List of usage examples for org.apache.pdfbox.cos COSDocument close
@Override public void close() throws IOException
From source file:ExtractTextFromPdf.java
public static void main(String[] args) { PDFParser parser = null;//from ww w. ja v a 2s . com PDDocument pdDoc = null; COSDocument cosDoc = null; PDFTextStripper pdfStripper; String parsedText; String fileName = "C:/Users/Kavya Gupta/Desktop/Texas_Title.pdf"; File file = new File(fileName); try { byte data[] = new byte[1024]; ((RandomAccessRead) file).read(data, 0, 1024); pdDoc = PDDocument.load(new File(fileName)); pdfStripper = new PDFTextStripper(); parsedText = pdfStripper.getText(pdDoc); System.out.println(parsedText.replaceAll("[^A-Za-z0-9. ]+", "")); } catch (Exception e) { e.printStackTrace(); try { if (cosDoc != null) cosDoc.close(); if (pdDoc != null) pdDoc.close(); } catch (Exception e1) { e.printStackTrace(); } } }
From source file:at.gv.egiz.pdfas.lib.impl.signing.pdfbox.PADESPDFBOXSigner.java
License:EUPL
public void signPDF(PDFObject genericPdfObject, RequestedSignature requestedSignature, PDFASSignatureInterface genericSigner) throws PdfAsException { String fisTmpFile = null;// ww w . j a va 2s . c om PDFAsVisualSignatureProperties properties = null; if (!(genericPdfObject instanceof PDFBOXObject)) { // tODO: throw new PdfAsException(); } PDFBOXObject pdfObject = (PDFBOXObject) genericPdfObject; if (!(genericSigner instanceof PDFASPDFBOXSignatureInterface)) { // tODO: throw new PdfAsException(); } PDFASPDFBOXSignatureInterface signer = (PDFASPDFBOXSignatureInterface) genericSigner; TempFileHelper helper = pdfObject.getStatus().getTempFileHelper(); PDDocument doc = pdfObject.getDocument(); SignatureOptions options = new SignatureOptions(); COSDocument visualSignatureDocumentGuard = null; try { fisTmpFile = helper.getStaticFilename(); FileOutputStream tmpOutputStream = null; try { // write to temporary file tmpOutputStream = new FileOutputStream(new File(fisTmpFile)); InputStream tmpis = null; try { tmpis = pdfObject.getOriginalDocument().getInputStream(); IOUtils.copy(tmpis, tmpOutputStream); tmpis.close(); } finally { IOUtils.closeQuietly(tmpis); } SignaturePlaceholderData signaturePlaceholderData = PlaceholderFilter .checkPlaceholderSignature(pdfObject.getStatus(), pdfObject.getStatus().getSettings()); TablePos tablePos = null; if (signaturePlaceholderData != null) { // Placeholder found! logger.info("Placeholder data found."); if (signaturePlaceholderData.getProfile() != null) { logger.debug("Placeholder Profile set to: " + signaturePlaceholderData.getProfile()); requestedSignature.setSignatureProfileID(signaturePlaceholderData.getProfile()); } tablePos = signaturePlaceholderData.getTablePos(); if (tablePos != null) { SignatureProfileConfiguration signatureProfileConfiguration = pdfObject.getStatus() .getSignatureProfileConfiguration(requestedSignature.getSignatureProfileID()); float minWidth = signatureProfileConfiguration.getMinWidth(); if (minWidth > 0) { if (tablePos.getWidth() < minWidth) { tablePos.width = minWidth; logger.debug("Correcting placeholder with to minimum width {}", minWidth); } } logger.debug("Placeholder Position set to: " + tablePos.toString()); } } PDSignature signature = new PDSignature(); signature.setFilter(COSName.getPDFName(signer.getPDFFilter())); // default // filter signature.setSubFilter(COSName.getPDFName(signer.getPDFSubFilter())); SignatureProfileSettings signatureProfileSettings = TableFactory.createProfile( requestedSignature.getSignatureProfileID(), pdfObject.getStatus().getSettings()); ValueResolver resolver = new ValueResolver(requestedSignature, pdfObject.getStatus()); String signerName = resolver.resolve("SIG_SUBJECT", signatureProfileSettings.getValue("SIG_SUBJECT"), signatureProfileSettings); signature.setName(signerName); // take signing time from provided signer... signature.setSignDate(signer.getSigningDate()); // ...and update operation status in order to use exactly this date for the complete signing process requestedSignature.getStatus().setSigningDate(signer.getSigningDate()); String signerReason = signatureProfileSettings.getSigningReason(); if (signerReason == null) { signerReason = "PAdES Signature"; } signature.setReason(signerReason); logger.debug("Signing reason: " + signerReason); logger.debug("Signing @ " + signer.getSigningDate().getTime().toString()); // the signing date, needed for valid