Closes #53: Build thumbnails for DOCX and ODT files

README.md

@@ -21,7 +21,7 @@ Features
 
 - Responsive user interface
 - Optical character recognition
-- Support image and PDF files
+- Support image, PDF, ODT and DOCX files
 - Flexible search engine
 - Full text search in image and PDF
 - 256-bit AES encryption

docs-core/src/main/java/com/sismics/docs/core/event/FileCreatedAsyncEvent.java

@@ -28,58 +28,43 @@ public class FileCreatedAsyncEvent {
     private InputStream inputStream;
     
     /**
-     * Getter of file.
+     * Unencrypted input stream containing a PDF representation
-     *
+     * of the file. May be null if the PDF conversion is not
-     * @return the file
+     * necessary or not possible.
      */
+    private InputStream pdfInputStream;
+    
     public File getFile() {
         return file;
     }
 
-    /**
-     * Setter of file.
-     *
-     * @param file file
-     */
     public void setFile(File file) {
         this.file = file;
     }
     
-    /**
-     * Getter of document.
-     *
-     * @return the document
-     */
     public Document getDocument() {
         return document;
     }
 
-    /**
-     * Setter of document.
-     *
-     * @param document document
-     */
     public void setDocument(Document document) {
         this.document = document;
     }
     
-    /**
-     * Getter of inputStream.
-     *
-     * @return the inputStream
-     */
     public InputStream getInputStream() {
         return inputStream;
     }
 
-    /**
-     * Setter de inputStream.
-     *
-     * @param inputStream inputStream
-     */
     public void setInputStream(InputStream inputStream) {
         this.inputStream = inputStream;
     }
+    
+    public InputStream getPdfInputStream() {
+        return pdfInputStream;
+    }
+
+    public void setPdfInputStream(InputStream pdfInputStream) {
+        this.pdfInputStream = pdfInputStream;
+    }
 
     @Override
     public String toString() {

docs-core/src/main/java/com/sismics/docs/core/listener/async/FileCreatedAsyncListener.java

@@ -12,7 +12,6 @@ import com.sismics.docs.core.event.FileCreatedAsyncEvent;
 import com.sismics.docs.core.model.jpa.File;
 import com.sismics.docs.core.util.FileUtil;
 import com.sismics.docs.core.util.TransactionUtil;
-import com.sismics.util.mime.MimeTypeUtil;
 
 /**
  * Listener on file created.
@@ -39,12 +38,15 @@ public class FileCreatedAsyncListener {
 
         // Guess the mime type a second time, for open document format (first detected as simple ZIP file)
         final File file = fileCreatedAsyncEvent.getFile();
-        file.setMimeType(MimeTypeUtil.guessOpenDocumentFormat(file, fileCreatedAsyncEvent.getInputStream()));
         
         // Extract text content from the file
         long startTime = System.currentTimeMillis();
-        final String content = FileUtil.extractContent(fileCreatedAsyncEvent.getDocument(), file, fileCreatedAsyncEvent.getInputStream());
+        final String content = FileUtil.extractContent(fileCreatedAsyncEvent.getDocument(), file,
+                fileCreatedAsyncEvent.getInputStream(), fileCreatedAsyncEvent.getPdfInputStream());
         fileCreatedAsyncEvent.getInputStream().close();
+        if (fileCreatedAsyncEvent.getPdfInputStream() != null) {
+            fileCreatedAsyncEvent.getPdfInputStream().close();
+        }
         log.info(MessageFormat.format("File content extracted in {0}ms", System.currentTimeMillis() - startTime));
         
