public static void writeText(String pdfPath, String outputPath) throws java.io.IOException {
Document pdf = PDF.open(pdfPath);
StringBuilder text = new StringBuilder(1024);
// pdf.pipe(new OutputTarget(text));
pdf.pipe(new VisualOutputTarget(text));
pdf.close();
File file = new File(outputPath);
FileOutputStream outputStream = new FileOutputStream(file);
outputStream.write(text.toString().getBytes());
outputStream.flush();
outputStream.close();
}
public void index(Path file, IndexWriter indexWriter)
throws IOException, InvalidFormatException, OpenXML4JException, XmlException {
logger.info("Indexing PDF document " + file);
try (com.snowtide.pdf.Document pdf = PDF.open(file.toFile())) {
LucenePDFConfiguration configuration = new LucenePDFConfiguration();
configuration.setBodyTextFieldName("content");
Document pdfDocument = LucenePDFDocumentFactory.buildPDFDocument(pdf, configuration);
pdfDocument.add(new StoredField("name", file.toFile().getName()));
pdfDocument.add(new StoredField("path", file.toFile().getAbsolutePath()));
indexWriter.addDocument(pdfDocument);
}
}
