Example usage for org.apache.pdfbox.io RandomAccessBufferedFileInputStream RandomAccessBufferedFileInputStream

Introduction

In this page you can find the example usage for org.apache.pdfbox.io RandomAccessBufferedFileInputStream RandomAccessBufferedFileInputStream.

Prototype

public RandomAccessBufferedFileInputStream(InputStream input) throws IOException

Source Link

Document

Create a random access input stream for the given input stream by copying the data to a temporary file.

Usage

From source file:com.sastix.cms.common.services.htmltopdf.PdfTest.java

License:Apache License

@Test
public void testPdfFromStringTo() throws Exception {

    // GIVEN an html template containing special characters that java stores in utf-16 internally
    Pdf pdf = pdfBuilder.build();//w ww.  jav  a  2s .com
    pdf.addPage("<html><head><meta charset=\"utf-8\"></head><h1>Mller</h1></html>", PageType.htmlAsString);

    String tempFolder = temporaryFolder.newFolder().getPath();
    pdf.saveAs(tempFolder + "/output.pdf");

    // WHEN
    byte[] pdfBytes = pdf.getPDF();

    PDFParser parser = new PDFParser(
            new RandomAccessBufferedFileInputStream(new ByteArrayInputStream(pdfBytes)));

    // that is a valid PDF (otherwise an IOException occurs)
    parser.parse();
    PDFTextStripper pdfTextStripper = new PDFTextStripper();
    String pdfText = pdfTextStripper.getText(new PDDocument(parser.getDocument()));

    assertThat("document should contain the creditorName", pdfText, containsString("Mller"));
}

From source file:com.validation.manager.core.server.core.AttachmentServerTest.java

License:Apache License

/**
 * Test of addFile method, of class AttachmentServer.
 *//*from   w ww  .  j a  v  a  2  s . com*/
@Test
public void testAddRetrieveTextFile() {
    try {
        System.out.println("add text File");
        File f = new File("target/Test.txt");
        f.deleteOnExit();
        List<String> lines = Arrays.asList("The first line", "The second line");
        Path file = Paths.get(f.getAbsolutePath());
        Files.write(file, lines, Charset.forName("UTF-8"));
        AttachmentServer instance = new AttachmentServer();
        instance.addFile(f, f.getName());
        instance.write2DB();
        //Delete the file
        FileUtils.delete(f.getAbsolutePath());
        assertEquals(1, (int) instance.getAttachmentType().getId());//Text file
        System.out.println("retrieveFile");
        AttachmentServer temp = new AttachmentServer(instance.getAttachmentPK());
        File loadedFile = temp.getAttachedFile("target/loaded/");
        BufferedReader br = new BufferedReader(new FileReader(loadedFile));
        String line;
        int count = 0;
        while ((line = br.readLine()) != null) {
            assertEquals(lines.get(count), line);
            System.out.println(line);
            count++;
        }
        assertEquals(lines.size(), count);
        //Create pdf file
        System.out.println("add pdf File");
        File pdf = Tool.convertToPDF(loadedFile, "target/Text.pdf");
        pdf.deleteOnExit();
        instance = new AttachmentServer();
        instance.addFile(pdf, pdf.getName());
        instance.write2DB();
        //Delete the file
        FileUtils.delete(pdf.getAbsolutePath());
        assertEquals(2, (int) instance.getAttachmentType().getId());//PDF file
        System.out.println("retrieveFile");
        temp = new AttachmentServer(instance.getAttachmentPK());
        loadedFile = temp.getAttachedFile("target/loaded/");
        PDFTextStripper pdfStripper;
        PDDocument pdDoc = null;
        COSDocument cosDoc = null;
        try {
            PDFParser parser = new PDFParser(new RandomAccessBufferedFileInputStream(loadedFile));
            parser.parse();
            cosDoc = parser.getDocument();
            pdfStripper = new PDFTextStripper();
            pdDoc = new PDDocument(cosDoc);
            pdfStripper.setStartPage(1);
            pdfStripper.setEndPage(1);
            String parsedText = pdfStripper.getText(pdDoc);
            System.out.println(parsedText);
        } catch (IOException ex) {
            Exceptions.printStackTrace(ex);
            fail();
        } finally {
            if (cosDoc != null) {
                cosDoc.close();
            }
            if (pdDoc != null) {
                pdDoc.close();
            }
        }
    } catch (IOException | VMException ex) {
        Exceptions.printStackTrace(ex);
        fail();
    }
}

From source file:cz.mzk.editor.server.handler.GetOcrFromPdfHandler.java

License:Open Source License

private String pdftoText(String fileName) throws ActionException {

    File pdfFile = new File(fileName);

    if (!pdfFile.isFile()) {
        LOGGER.error("The file: " + fileName + " does not exist.");
        throw new ActionException("Unable to parse the pdf file.");
    }/*from  w  ww . j a  v  a  2 s .  c  o m*/

    PDFParser parser = null;
    COSDocument cosDoc = null;
    PDFTextStripper pdfStripper;
    PDDocument pdDoc = null;
    String parsedText;
    try {
        parser = new PDFParser(new RandomAccessBufferedFileInputStream(new FileInputStream(pdfFile)));
    } catch (Exception e) {
        LOGGER.error("Unable to open PDF Parser.: " + e);
        e.printStackTrace();
        throw new ActionException("Unable to parse the pdf file.");
    }

    try {
        parser.parse();
        cosDoc = parser.getDocument();
        pdfStripper = new PDFTextStripper();
        pdDoc = new PDDocument(cosDoc);
        parsedText = pdfStripper.getText(pdDoc);
    } catch (Exception e) {
        LOGGER.error("An exception occured in parsing the PDF Document.");
        e.printStackTrace();
        throw new ActionException("Unable to parse the pdf file. " + e);
    } finally {
        try {
            if (cosDoc != null)
                cosDoc.close();
            if (pdDoc != null)
                pdDoc.close();
        } catch (Exception e) {
            e.printStackTrace();
        }
    }

    return parsedText;
}

From source file:org.opensingular.lib.commons.pdf.TestPDFUtil.java

License:Apache License

private int countPages(File file) {
    try (RandomAccessBufferedFileInputStream in = new RandomAccessBufferedFileInputStream(file)) {
        PDFParser parser = new PDFParser(in);
        parser.parse();/*from w  w  w  . java2  s . com*/
        PDDocument doc = parser.getPDDocument();
        int pages = doc.getNumberOfPages();
        doc.close();
        return pages;
    } catch (IOException e) {
        throw new RuntimeException(e);
    }
}