Examples with PDDocument org.apache.pdfbox.pdmodel.PDDocument used on opensource projects

Example 1 with PDDocument

use of org.apache.pdfbox.pdmodel.PDDocument in project paper2ebook by ogrisel.

the class Transformer method main.

public static void main(String[] args) throws IOException, COSVisitorException {
    String original_pdf;
    if (args.length < 1 || args.length > 2) {
        System.err.println("Usage: java -jar paper2ebook-*.jar input.pdf [output.pdf]");
        return;
    } else {
        original_pdf = args[0];
    }
    Transformer transformer = new Transformer(PDDocument.load(original_pdf));
    PDDocument output = transformer.extract();
    if (args.length == 1) {
        String orig_no_pdf = original_pdf.substring(0, original_pdf.length() - 4);
        output.save(orig_no_pdf + "_ebook.pdf");
    } else {
        output.save(args[1]);
    }
}

Example 2 with PDDocument

use of org.apache.pdfbox.pdmodel.PDDocument in project mustangproject by ZUGFeRD.

the class ZUGFeRDExporterFromA1Factory method loadFromPDFA1.

/**
     * Makes A PDF/A3a-compliant document from a PDF-A1 compliant document (on
     * the metadata level, this will not e.g. convert graphics to JPG-2000)
     *
     * @param pdfBinary binary of a PDF/A1 compliant document
     */
public ZUGFeRDExporter loadFromPDFA1(byte[] pdfBinary) throws IOException, TransformerException {
    ensurePDFIsValidA1(new ByteArrayDataSource(new ByteArrayInputStream(pdfBinary)));
    PDDocument doc = PDDocument.load(pdfBinary);
    makePDFA3compliant(doc);
    return new ZUGFeRDExporter(doc);
}

Example 3 with PDDocument

use of org.apache.pdfbox.pdmodel.PDDocument in project camel by apache.

the class PdfProducer method doExtractText.

private String doExtractText(Exchange exchange) throws IOException, CryptographyException, InvalidPasswordException, BadSecurityHandlerException {
    LOG.debug("Got {} operation, going to extract text from provided pdf.", pdfConfiguration.getOperation());
    PDDocument document = exchange.getIn().getBody(PDDocument.class);
    if (document.isEncrypted()) {
        DecryptionMaterial decryptionMaterial = exchange.getIn().getHeader(DECRYPTION_MATERIAL_HEADER_NAME, DecryptionMaterial.class);
        if (decryptionMaterial == null) {
            throw new IllegalArgumentException(String.format("%s header is expected for %s operation " + "on encrypted document", DECRYPTION_MATERIAL_HEADER_NAME, pdfConfiguration.getOperation()));
        }
        document.openProtection(decryptionMaterial);
    }
    PDFTextStripper pdfTextStripper = new PDFTextStripper();
    return pdfTextStripper.getText(document);
}

Example 4 with PDDocument

use of org.apache.pdfbox.pdmodel.PDDocument in project camel by apache.

the class PdfAppendTest method testAppend.

@Test
public void testAppend() throws Exception {
    final String originalText = "Test";
    final String textToAppend = "Append";
    PDDocument document = new PDDocument();
    PDPage page = new PDPage(PDPage.PAGE_SIZE_A4);
    document.addPage(page);
    PDPageContentStream contentStream = new PDPageContentStream(document, page);
    contentStream.setFont(PDType1Font.HELVETICA, 12);
    contentStream.beginText();
    contentStream.moveTextPositionByAmount(20, 400);
    contentStream.drawString(originalText);
    contentStream.endText();
    contentStream.close();
    template.sendBodyAndHeader("direct:start", textToAppend, PdfHeaderConstants.PDF_DOCUMENT_HEADER_NAME, document);
    resultEndpoint.setExpectedMessageCount(1);
    resultEndpoint.expectedMessagesMatches(new Predicate() {

        @Override
        public boolean matches(Exchange exchange) {
            Object body = exchange.getIn().getBody();
            assertThat(body, instanceOf(ByteArrayOutputStream.class));
            try {
                PDDocument doc = PDDocument.load(new ByteArrayInputStream(((ByteArrayOutputStream) body).toByteArray()));
                PDFTextStripper pdfTextStripper = new PDFTextStripper();
                String text = pdfTextStripper.getText(doc);
                assertEquals(2, doc.getNumberOfPages());
                assertThat(text, containsString(originalText));
                assertThat(text, containsString(textToAppend));
            } catch (IOException e) {
                throw new RuntimeException(e);
            }
            return true;
        }
    });
    resultEndpoint.assertIsSatisfied();
}

Example 5 with PDDocument

use of org.apache.pdfbox.pdmodel.PDDocument in project camel by apache.

the class PdfAppendTest method testAppendEncrypted.

@Test
public void testAppendEncrypted() throws Exception {
    final String originalText = "Test";
    final String textToAppend = "Append";
    PDDocument document = new PDDocument();
    PDPage page = new PDPage(PDPage.PAGE_SIZE_A4);
    document.addPage(page);
    PDPageContentStream contentStream = new PDPageContentStream(document, page);
    contentStream.setFont(PDType1Font.HELVETICA, 12);
    contentStream.beginText();
    contentStream.moveTextPositionByAmount(20, 400);
    contentStream.drawString(originalText);
    contentStream.endText();
    contentStream.close();
    final String ownerPass = "ownerPass";
    final String userPass = "userPass";
    AccessPermission accessPermission = new AccessPermission();
    accessPermission.setCanExtractContent(false);
    StandardProtectionPolicy protectionPolicy = new StandardProtectionPolicy(ownerPass, userPass, accessPermission);
    protectionPolicy.setEncryptionKeyLength(128);
    document.protect(protectionPolicy);
    ByteArrayOutputStream output = new ByteArrayOutputStream();
    document.save(output);
    // Encryption happens after saving.
    PDDocument encryptedDocument = PDDocument.load(new ByteArrayInputStream(output.toByteArray()));
    Map<String, Object> headers = new HashMap<String, Object>();
    headers.put(PdfHeaderConstants.PDF_DOCUMENT_HEADER_NAME, encryptedDocument);
    headers.put(PdfHeaderConstants.DECRYPTION_MATERIAL_HEADER_NAME, new StandardDecryptionMaterial(userPass));
    template.sendBodyAndHeaders("direct:start", textToAppend, headers);
    resultEndpoint.setExpectedMessageCount(1);
    resultEndpoint.expectedMessagesMatches(new Predicate() {

        @Override
        public boolean matches(Exchange exchange) {
            Object body = exchange.getIn().getBody();
            assertThat(body, instanceOf(ByteArrayOutputStream.class));
            try {
                PDDocument doc = PDDocument.load(new ByteArrayInputStream(((ByteArrayOutputStream) body).toByteArray()));
                PDFTextStripper pdfTextStripper = new PDFTextStripper();
                String text = pdfTextStripper.getText(doc);
                assertEquals(2, doc.getNumberOfPages());
                assertThat(text, containsString(originalText));
                assertThat(text, containsString(textToAppend));
            } catch (IOException e) {
                throw new RuntimeException(e);
            }
            return true;
        }
    });
    resultEndpoint.assertIsSatisfied();
}