signature // signature.setSignDate(signer.getSigningDate()); signer.setPDSignature(signature); int signatureSize = 0x1000; try { String reservedSignatureSizeString = signatureProfileSettings.getValue(SIG_RESERVED_SIZE); if (reservedSignatureSizeString != null) { signatureSize = Integer.parseInt(reservedSignatureSizeString); } logger.debug("Reserving {} bytes for signature", signatureSize); } catch (NumberFormatException e) { logger.warn("Invalid configuration value: {} should be a number using 0x1000", SIG_RESERVED_SIZE); } options.setPreferedSignatureSize(signatureSize); // Is visible Signature if (requestedSignature.isVisual()) { logger.debug("Creating visual signature block"); SignatureProfileConfiguration signatureProfileConfiguration = pdfObject.getStatus() .getSignatureProfileConfiguration(requestedSignature.getSignatureProfileID()); if (tablePos == null) { // ================================================================ // PositioningStage (visual) -> find position or use // fixed // position String posString = pdfObject.getStatus().getSignParamter().getSignaturePosition(); TablePos signaturePos = null; String signaturePosString = signatureProfileConfiguration.getDefaultPositioning(); if (signaturePosString != null) { logger.debug("using signature Positioning: " + signaturePos); signaturePos = new TablePos(signaturePosString); } logger.debug("using Positioning: " + posString); if (posString != null) { // Merge Signature Position tablePos = new TablePos(posString, signaturePos); } else { // Fallback to signature Position! tablePos = signaturePos; } if (tablePos == null) { // Last Fallback default position tablePos = new TablePos(); } } boolean legacy32Position = signatureProfileConfiguration.getLegacy32Positioning(); boolean legacy40Position = signatureProfileConfiguration.getLegacy40Positioning(); // create Table describtion Table main = TableFactory.createSigTable(signatureProfileSettings, MAIN, pdfObject.getStatus(), requestedSignature); IPDFStamper stamper = StamperFactory.createDefaultStamper(pdfObject.getStatus().getSettings()); IPDFVisualObject visualObject = stamper.createVisualPDFObject(pdfObject, main); /* * PDDocument originalDocument = PDDocument .load(new * ByteArrayInputStream(pdfObject.getStatus() * .getPdfObject().getOriginalDocument())); */ PositioningInstruction positioningInstruction = Positioning.determineTablePositioning(tablePos, "", doc, visualObject, legacy32Position, legacy40Position); logger.debug("Positioning: {}", positioningInstruction.toString()); if (positioningInstruction.isMakeNewPage()) { int last = doc.getNumberOfPages() - 1; PDDocumentCatalog root = doc.getDocumentCatalog(); PDPageNode rootPages = root.getPages(); List<PDPage> kids = new ArrayList<PDPage>(); rootPages.getAllKids(kids); PDPage lastPage = kids.get(last); rootPages.getCOSObject().setNeedToBeUpdate(true); PDPage p = new PDPage(lastPage.findMediaBox()); p.setResources(new PDResources()); p.setRotation(lastPage.findRotation()); doc.addPage(p); } // handle rotated page PDDocumentCatalog documentCatalog = doc.getDocumentCatalog(); PDPageNode documentPages = documentCatalog.getPages(); List<PDPage> documentPagesKids = new ArrayList<PDPage>(); documentPages.getAllKids(documentPagesKids); int targetPageNumber = positioningInstruction.getPage(); logger.debug("Target Page: " + targetPageNumber); // rootPages.getAllKids(kids); PDPage targetPage = documentPagesKids.get(targetPageNumber - 1); int rot = targetPage.findRotation(); logger.debug("Page rotation: " + rot); // positioningInstruction.setRotation(positioningInstruction.getRotation() // + rot); logger.debug("resulting Sign rotation: " + positioningInstruction.getRotation()); SignaturePositionImpl position = new SignaturePositionImpl(); position.setX(positioningInstruction.getX()); position.setY(positioningInstruction.getY()); position.setPage(positioningInstruction.getPage()); position.setHeight(visualObject.getHeight()); position.setWidth(visualObject.getWidth()); requestedSignature.setSignaturePosition(position); properties = new PDFAsVisualSignatureProperties(pdfObject.getStatus().getSettings(), pdfObject, (PdfBoxVisualObject) visualObject, positioningInstruction, signatureProfileSettings); properties.buildSignature(); /* * ByteArrayOutputStream sigbos = new * ByteArrayOutputStream(); * sigbos.write(StreamUtils.inputStreamToByteArray * (properties .getVisibleSignature())); sigbos.close(); */ if (signaturePlaceholderData != null) { // Placeholder found! // replace placeholder InputStream is = null; try { is = PADESPDFBOXSigner.class.getResourceAsStream("/placeholder/empty.jpg"); PDJpeg img = new PDJpeg(doc, is); img.getCOSObject().setNeedToBeUpdate(true); PDDocumentCatalog root = doc.getDocumentCatalog(); PDPageNode rootPages = root.getPages(); List<PDPage> kids = new ArrayList<PDPage>(); rootPages.getAllKids(kids); int pageNumber = positioningInstruction.getPage(); // rootPages.getAllKids(kids); PDPage page = kids.get(pageNumber - 1); logger.info("Placeholder name: " + signaturePlaceholderData.getPlaceholderName()); COSDictionary xobjectsDictionary = (COSDictionary) page.findResources() .getCOSDictionary().getDictionaryObject(COSName.XOBJECT); xobjectsDictionary.setItem(signaturePlaceholderData.getPlaceholderName(), img); xobjectsDictionary.setNeedToBeUpdate(true); page.findResources().getCOSObject().setNeedToBeUpdate(true); logger.info("Placeholder name: " + signaturePlaceholderData.getPlaceholderName()); } finally { IOUtils.closeQuietly(is); } } if (signatureProfileSettings.isPDFA()) { PDDocumentCatalog root = doc.getDocumentCatalog(); COSBase base = root.getCOSDictionary().getItem(COSName.OUTPUT_INTENTS); if (base == null) { InputStream colorProfile = null; try { colorProfile = PDDocumentCatalog.class .getResourceAsStream("/icm/sRGB Color Space Profile.icm"); try { PDOutputIntent oi = new PDOutputIntent(doc, colorProfile); oi.setInfo("sRGB IEC61966-2.1"); oi.setOutputCondition("sRGB IEC61966-2.1"); oi.setOutputConditionIdentifier("sRGB IEC61966-2.1"); oi.setRegistryName("http://www.color.org"); root.addOutputIntent(oi); root.getCOSObject().setNeedToBeUpdate(true); logger.info("added Output Intent"); } catch (Throwable e) { throw new PdfAsException("Failed to add Output Intent", e); } } finally { IOUtils.closeQuietly(colorProfile); } } } options.setPage(positioningInstruction.getPage()); options.setVisualSignature(properties.getVisibleSignature()); } visualSignatureDocumentGuard = options.getVisualSignature(); doc.addSignature(signature, signer, options); // set need to update indirect fields array in acro form COSDictionary trailer = doc.getDocument().getTrailer(); if (trailer != null) { COSDictionary troot = (COSDictionary) trailer.getDictionaryObject(COSName.ROOT); if (troot != null) { COSDictionary acroForm = (COSDictionary) troot.getDictionaryObject(COSName.ACRO_FORM); if (acroForm != null) { COSArray tfields = (COSArray) acroForm.getDictionaryObject(COSName.FIELDS); if (tfields != null && !tfields.isDirect()) { tfields.setNeedToBeUpdate(true); } } } } String sigFieldName = signatureProfileSettings.getSignFieldValue(); if (sigFieldName == null) { sigFieldName = "PDF-AS Signatur"; } int count = PdfBoxUtils.countSignatures(doc, sigFieldName); sigFieldName = sigFieldName + count; PDAcroForm acroFormm = doc.getDocumentCatalog().getAcroForm(); PDSignatureField signatureField = null; if (acroFormm != null) { @SuppressWarnings("unchecked") List<PDField> fields = acroFormm.getFields(); if (fields != null) { for (PDField pdField : fields) { if (pdField != null) { if (pdField instanceof PDSignatureField) { PDSignatureField tmpSigField = (PDSignatureField) pdField; if (tmpSigField.getSignature() != null && tmpSigField.getSignature().getDictionary() != null) { if (tmpSigField.getSignature().getDictionary() .equals(signature.getDictionary())) { signatureField = (PDSignatureField) pdField; } } } } } } else { logger.warn("Failed to name Signature Field! [Cannot find Field list in acroForm!]"); } if (signatureField != null) { signatureField.setPartialName(sigFieldName); } if (properties != null) { signatureField.setAlternateFieldName(properties.getAlternativeTableCaption()); } else { signatureField.setAlternateFieldName(sigFieldName); } } else { logger.warn("Failed to name Signature Field! [Cannot find acroForm!]"); } // PDF-UA logger.info("Adding pdf/ua content."); try { PDDocumentCatalog root = doc.getDocumentCatalog(); PDStructureTreeRoot structureTreeRoot = root.getStructureTreeRoot(); if (structureTreeRoot != null) { logger.info("Tree Root: {}", structureTreeRoot.toString()); List<Object> kids = structureTreeRoot.getKids(); if (kids == null) { logger.info("No kid-elements in structure tree Root, maybe not PDF/UA document"); } PDStructureElement docElement = null; for (Object k : kids) { if (k instanceof PDStructureElement) { docElement = (PDStructureElement) k; break; } } PDStructureElement sigBlock = new PDStructureElement("Form", docElement); // create object dictionary and add as child element COSDictionary objectDic = new COSDictionary(); objectDic.setName("Type", "OBJR"); objectDic.setItem("Pg", signatureField.getWidget().getPage()); objectDic.setItem("Obj", signatureField.getWidget()); List<Object> l = new ArrayList<Object>(); l.add(objectDic); sigBlock.setKids(l); sigBlock.setPage(signatureField.getWidget().getPage()); sigBlock.setTitle("Signature Table"); sigBlock.setParent(docElement); docElement.appendKid(sigBlock); // Create and add Attribute dictionary to mitigate PAC // warning COSDictionary sigBlockDic = (COSDictionary) sigBlock.getCOSObject(); COSDictionary sub = new COSDictionary(); sub.setName("O", "Layout"); sub.setName("Placement", "Block"); sigBlockDic.setItem(COSName.A, sub); sigBlockDic.setNeedToBeUpdate(true); // Modify number tree PDNumberTreeNode ntn = structureTreeRoot.getParentTree(); int parentTreeNextKey = structureTreeRoot.getParentTreeNextKey(); if (ntn == null) { ntn = new PDNumberTreeNode(objectDic, null); logger.info("No number-tree-node found!"); } COSArray ntnKids = (COSArray) ntn.getCOSDictionary().getDictionaryObject(COSName.KIDS); COSArray ntnNumbers = (COSArray) ntn.getCOSDictionary().getDictionaryObject(COSName.NUMS); if (ntnNumbers == null && ntnKids != null) {//no number array, so continue with the kids array //create dictionary with limits and nums array COSDictionary pTreeEntry = new COSDictionary(); COSArray limitsArray = new COSArray(); //limits for exact one entry limitsArray.add(COSInteger.get(parentTreeNextKey)); limitsArray.add(COSInteger.get(parentTreeNextKey)); COSArray numsArray = new COSArray(); numsArray.add(COSInteger.get(parentTreeNextKey)); numsArray.add(sigBlock); pTreeEntry.setItem(COSName.NUMS, numsArray); pTreeEntry.setItem(COSName.LIMITS, limitsArray); PDNumberTreeNode newKidsElement = new PDNumberTreeNode(pTreeEntry, PDNumberTreeNode.class); ntnKids.add(newKidsElement); ntnKids.setNeedToBeUpdate(true); //working // List<PDNumberTreeNode> treeRootKids = structureTreeRoot.getParentTree().getKids(); // PDNumberTreeNode last = (PDNumberTreeNode)treeRootKids.get(treeRootKids.size()-1); // COSArray lim1 = (COSArray) last.getCOSDictionary().getDictionaryObject(COSName.LIMITS); // lim1.remove(1); // lim1.add(1, COSInteger.get(parentTreeNextKey)); // PDNumberTreeNode verylast = (PDNumberTreeNode)last.getKids().get(last.getKids().size()-1); // COSArray numa = (COSArray) verylast.getCOSDictionary().getDictionaryObject(COSName.NUMS); // COSArray lim = (COSArray) verylast.getCOSDictionary().getDictionaryObject(COSName.LIMITS); // lim.remove(1); // lim.add(1, COSInteger.get(parentTreeNextKey)); // // int size = numa.size(); // numa.add(size, COSInteger.get(parentTreeNextKey)); // numa.add(sigBlock); //working end } else if (ntnNumbers != null && ntnKids == null) { int arrindex = ntnNumbers.size(); ntnNumbers.add(arrindex, COSInteger.get(parentTreeNextKey)); ntnNumbers.add(arrindex + 1, sigBlock.getCOSObject()); ntnNumbers.getCOSObject().setNeedToBeUpdate(true); structureTreeRoot.setParentTree(ntn); } else if (ntnNumbers == null && ntnKids == null) { //document is not pdfua conform before signature creation throw new PdfAsException("error.pdf.sig.pdfua.1"); } else { //this is not allowed throw new PdfAsException("error.pdf.sig.pdfua.1"); } // set StructureParent for signature field annotation signatureField.getWidget().setStructParent(parentTreeNextKey); //Increase the next Key value in the structure tree root structureTreeRoot.setParentTreeNextKey(parentTreeNextKey + 1); // add the Tabs /S Element for Tabbing through annots PDPage p = signatureField.getWidget().getPage(); p.getCOSDictionary().setName("Tabs", "S"); p.getCOSObject().setNeedToBeUpdate(true); //check alternative signature field name if (signatureField != null) { if (signatureField.getAlternateFieldName().equals("")) signatureField.setAlternateFieldName(sigFieldName); } ntn.getCOSDictionary().setNeedToBeUpdate(true); sigBlock.getCOSObject().setNeedToBeUpdate(true); structureTreeRoot.getCOSObject().setNeedToBeUpdate(true); objectDic.getCOSObject().setNeedToBeUpdate(true); docElement.getCOSObject().setNeedToBeUpdate(true); } } catch (Throwable e) { if (signatureProfileSettings.isPDFUA() == true) { logger.error("Could not create PDF-UA conform document!"); throw new PdfAsException("error.pdf.sig.pdfua.1", e); } else { logger.info("Could not create PDF-UA conform signature"); } } try { applyFilter(doc, requestedSignature); } catch (PDFASError e) { throw new PdfAsErrorCarrier(e); } FileInputStream tmpFileIs = null; try { tmpFileIs = new FileInputStream(new File(fisTmpFile)); synchronized (doc) { doc.saveIncremental(tmpFileIs, tmpOutputStream); } tmpFileIs.close(); } finally { IOUtils.closeQuietly(tmpFileIs); if (options != null) { if (options.getVisualSignature() != null) { options.getVisualSignature().close(); } } } tmpOutputStream.flush(); tmpOutputStream.close(); } finally { IOUtils.closeQuietly(tmpOutputStream); COSDocument visualSignature = options.getVisualSignature(); if (visualSignature != null) { visualSignature.close(); } } FileInputStream readReadyFile = null; try { readReadyFile = new FileInputStream(new File(fisTmpFile)); // write to resulting output stream // ByteArrayOutputStream bos = new ByteArrayOutputStream(); // bos.write(); // bos.close(); pdfObject.setSignedDocument(StreamUtils.inputStreamToByteArray(readReadyFile)); readReadyFile.close(); } finally { IOUtils.closeQuietly(readReadyFile); } } catch (IOException e) { logger.info(MessageResolver.resolveMessage("error.pdf.sig.01") + ": {}", e.toString()); throw new PdfAsException("error.pdf.sig.01", e); } catch (SignatureException e) { logger.info(MessageResolver.resolveMessage("error.pdf.sig.01") + ": {}", e.toString()); throw new PdfAsException("error.pdf.sig.01", e); } catch (COSVisitorException e) { logger.info(MessageResolver.resolveMessage("error.pdf.sig.01") + ": {}", e.toString()); throw new PdfAsException("error.pdf.sig.01", e); } finally { if (doc != null) { try { doc.close(); } catch (IOException e) { logger.debug("Failed to close COS Doc!", e); // Ignore } } if (fisTmpFile != null) { helper.deleteFile(fisTmpFile); } logger.debug("Signature done!"); } }
From source file:com.cisco.iwe.services.util.EmailMonitor.java
/** * /*ww w . ja v a2 s .co m*/ * @param fileDir * @return */ /* This method is used to scan the uploaded expense receipt in .pdf format and extract the text embedded in it. */ public String scanPDF(String fileDir) { PDFParser parser; String parsedText = null; PDFTextStripper pdfStripper = null; PDDocument pdDoc = null; COSDocument cosDoc = null; File file = new File(fileDir); if (!file.isFile()) { System.err.println("File " + fileDir + " does not exist."); return null; } try { parser = new PDFParser(new FileInputStream(file)); } catch (IOException e) { System.err.println("Unable to open PDF Parser. " + e.getMessage()); return null; } try { parser.parse(); cosDoc = parser.getDocument(); pdfStripper = new PDFTextStripper(); pdDoc = new PDDocument(cosDoc); pdfStripper.setStartPage(1); pdfStripper.setEndPage(pdDoc.getNumberOfPages()); parsedText = pdfStripper.getText(pdDoc); } catch (Exception e) { System.err.println("An exception occured in parsing the PDF Document." + e.getMessage()); } finally { try { if (cosDoc != null) cosDoc.close(); if (pdDoc != null) pdDoc.close(); } catch (Exception e) { e.printStackTrace(); } } return parsedText; }
From source file:com.lanacion.adminsiteln.services.PdfIndexerService.PdfIndexerService.java
/** * Metodos privados para la indexacin//from w w w . ja v a 2 s . c o m */ private String pdftoText(String fileName, int pagina) { PDFParser parser; String parsedText = null; ; PDFTextStripper pdfStripper = null; //pdfStripper.setStartPage(0); //pdfStripper.setEndPage(0); PDDocument pdDoc = null; COSDocument cosDoc = null; File file = new File(fileName); if (!file.isFile()) { System.err.println("File " + fileName + " does not exist."); return null; } try { parser = new PDFParser(new FileInputStream(file)); } catch (IOException e) { System.err.println("Unable to open PDF Parser. " + e.getMessage()); return null; } try { parser.parse(); cosDoc = parser.getDocument(); pdfStripper = new PDFTextStripper(); pdDoc = new PDDocument(cosDoc); pdfStripper.setStartPage(pagina); pdfStripper.setEndPage(pagina); parsedText = pdfStripper.getText(pdDoc); } catch (Exception e) { System.err.println("An exception occured in parsing the PDF Document." + e.getMessage()); } finally { try { if (cosDoc != null) { cosDoc.close(); } if (pdDoc != null) { pdDoc.close(); } } catch (Exception e) { e.printStackTrace(); } } return parsedText; }
