Example usage for org.apache.pdfbox.pdmodel PDDocument load

List of usage examples for org.apache.pdfbox.pdmodel PDDocument load

  1. HOME
  2. Java
  3. org.apache.pdfbox
  4. org.apache.pdfbox.pdmodel.*
  5. PDDocument
  6. load

Introduction

In this page you can find the example usage for org.apache.pdfbox.pdmodel PDDocument load.

Prototype

public static PDDocument load(byte[] input) throws IOException

Source Link

Document

Parses a PDF.

Usage

From source file:indexer.PDFTextExtractor.java

License:Open Source License

private PDDocument getPDDocument(String filename) throws IOException {
    if (stripper == null)
        throw new IOException("ERROR: PDFStripper was not created");
    File file = new File(filename);
    if (!file.exists())
        throw new IOException("ERROR: " + filename + " doesn't exist");
    return PDDocument.load(file);
}

From source file:io.cloudslang.content.utilities.services.PdfParseService.java

License:Open Source License

private static PDDocument getPdfDocument(final Path path, final String password) throws IOException {
    if (isEmpty(password))
        return PDDocument.load(path.toFile());
    return PDDocument.load(path.toFile(), password);
}

From source file:io.inkstand.scribble.pdf.BasePDFMatcher.java

License:Apache License

/**
 * Is invoked by the matches method when the type of the target object is verified. Override this method to add
 * verifications on the raw data instead of a loaded document.
 *
 * @param pdf//  w w  w.  j a  va 2s  .c o m
 *         the handle for the PDF data
 *
 * @return <code>true</code> if the pdf document is a valid PDF document.
 */
protected boolean matches(PDF pdf) {
    try (InputStream inStream = pdf.openStream()) {
        final PDDocument doc = PDDocument.load(inStream);
        return matchesPDF(doc);
    } catch (IOException e) {
        LOG.debug("Could not load PDF document", e);
        return false;
    }
}

From source file:io.konik.carriage.pdfbox.PDFBoxInvoiceAppender.java

License:Open Source License

@Override
public void append(AppendParameter appendParameter) {
    InputStream inputPdf = appendParameter.inputPdf();
    try {/*www . j a  va  2s .  com*/
        PDDocument doc = PDDocument.load(inputPdf);
        setMetadata(doc, appendParameter);
        attachZugferdFile(doc, appendParameter.attachmentFile());
        doc.getDocument().setVersion(1.7f);
        doc.save(appendParameter.resultingPdf());
        doc.close();
    } catch (Exception e) {
        throw new InvoiceAppendError("Error appending Invoice", e);
    }

}

From source file:io.konik.carriage.pdfbox.PDFBoxInvoiceExtractor.java

License:Open Source License

private static final InputStream extractIntern(InputStream pdfStream) throws IOException {
    PDDocument doc = PDDocument.load(pdfStream);
    InputStream inputStream = extractZugferdFileAttachment(doc);
    return new CallBackInputStream(inputStream, doc);
}

From source file:it.myideas.bancamarcheextractor.Distinta.java

public static Distinta parse(Path file) {

    try (PDDocument doc = PDDocument.load(file.toFile())) {

        Distinta distinta = new Distinta();

        PDFTextStripper stripper = new PDFTextStripper();
        String contents = stripper.getText(doc);
        Stream<String> lines = Arrays.stream(contents.split(stripper.getLineSeparator()));

        log.debug("FILE:" + file.toString());
        log.debug(contents);/*  w w w .j a  v a2  s. com*/

        lines.forEach(line -> {

            if (line.startsWith("Tipo disposizione")) {
                distinta.tipoDisposizione = line.replace("Tipo disposizione", "").trim().toLowerCase();
            } else if (line.startsWith("1 Esecuzione")) {
                String[] p = line.split(" ");

                distinta.beneficiario = Arrays.stream(Arrays.copyOfRange(p, 4, p.length))
                        .map(String::toLowerCase).collect(Collectors.joining("_"));

                distinta.data = LocalDate.parse(p[2], DateTimeFormatter.ofPattern("dd/MM/yyyy"));
            }

        });

        if (!isOk(distinta.beneficiario) || !isOk(distinta.tipoDisposizione) || distinta.data == null) {
            throw new IOException("Parser failure for file " + file.toString());
        }

        return distinta;
    } catch (IOException e) {
        log.error("Error parsing PDF", e);
        return null;
    }
}

From source file:javaapplication2.NewJFrame.java

private void jButton1MouseClicked(java.awt.event.MouseEvent evt) {//GEN-FIRST:event_jButton1MouseClicked
    /* JFileChooser chooser = new JFileChooser();
     FileNameExtensionFilter filter = new FileNameExtensionFilter(
    "ficheiros PDF", "pdf");/*ww  w . jav a  2s.  co  m*/
     chooser.setFileFilter(filter);
     int returnVal;
     returnVal = chooser.showOpenDialog(getParent());
     if(returnVal == JFileChooser.APPROVE_OPTION) {*/
    if (!jTextField1.getText().isEmpty()) {

