License:Open Source License

/**
 * DOCUMENT ME!// w w w . ja  va 2  s.  c  om
 *
 * @return DOCUMENT ME!
 *
 * @throws ParserException DOCUMENT ME!
 */
private String getContents() throws ParserException {
    String contents = "";

    try {
        POIFSReader reader = new POIFSReader();
        writer = new ByteArrayOutputStream();
        reader.registerListener(this);
        reader.read(in);
        contents = writer.toString();
    } catch (Exception ex) {
        throw new ParserException(ex);
    }

    return contents;
}

---

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

License:Open Source License

/**
 * Extracts the text informations from the excel file.
 *
 * @param in the input stream to read from
 * @return the extraxted informations, or null if no text extraction was possible
 *///from w  w  w  . ja  v  a 2s. c om
public ExtractedData extract(final InputStream in) {

    BufferedInputStream bis = null;
    try {

        writer = new ByteArrayOutputStream();

        // We need to read the stream 2 times, so we use a buffered input stream and mark the
        // beginning
        bis = new BufferedInputStream(in);
        bis.mark(Integer.MAX_VALUE);

        // Retrieve summary information
        POIFSReader r = new POIFSReader();
        r.registerListener(this, "\005SummaryInformation");
        r.read(bis);
        bis.reset();

        // Retrieve text by processing all sheets
        HSSFWorkbook wb = new HSSFWorkbook(bis);
        for (int i = 0; i < wb.getNumberOfSheets(); i++) {
            HSSFSheet sheet = wb.getSheetAt(i);
            processSheet(sheet);
        }

        // Append summary info to text
        if (fxsi != null) {
            writer.write(FxSharedUtils.getBytes(fxsi.getFTIndexInformations()));
        }
        writer.flush();

        return new ExtractedData(fxsi, writer.toString());
    } catch (Exception exc) {
        exc.printStackTrace();
        return null;
    } finally {
        try {
            if (writer != null)
                writer.close();
        } catch (Exception exc) {
            /*ignore*/}
        try {
            if (bis != null)
                bis.close();
        } catch (Exception exc) {
            /*ignore*/}
    }
}

---

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

License:Open Source License

/**
 * Reads the summary information from a document.
 *
 * @param input the input stream to read from, will not be closed at the end
 * @return the summary information//from w  w  w.  j a va  2  s.  c  o  m
 */
public static FxSummaryInformation getSummaryInformation(InputStream input) {
    class SummaryStore implements POIFSReaderListener {
        private FxSummaryInformation fxsi = null;

        /**
         * Processes the Summary section.
         *
         * @param event the summary section event.
         */
        @Override
        public void processPOIFSReaderEvent(POIFSReaderEvent event) {
            try {
                SummaryInformation si = (SummaryInformation) PropertySetFactory.create(event.getStream());
                fxsi = new FxSummaryInformation(si);
            } catch (Exception ex) {
                /* ignore */
            }
        }

        protected FxSummaryInformation getFxSummaryInformation() {
            return fxsi;
        }
    }
    try {
        POIFSReader reader = new POIFSReader();
        SummaryStore st = new SummaryStore();
        reader.registerListener(st, "\005SummaryInformation");
        reader.read(input);
        return st.getFxSummaryInformation();
    } catch (Exception ex) {
        return null;
    }
}

---

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

License:Open Source License

/**
 * Extracts the text informations from the powerpoint file.
 *
 * @param in the input stream to read from
 * @return the extraxted informations, or null if no text extraction was possible
 *///from w w w .j  a  v a2 s.c  o  m
public ExtractedData extract(final InputStream in) {
    try {
        writer = new ByteArrayOutputStream();
        POIFSReader reader = new POIFSReader();
        reader.registerListener(this);
        //FxSummaryInformation.getSummaryInformation(fileName);
        reader.read(in);
        if (fxsi != null) {
            writer.write(FxSharedUtils.getBytes(fxsi.getFTIndexInformations()));
        }
        writer.flush();
        return new ExtractedData(fxsi, writer.toString());
    } catch (Exception ex) {
        return null;
    } finally {
        try {
            writer.close();
        } catch (Exception exc) {
            /*ignore*/}
    }
}

---

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

License:Open Source License

/**
 * Extracts the text informations from the word file.
 *
 * @param in the input stream to read from
 * @return the extraxted informations, or null if no text extraction was possible
 *///from  ww  w.ja  v a2s .co  m
public ExtractedData extract(final InputStream in) {
    ByteArrayOutputStream baos = null;
    PrintWriter writer = null;
    BufferedInputStream bis = null;
    try {

        baos = new ByteArrayOutputStream();
        writer = new PrintWriter(baos);

        // We need to read the stream 2 times, so we use a buffered input stream and mark the
        // beginning
        bis = new BufferedInputStream(in);
        bis.mark(Integer.MAX_VALUE);

        // Retrieve summary information
        POIFSReader r = new POIFSReader();
        r.registerListener(this, "\005SummaryInformation");
        r.read(bis);
        bis.reset();

        // Retrieve text
        WordDocument wd = new WordDocument(bis);
        wd.writeAllText(writer);
        if (fxsi != null) {
            writer.write(fxsi.getFTIndexInformations());
        }
        writer.flush();

        return new ExtractedData(fxsi, baos.toString());
    } catch (Exception exc) {
        return null;
    } finally {
        try {
            if (writer != null)
                writer.close();
        } catch (Exception exc) {
            /*ignore*/}
        try {
            if (baos != null)
                baos.close();
        } catch (Exception exc) {
            /*ignore*/}
        try {
            if (bis != null)
                bis.close();
        } catch (Exception exc) {
            /*ignore*/}
    }
}

---

From source file:com.knowgate.ole.OLEDocument.java

License:Open Source License

public OLEDocument() {

    oReader = new POIFSReader();
    oListener = new OLEListener();

    oReader.registerListener(oListener, "\005SummaryInformation");

}

---

From source file:com.krawler.esp.fileparser.word.ExtractWordFile.java

License:Open Source License

public Properties extractProperties(InputStream in) throws IOException {

    PropertiesBroker propertiesBroker = new PropertiesBroker();
    POIFSReader reader = new POIFSReader();
    reader.registerListener(new PropertiesReaderListener(propertiesBroker), "\005SummaryInformation");
    reader.read(in);//from  w  w  w .  j  ava 2  s. c o m
    return propertiesBroker.getProperties();
}

---

From source file:com.villemos.ispace.aperture.enricher.MicrosoftPropertyReader.java

License:Open Source License

@Handler
public void addMSProperties(@Body InformationObject io, @Headers Map<String, Object> headers) {

    File file = new File(io.hasUri);

    if (file.exists() && io.hasUri.endsWith(".doc")) {
        POIFSReader r = new POIFSReader();
        r.registerListener(this);
        try {//from w w w. ja va  2 s  . c o  m
            FileInputStream inStream = new FileInputStream(file);
            r.read(inStream);

            Iterator<Entry<String, String>> it = msProperties.entrySet().iterator();
            while (it.hasNext()) {
                Entry<String, String> entry = it.next();
                io.metadata.put(entry.getKey(), entry.getValue());
            }
            inStream.close();
        } catch (Exception e) {
            e.printStackTrace();
            LOG.error("Failed to get properties for .doc file '" + file.getName() + "'.");
        }
    }
}

---

• «
• 1
• 2
• 3
• 4
• 5
• 6
• »