From source file:com.lanacion.adminsiteln.services.PdfIndexerService.PdfIndexerService.java
private int pdfgetPages(String fileName) { int numero_paginas = 0; PDFParser parser;//from w ww . jav a 2s . c o m String parsedText = null; ; PDFTextStripper pdfStripper = null; //pdfStripper.setStartPage(0); //pdfStripper.setEndPage(0); PDDocument pdDoc = null; COSDocument cosDoc = null; File file = new File(fileName); if (!file.isFile()) { System.err.println("File " + fileName + " does not exist."); return 0; } try { parser = new PDFParser(new FileInputStream(file)); } catch (IOException e) { System.err.println("Unable to open PDF Parser. " + e.getMessage()); return 0; } try { parser.parse(); cosDoc = parser.getDocument(); pdfStripper = new PDFTextStripper(); pdDoc = new PDDocument(cosDoc); numero_paginas = pdDoc.getNumberOfPages(); } catch (Exception e) { System.err.println("An exception occured in parsing the PDF Document." + e.getMessage()); } finally { try { if (cosDoc != null) { cosDoc.close(); } if (pdDoc != null) { pdDoc.close(); } } catch (Exception e) { e.printStackTrace(); } } return numero_paginas; }
From source file:com.validation.manager.core.server.core.AttachmentServerTest.java
License:Apache License
/** * Test of addFile method, of class AttachmentServer. *///from w w w. j a va 2 s .c o m @Test public void testAddRetrieveTextFile() { try { System.out.println("add text File"); File f = new File("target/Test.txt"); f.deleteOnExit(); List<String> lines = Arrays.asList("The first line", "The second line"); Path file = Paths.get(f.getAbsolutePath()); Files.write(file, lines, Charset.forName("UTF-8")); AttachmentServer instance = new AttachmentServer(); instance.addFile(f, f.getName()); instance.write2DB(); //Delete the file FileUtils.delete(f.getAbsolutePath()); assertEquals(1, (int) instance.getAttachmentType().getId());//Text file System.out.println("retrieveFile"); AttachmentServer temp = new AttachmentServer(instance.getAttachmentPK()); File loadedFile = temp.getAttachedFile("target/loaded/"); BufferedReader br = new BufferedReader(new FileReader(loadedFile)); String line; int count = 0; while ((line = br.readLine()) != null) { assertEquals(lines.get(count), line); System.out.println(line); count++; } assertEquals(lines.size(), count); //Create pdf file System.out.println("add pdf File"); File pdf = Tool.convertToPDF(loadedFile, "target/Text.pdf"); pdf.deleteOnExit(); instance = new AttachmentServer(); instance.addFile(pdf, pdf.getName()); instance.write2DB(); //Delete the file FileUtils.delete(pdf.getAbsolutePath()); assertEquals(2, (int) instance.getAttachmentType().getId());//PDF file System.out.println("retrieveFile"); temp = new AttachmentServer(instance.getAttachmentPK()); loadedFile = temp.getAttachedFile("target/loaded/"); PDFTextStripper pdfStripper; PDDocument pdDoc = null; COSDocument cosDoc = null; try { PDFParser parser = new PDFParser(new RandomAccessBufferedFileInputStream(loadedFile)); parser.parse(); cosDoc = parser.getDocument(); pdfStripper = new PDFTextStripper(); pdDoc = new PDDocument(cosDoc); pdfStripper.setStartPage(1); pdfStripper.setEndPage(1); String parsedText = pdfStripper.getText(pdDoc); System.out.println(parsedText); } catch (IOException ex) { Exceptions.printStackTrace(ex); fail(); } finally { if (cosDoc != null) { cosDoc.close(); } if (pdDoc != null) { pdDoc.close(); } } } catch (IOException | VMException ex) { Exceptions.printStackTrace(ex); fail(); } }
From source file:cz.muni.pdfjbim.PdfImageExtractor.java
License:Apache License
/** * This method extracts images by going through all COSObjects pointed from xref table * @param is input stream containing PDF file * @param prefix output basename for images * @param password password for access to PDF if needed * @param pagesToProcess list of pages which should be processed if null given => processed all pages * -- not working yet/* ww w . ja v a 2 s. c o m*/ * @param binarize -- enables processing of nonbitonal images as well (LZW is still not * processed because of output with inverted colors) * @throws PdfRecompressionException if problem to extract images from PDF */ public void extractImagesUsingPdfParser(InputStream is, String prefix, String password, Set<Integer> pagesToProcess, Boolean binarize) throws PdfRecompressionException { // checking arguments and setting appropriate variables