        System.out.println("You chose to open this file: " + jTextField1.getText());
        String text = null;
        String[] linhas = null;

        try {
            PDDocument doc = PDDocument.load(jTextField1.getText());
            PDFTextStripper stripper = new PDFTextStripper();
            text = stripper.getText(doc);
            doc.close();
        } catch (IOException ex) {
            Logger.getLogger(NewJFrame.class.getName()).log(Level.SEVERE, null, ex);
        }
        linhas = text.split("\n", -1);

        ArrayList<Dia> dias = new ArrayList<>();
        Refeicao r = null;
        Dia d = null;
        Boolean almoco = false;
        Boolean ignorarDia = false; //dias sem refeio ex. feriados
        for (String linha : linhas) {
            if (linha.contains("Segunda-feira") || linha.contains("Tera-feira")
                    || linha.contains("Quarta-feira") || linha.contains("Quinta-feira")
                    || linha.contains("Sexta-feira")) {
                if (d != null) {// && !ignorarDia){
                    dias.add(d);
                    // ignorarDia=false;
                }
                d = new Dia();
                almoco = true;
                r = new Refeicao();
                jTextArea1.setText(jTextArea1.getText() + linha.trim() + " --NOVO DIA \n");
            }
            /* else if(linha.contains("ENCERRADO")){
            ignorarDia=true;
             }*/
            else if (isValidDate(linha)) {
                DateFormat format = new SimpleDateFormat("dd-MM-yyyy", Locale.ENGLISH);
                try {
                    d.setDia(format.parse(linha));
                } catch (ParseException ex) {
                    Logger.getLogger(NewJFrame.class.getName()).log(Level.SEVERE, null, ex);
                }
                jTextArea1.setText(jTextArea1.getText() + linha.trim() + " --DATA \n");
            } else if (linha.contains("Sopa")) {

                if (almoco)
                    jTextArea1.setText(jTextArea1.getText() + " ALMOO \n");
                else
                    jTextArea1.setText(jTextArea1.getText() + " JANTAR \n");

                r.setSopa(linha);
                jTextArea1.setText(jTextArea1.getText() + linha.trim() + " --SOPA \n");
            }

            else if (linha.contains("Carne")) {
                r.setCarne(linha);
                jTextArea1.setText(jTextArea1.getText() + linha.trim() + " --CARNE \n");
            }

            else if (linha.contains("Peixe")) {
                r.setPeixe(linha);
                jTextArea1.setText(jTextArea1.getText() + linha.trim() + " --PEIXE \n");
            }

            else if (linha.contains("Sobremesa")) {

                jTextArea1.setText(jTextArea1.getText() + linha.trim() + " --SOBREMESA \n");
                r.setSobremesa(linha);
                if (almoco) {
                    d.setAlmoco(r);
                    r = new Refeicao();
                    almoco = false;
                } else {
                    d.setJantar(r);
                    r = new Refeicao();
                    almoco = true;
                }

            } else {
                jTextArea1.setText(jTextArea1.getText() + linha.trim() + " --Ignorado \n");
            }

            // jTextArea1.setText(jTextArea1.getText()+t+" a testar\n ");
        }

        if (d != null)
            dias.add(d);

        jTextArea1.setText(" ");
        System.out.println("nmero de dias: " + dias.size());
        getSelectedButtonText();
        getConnection();
        for (Dia dia : dias) {
            //if(dia.dia.compareTo(new Date(2015, 06, 12))==0){
            if (dia.dia != null) {
                System.out.println("dia: " + dia.dia);
                jTextArea1.setText(jTextArea1.getText() + "\n" + dia.dia);
                jTextArea1.setText(jTextArea1.getText() + "\n" + "-----ALMOO-----\n");
                if (dia.almoco != null)
                    jTextArea1
                            .setText(jTextArea1.getText() + dia.almoco.getSopa() + "\n" + dia.almoco.getCarne()
                                    + "\n" + dia.almoco.getPeixe() + "\n" + dia.almoco.getSobremesa() + "\n");
                else
                    jTextArea1.setText(jTextArea1.getText() + "Nao definido" + "\n");

                jTextArea1.setText(jTextArea1.getText() + "\n" + "-----JANTAR-----\n");
                if (dia.jantar != null)
                    jTextArea1
                            .setText(jTextArea1.getText() + dia.jantar.getSopa() + "\n" + dia.jantar.getCarne()
                                    + "\n" + dia.jantar.getPeixe() + "\n" + dia.jantar.getSobremesa() + "\n");
                else
                    jTextArea1.setText(jTextArea1.getText() + "Nao definido" + "\n");

                try {
                    String query = " insert into " + getSelectedButtonText()
                            + " (data,temAlmoco,almoco_sopa,almoco_carne,almoco_peixe,almoco_sobremesa,temJantar,jantar_sopa,jantar_carne,jantar_peixe,jantar_sobremesa)"
                            + " values (?,?,?,?,?,?,?,?,?,?,?)";