         // Store the text content in the database

docs-core/src/main/java/com/sismics/docs/core/util/FileUtil.java

@@ -1,6 +1,8 @@
 package com.sismics.docs.core.util;
 
 import java.awt.image.BufferedImage;
+import java.io.ByteArrayInputStream;
+import java.io.ByteArrayOutputStream;
 import java.io.IOException;
 import java.io.InputStream;
 import java.io.OutputStream;
@@ -48,19 +50,16 @@ public class FileUtil {
      * @param document Document linked to the file
      * @param file File to extract
      * @param inputStream Unencrypted input stream
+     * @param pdfInputStream Unencrypted PDF input stream
      * @return Content extract
      */
-    public static String extractContent(Document document, File file, InputStream inputStream) {
+    public static String extractContent(Document document, File file, InputStream inputStream, InputStream pdfInputStream) {
         String content = null;
         
         if (ImageUtil.isImage(file.getMimeType())) {
             content = ocrFile(inputStream, document);
-        } else if (file.getMimeType().equals(MimeType.APPLICATION_PDF)) {
+        } else if (pdfInputStream != null) {
-            content = extractPdf(inputStream);
+            content = extractPdf(pdfInputStream);
-        } else if (file.getMimeType().equals(MimeType.OPEN_DOCUMENT_TEXT)) {
-            content = extractOpenDocumentText(inputStream);
-        } else if (file.getMimeType().equals(MimeType.OFFICE_DOCUMENT)) {
-            content = extractOfficeDocument(inputStream);
         }
         
         return content;
@@ -129,92 +128,80 @@ public class FileUtil {
     }
     
     /**
-     * Extract text from an open document text file.
+     * Convert a file to PDF if necessary.
      * 
-     * @param inputStream Unencrypted input stream
+     * @param inputStream InputStream
-     * @return Content extracted
+     * @param file File
+     * @return PDF input stream
+     * @throws Exception 
      */
-    private static String extractOpenDocumentText(InputStream inputStream) {
+    public static InputStream convertToPdf(InputStream inputStream, File file) throws Exception {
-        String content = null;
+        if (file.getMimeType().equals(MimeType.APPLICATION_PDF)) {
-        Path tempFile = null;
+            // It's already PDF, just return the input
-        try {
+            return inputStream;
-            // Convert the ODT file to a temporary PDF file
-            tempFile = Files.createTempFile("sismicsdocs_", ".pdf");
-            try (OutputStream out = Files.newOutputStream(tempFile)) {
-                OdfTextDocument document = OdfTextDocument.loadDocument(inputStream);
-                PdfOptions options = PdfOptions.create();
-                PdfConverter.getInstance().convert(document, out, options);
-            }
-            
-            // Extract content from the PDF file
-            try (InputStream pdfInputStream = Files.newInputStream(tempFile)) {
-                content = extractPdf(pdfInputStream);
-            }
-            
-        } catch (Exception e) {
-            log.error("Error while extracting text from the ODT", e);
-        } finally {
-            try {
-                Files.delete(tempFile); // Delete the temporary PDF file
-            } catch (IOException e) {
-                // Should not happen
-            }
         }
-        return content;
+        
+        if (file.getMimeType().equals(MimeType.OFFICE_DOCUMENT)) {
+            return convertOfficeDocument(inputStream);
+        }
+        
+        if (file.getMimeType().equals(MimeType.OPEN_DOCUMENT_TEXT)) {
+            return convertOpenDocumentText(inputStream);
+        }
+        
+        // PDF conversion not necessary/possible
+        return null;
     }
     