if (binarize == null) { binarize = false; } log.debug("Extracting images (binarize set to {})", binarize); InputStream inputStream = null; if (password != null) { try (ByteArrayOutputStream decryptedOutputStream = new ByteArrayOutputStream()) { PdfReader reader = new PdfReader(is, password.getBytes(StandardCharsets.UTF_8)); PdfStamper stamper = new PdfStamper(reader, decryptedOutputStream); if (stamper != null) { stamper.close(); } inputStream = new ByteArrayInputStream(decryptedOutputStream.toByteArray()); } catch (DocumentException ex) { throw new PdfRecompressionException(ex); } catch (IOException ex) { throw new PdfRecompressionException("Reading file caused exception", ex); } } else { inputStream = is; } PDFParser parser = null; COSDocument doc = null; try { parser = new PDFParser(inputStream); parser.parse(); doc = parser.getDocument(); List<COSObject> objs = doc.getObjectsByType(COSName.XOBJECT); if (objs != null) { for (COSObject obj : objs) { COSBase subtype = obj.getItem(COSName.SUBTYPE); if (subtype.toString().equalsIgnoreCase("COSName{Image}")) { COSBase imageObj = obj.getObject(); COSBase cosNameObj = obj.getItem(COSName.NAME); String key; if (cosNameObj != null) { String cosNameKey = cosNameObj.toString(); int startOfKey = cosNameKey.indexOf("{") + 1; key = cosNameKey.substring(startOfKey, cosNameKey.length() - 1); } else { key = "im0"; } int objectNum = obj.getObjectNumber().intValue(); int genNum = obj.getGenerationNumber().intValue(); PDXObjectImage image = (PDXObjectImage) PDXObjectImage.createXObject(imageObj); PDStream pdStr = new PDStream(image.getCOSStream()); List<COSName> filters = pdStr.getFilters(); log.debug("Detected image with color depth: {} bits", image.getBitsPerComponent()); if (filters == null) { continue; } log.debug("Detected filters: {}", filters.toString()); if ((image.getBitsPerComponent() > 1) && (!binarize)) { log.info("It is not a bitonal image => skipping"); continue; } // at this moment for preventing bad output (bad coloring) from LZWDecode filter if (filters.contains(COSName.LZW_DECODE)) { log.info("This is LZWDecoded => skipping"); continue; } if (filters.contains(COSName.FLATE_DECODE)) { log.debug("FlateDecoded image detected"); } if (filters.contains(COSName.JBIG2_DECODE)) { if (skipJBig2Images) { log.warn("Allready compressed according to JBIG2 standard => skipping"); continue; } else { log.debug("JBIG2 image detected"); } } // detection of unsupported filters by pdfBox library if (filters.contains(COSName.JPX_DECODE)) { log.warn("Unsupported filter JPXDecode => skipping"); continue; } String name = getUniqueFileName(prefix, image.getSuffix()); log.info("Writing image: {}", name); image.write2file(name); PdfImageInformation pdfImageInfo = new PdfImageInformation(key, image.getWidth(), image.getHeight(), objectNum, genNum); originalImageInformations.add(pdfImageInfo); namesOfImages.add(name + "." + image.getSuffix()); } } } } catch (IOException ex) { Tools.deleteFilesFromList(namesOfImages); throw new PdfRecompressionException("Unable to parse PDF document", ex); } catch (Exception ex) { Tools.deleteFilesFromList(namesOfImages); } finally { if (doc != null) { try { doc.close(); } catch (IOException ex) { throw new PdfRecompressionException(ex); } } } }
From source file:cz.muni.pdfjbim.PdfImageProcessor.java
License:Apache License
/** * This method extracts images by going through all COSObjects pointed from xref table * @param is input stream containing PDF file * @param password password for access to PDF if needed * @param pagesToProcess list of pages which should be processed if null given => processed all pages * -- not working yet/* ww w.j a va2 s . c o m*/ * @param binarize -- enables processing of nonbitonal images as well (LZW is still not * processed because of output with inverted colors) * @throws PdfRecompressionException if problem to extract images from PDF */ public void extractImagesUsingPdfParser(InputStream is, String prefix, String password, Set<Integer> pagesToProcess, Boolean binarize) throws PdfRecompressionException { // checking arguments and setting appropriate variables if (binarize == null) { binarize = false; } InputStream inputStream = null; if (password != null) { try { ByteArrayOutputStream decryptedOutputStream = null; PdfReader reader = new PdfReader(is, password.getBytes()); PdfStamper stamper = new PdfStamper(reader, decryptedOutputStream); stamper.close(); inputStream = new ByteArrayInputStream(decryptedOutputStream.toByteArray()); } catch (DocumentException ex) { throw new PdfRecompressionException(ex); } catch (IOException ex) { throw new PdfRecompressionException("Reading file caused exception", ex); } } else { inputStream = is; } PDFParser parser = null; COSDocument doc = null; try { parser = new PDFParser(inputStream); parser.parse(); doc = parser.getDocument(); List<COSObject> objs = doc.getObjectsByType(COSName.XOBJECT); if (objs != null) { for (COSObject obj : objs) { COSBase subtype = obj.getItem(COSName.SUBTYPE); if (subtype.toString().equalsIgnoreCase("COSName{Image}")) { COSBase imageObj = obj.getObject(); COSBase cosNameObj = obj.getItem(COSName.NAME); String key; if (cosNameObj != null) { String cosNameKey = cosNameObj.toString(); int startOfKey = cosNameKey.indexOf("{") + 1; key = cosNameKey.substring(startOfKey, cosNameKey.length() - 1); } else { key = "im0"; } int objectNum = obj.getObjectNumber().intValue(); int genNum = obj.getGenerationNumber().intValue(); PDXObjectImage image = (PDXObjectImage) PDXObjectImage.createXObject(imageObj); PDStream pdStr = new PDStream(image.getCOSStream()); List filters = pdStr.getFilters(); if ((image.getBitsPerComponent() > 1) && (!binarize)) { log.info("It is not a bitonal image => skipping"); continue; } // at this moment for preventing bad output (bad coloring) from LZWDecode filter if (filters.contains(COSName.LZW_DECODE.getName())) { log.info("This is LZWDecoded => skipping"); continue; } // detection of unsupported filters by pdfBox library if (filters.contains("JBIG2Decode")) { log.warn("Allready compressed according to JBIG2 standard => skipping"); continue; } if (filters.contains("JPXDecode")) { log.warn("Unsupported filter JPXDecode => skipping"); continue; } String name = getUniqueFileName(prefix, image.getSuffix()); log.info("Writing image:" + name); image.write2file(name); PdfImageInformation pdfImageInfo = new PdfImageInformation(key, image.getWidth(), image.getHeight(), objectNum, genNum); originalImageInformations.add(pdfImageInfo); namesOfImages.add(name + "." + image.getSuffix()); } // } } } } catch (IOException ex) { throw new PdfRecompressionException("Unable to parse PDF document", ex); } finally { if (doc != null) { try { doc.close(); } catch (IOException ex) { throw new PdfRecompressionException(ex); } } } }
From source file:cz.mzk.editor.server.handler.GetOcrFromPdfHandler.java
License:Open Source License
private String pdftoText(String fileName) throws ActionException { File pdfFile = new File(fileName); if (!pdfFile.isFile()) { LOGGER.error("The file: " + fileName + " does not exist."); throw new ActionException("Unable to parse the pdf file."); }/*from ww w. j a v a 2s . c o m*/ PDFParser parser = null; COSDocument cosDoc = null; PDFTextStripper pdfStripper; PDDocument pdDoc = null; String parsedText; try { parser = new PDFParser(new RandomAccessBufferedFileInputStream(new FileInputStream(pdfFile))); } catch (Exception e) { LOGGER.error("Unable to open PDF Parser.: " + e); e.printStackTrace(); throw new ActionException("Unable to parse the pdf file."); } try { parser.parse(); cosDoc = parser.getDocument(); pdfStripper = new PDFTextStripper(); pdDoc = new PDDocument(cosDoc); parsedText = pdfStripper.getText(pdDoc); } catch (Exception e) { LOGGER.error("An exception occured in parsing the PDF Document."); e.printStackTrace(); throw new ActionException("Unable to parse the pdf file. " + e); } finally { try { if (cosDoc != null) cosDoc.close(); if (pdDoc != null) pdDoc.close(); } catch (Exception e) { e.printStackTrace(); } } return parsedText; }
From source file:de.hsmannheim.ss15.alr.searchengine.PDFParser.java
public String getTextOfPDF(byte[] in) throws Exception { ByteArrayInputStream input = new ByteArrayInputStream(in); org.apache.pdfbox.pdfparser.PDFParser parser; String parsedText = null;//from w w w. j a v a2s . c om ; PDFTextStripper pdfStripper = null; PDDocument pdDoc = null; COSDocument cosDoc = null; parser = new NonSequentialPDFParser(input); //parse PDF try { parser.parse(); cosDoc = parser.getDocument(); pdfStripper = new PDFTextStripper(); pdDoc = new PDDocument(cosDoc); parsedText = pdfStripper.getText(pdDoc); } catch (Exception e) { throw (e); } finally { if (cosDoc != null) { cosDoc.close(); } if (pdDoc != null) { pdDoc.close(); } } return parsedText; }