                    // create the mysql insert preparedstatement
                    PreparedStatement preparedStmt = conexao.prepareStatement(query);
                    java.sql.Date sqldate = new Date(dia.dia.getTime());
                    preparedStmt.setDate(1, sqldate);
                    if (dia.almoco != null) {
                        preparedStmt.setInt(2, 1);
                        preparedStmt.setString(3, dia.almoco.getSopa());
                        preparedStmt.setString(4, dia.almoco.getCarne());
                        preparedStmt.setString(5, dia.almoco.getPeixe());
                        preparedStmt.setString(6, dia.almoco.getSobremesa());
                    } else {
                        preparedStmt.setInt(2, 0);
                        preparedStmt.setString(3, "No Definido");
                        preparedStmt.setString(4, "No Definido");
                        preparedStmt.setString(5, "No Definido");
                        preparedStmt.setString(6, "No Definido");
                    }
                    if (dia.jantar != null) {
                        preparedStmt.setInt(7, 1);
                        preparedStmt.setString(8, dia.jantar.getSopa());
                        preparedStmt.setString(9, dia.jantar.getCarne());
                        preparedStmt.setString(10, dia.jantar.getPeixe());
                        preparedStmt.setString(11, dia.jantar.getSobremesa());
                    } else {
                        preparedStmt.setInt(7, 0);
                        preparedStmt.setString(8, "No Definido");
                        preparedStmt.setString(9, "No Definido");
                        preparedStmt.setString(10, "No Definido");
                        preparedStmt.setString(11, "No Definido");
                    }
                    // execute the preparedstatement
                    preparedStmt.execute();
                    //System.out.println(text);
                    //jTextArea1.setText(text);

                } catch (SQLException ex) {
                    Logger.getLogger(NewJFrame.class.getName()).log(Level.SEVERE, null, ex);
                }
            }
        }

    }
}

From source file:jeanderson.br.converte.Converte.java

public static void converter(String nome) {
    File arquivo = new File(
            "/Users/" + System.getProperty("user.name") + "/Google Drive/Notas Fiscais/PDF/" + nome + ".pdf");
    if (arquivo.exists()) {
        try {//www  .j a va2 s .c  o m
            PDDocument documento = PDDocument.load(arquivo);
            List<PDPage> lista = documento.getDocumentCatalog().getAllPages();
            int numeroDePagina = 1;
            for (PDPage paginas : lista) {
                BufferedImage imagem = paginas.convertToImage();
                File saida = new File("/Users/" + System.getProperty("user.name")
                        + "/Google Drive/Notas Fiscais/PNG/" + nome + numeroDePagina + ".png");
                ImageIO.write(imagem, "png", saida);
                numeroDePagina++;
            }
            documento.close();
        } catch (IOException ex) {
            Relatar.bug(Converte.class.getName(), ex.toString());
            Logger.getLogger(Converte.class.getName()).log(Level.SEVERE, null, ex);
        }
    }
}

From source file:khoji.PDFdoc.java

License:Apache License

/**
 *
 * @param doc_path/* www . java  2  s.  c o  m*/
 * @return
 * @throws Exception
 */
public String extractPDF(String doc_path) throws Exception {

    PDDocument document = null;
    try {
        document = PDDocument.load(doc_path);
        if (document.isEncrypted()) {
            try {
                document.decrypt("");
            } catch (InvalidPasswordException e) {
                System.err.println("Error: Document is encrypted with a password.");
                System.exit(1);
            }
        }

        PDFTextStripper stripper = new PDFTextStripper();
        String text = "";
        text = stripper.getText(document);
        //            System.out.println("text:"+text);
        return text;
    } finally {
        if (document != null) {
            document.close();
        }
    }

}

From source file:kz.technovision.technokitap.Book.java

/**
 * Imagify/*from w  w w  . j a  v a2 s  .c o m*/
 *
 * @param pdfFile
 */
public void imagify(String pdfFile) {
    try (PDDocument document = PDDocument.load(pdfFile)) {
        int imageType = BufferedImage.TYPE_INT_RGB;
        PDFImageWriter pw = new PDFImageWriter();
        boolean success = pw.writeImage(document, "jpg", "", 1, document.getNumberOfPages(), big, imageType,
                300);
        if (!success) {
            System.err.println("Error: no writer found for image format");
            System.exit(1);
        }
    } catch (Exception ex) {
        System.err.println(ex);
    }
}