SEEC-BOK-Backend/src/main/java/com/seec/bok/back/domainobject/Pdf.java

package com.seec.bok.back.domainobject;

import com.seec.bok.back.utils.EachPagePDFTextStripper;
import org.apache.pdfbox.cos.COSDocument;
import org.apache.pdfbox.io.RandomAccessFile;
import org.apache.pdfbox.pdfparser.PDFParser;
import org.apache.pdfbox.pdmodel.PDDocument;
import org.apache.pdfbox.pdmodel.PDPage;
import org.apache.pdfbox.text.PDFTextStripper;
import org.fit.pdfdom.PDFDomTree;

import javax.xml.parsers.ParserConfigurationException;
import java.io.File;
import java.io.IOException;
import java.io.PrintWriter;
import java.io.Writer;


public class Pdf {

    String filename;
    PDDocument  pdf;
    File file;
    public Pdf(String filename) throws IOException {
        this.file =new File("pdf/"+filename);
        this.pdf = PDDocument.load(file);
        this.filename=filename;

    }
    public void toHtml() throws IOException, ParserConfigurationException {
        Writer output = new PrintWriter("pdf/"+filename+".html", "utf-8");
        new PDFDomTree().writeText(pdf, output);
        output.close();
    }
    public void toText() throws IOException {
        String parsedText;
        PDFParser parser = new PDFParser(new RandomAccessFile(file, "r"));
        parser.parse();
        COSDocument cosDoc = parser.getDocument();
        PDFTextStripper pdfStripper = new EachPagePDFTextStripper();
        PDDocument pdDoc = new PDDocument(cosDoc);


        for (PDPage page : pdDoc.getPages()){
            System.out.println(page.getContents().toString());
        }

        parsedText = pdfStripper.getText(pdDoc);
        PrintWriter pw = new PrintWriter("pdf/"+filename+".text");
        pw.print(parsedText);
        pw.close();
        cosDoc.close();

    }
}