对加密的pdf文件进行解密，并获取pdf内的附件内容（为跳转网址以及附件内容另存指定位置）

原创已于 2024-11-26 16:30:17 修改 · 324 阅读

CC 4.0 BY-SA版权

文章标签：

于 2024-11-26 15:35:35 首次发布

邮件的附件是一个加密的pdf，需要对其进行密码解密，
解密后将未加密的文件另存新的存放地址，并对处理后的pdf内容进行解析，
pdf的附件是一个url下载地址，获取此下载地址，并跳转网址，
成功下载对应的文件，存放至本地；对pdf的内的附近也进行另存至指定地址

import org.apache.pdfbox.cos.COSDictionary;
import org.apache.pdfbox.pdmodel.PDDocument;
import com.spire.pdf.PdfDocument;
import com.spire.pdf.attachments.PdfAttachmentCollection;
import org.apache.pdfbox.pdmodel.interactive.annotation.PDAnnotation;

import java.io.*;
import java.util.List;
import java.util.regex.Matcher;
import java.util.regex.Pattern;

    public static void main(String[] args) throws Exception {
        String inputFile ="F:\\123.pdf";
        String outputFile = "F:\\321.pdf";
        String password = "password";
        boolean isEncrypted = false;

        try  {
            PDDocument document = PDDocument.load(new File(inputFile),password);
            isEncrypted = document.isEncrypted();

            if (isEncrypted) {
                System.out.println("There is a password");
                    document.setAllSecurityToBeRemoved(true);
                    document.save(outputFile);
                    System.out.println("PDF  File decrypted and saved successfully ！");
                    
				PdfDocument pdf = new PdfDocument();
                pdf.loadFromFile(outputFile);
                PdfAttachmentCollection attachmentCollection = pdf.getAttachments();
                for(int i = 0; i<attachmentCollection.getCount(); i++){

                    String fileName = attachmentCollection.get(i).getFileName();
                    fileName = "F:\\\\"+fileName;
                    File file = new File(fileName);
                    OutputStream outputStream = new FileOutputStream(file);
                    BufferedOutputStream bufferedOutputStream = new BufferedOutputStream(outputStream);
                    bufferedOutputStream.write(attachmentCollection.get(i).getData());
                    bufferedOutputStream.close();
                    System.out.println("PDF File decryption and save attachment successfully！");

                }
                    

                    //todo  Read the information from the file, obtain the URL to download the file
                    
                PDDocument doc = PDDocument.load(new File(outputFile));
                List<PDAnnotation> annotations = doc.getPage(0).getAnnotations();

                for (int i = 0; i < annotations.size(); i++) {
                    PDAnnotation pdAnnotation = annotations.get(i);
                    COSDictionary cosObject = pdAnnotation.getCOSObject();
                    String cosBase = cosObject.getItem("A").toString();
                    String url = null;
                    String pattern = String.format("%s(.+?)%s", Pattern.quote("COSString{"), Pattern.quote("};}"));
                    Matcher matcher = Pattern.compile(pattern).matcher(cosBase);
                    if (matcher.find()) {
                        url = matcher.group(1);
                    }
                Pattern patterns = Pattern.compile("https?:\\/\\/[^\\s]+");
                Matcher matchers = patterns.matcher(url);

                while (matchers.find()) {
                    System.out.println(matchers.group());
                }              
                }                 
            }
             else {
                System.out.println("No password");
            }

        } catch (IOException e) {
            e.printStackTrace();
        }
        
        }