     /**
-     * Extract text from an Office document.
+     * Convert an open document text file to PDF.
      * 
      * @param inputStream Unencrypted input stream
-     * @return Content extracted
+     * @return PDF input stream
+     * @throws Exception 
      */
-    private static String extractOfficeDocument(InputStream inputStream) {
+    private static InputStream convertOpenDocumentText(InputStream inputStream) throws Exception {
-        String content = null;
+        ByteArrayOutputStream pdfOutputStream = new ByteArrayOutputStream();
-        Path tempFile = null;
+        OdfTextDocument document = OdfTextDocument.loadDocument(inputStream);
-        try {
+        PdfOptions options = PdfOptions.create();
-            // Convert the DOCX file to a temporary PDF file
+        PdfConverter.getInstance().convert(document, pdfOutputStream, options);
-            tempFile = Files.createTempFile("sismicsdocs_", ".pdf");
+        inputStream.reset();
-            try (OutputStream out = Files.newOutputStream(tempFile)) {
+        return new ByteArrayInputStream(pdfOutputStream.toByteArray());
-                XWPFDocument document = new XWPFDocument(inputStream);
+    }
-                org.apache.poi.xwpf.converter.pdf.PdfOptions options = org.apache.poi.xwpf.converter.pdf.PdfOptions.create();
+    
-                org.apache.poi.xwpf.converter.pdf.PdfConverter.getInstance().convert(document, out, options);
+    /**
-            }
+     * Convert an Office document to PDF.
-            
+     * 
-            // Extract content from the PDF file
+     * @param inputStream Unencrypted input stream
-            try (InputStream pdfInputStream = Files.newInputStream(tempFile)) {
+     * @return PDF input stream
-                content = extractPdf(pdfInputStream);
+     * @throws Exception 
-            }
+     */
-            
+    private static InputStream convertOfficeDocument(InputStream inputStream) throws Exception {
-        } catch (Exception e) {
+        ByteArrayOutputStream pdfOutputStream = new ByteArrayOutputStream();
-            log.error("Error while extracting text from the DOCX", e);
+        XWPFDocument document = new XWPFDocument(inputStream);
-        } finally {
+        org.apache.poi.xwpf.converter.pdf.PdfOptions options = org.apache.poi.xwpf.converter.pdf.PdfOptions.create();
-            try {
+        org.apache.poi.xwpf.converter.pdf.PdfConverter.getInstance().convert(document, pdfOutputStream, options);
-                Files.delete(tempFile); // Delete the temporary PDF file
+        inputStream.reset();
-            } catch (IOException e) {
+        return new ByteArrayInputStream(pdfOutputStream.toByteArray());
-                // Should not happen
-            }
-        }
-        return content;
     }
     
     /**
      * Save a file on the storage filesystem.
      * 
      * @param inputStream Unencrypted input stream
+     * @param pdf
      * @param file File to save
      * @param privateKey Private key used for encryption
      * @throws Exception
      */
-    public static void save(InputStream inputStream, File file, String privateKey) throws Exception {
+    public static void save(InputStream inputStream, InputStream pdfInputStream, File file, String privateKey) throws Exception {
         Cipher cipher = EncryptionUtil.getEncryptionCipher(privateKey);
         Path path = DirectoryUtil.getStorageDirectory().resolve(file.getId());
         Files.copy(new CipherInputStream(inputStream, cipher), path);
+        inputStream.reset();
         
         // Generate file variations
-        inputStream.reset();
+        saveVariations(file, inputStream, pdfInputStream, cipher);
-        saveVariations(file, inputStream, cipher);
-        inputStream.reset();
     }
 
     /**
@@ -222,25 +209,27 @@ public class FileUtil {
      * 
      * @param file File from database
      * @param inputStream Unencrypted input stream
+     * @param pdfInputStream Unencrypted PDF input stream
      * @param cipher Cipher to use for encryption
      * @throws Exception
      */
-    public static void saveVariations(File file, InputStream inputStream, Cipher cipher) throws Exception {
+    public static void saveVariations(File file, InputStream inputStream, InputStream pdfInputStream, Cipher cipher) throws Exception {
         BufferedImage image = null;
         if (ImageUtil.isImage(file.getMimeType())) {
             image = ImageIO.read(inputStream);
-        } else if(file.getMimeType().equals(MimeType.APPLICATION_PDF)) {
+            inputStream.reset();
+        } else if(pdfInputStream != null) {
             // Generate preview from the first page of the PDF
             PDDocument pdfDocument = null;
             try {
-                pdfDocument = PDDocument.load(inputStream);
+                pdfDocument = PDDocument.load(pdfInputStream);
                 PDFRenderer renderer = new PDFRenderer(pdfDocument);
                 image = renderer.renderImage(0);
+                pdfInputStream.reset();
             } finally {
                 pdfDocument.close();
             }
         }
-        // TODO Generate thumbnails for DOCX/ODT documents (guess the MIME type earlier and build a PDF version now?)
         
         if (image != null) {
             // Generate thumbnails from image

docs-core/src/test/java/com/sismics/docs/core/util/TestFileUtil.java

@@ -1,9 +1,11 @@
 package com.sismics.docs.core.util;
 
+import java.io.ByteArrayInputStream;
 import java.io.InputStream;
 
 import junit.framework.Assert;
 
+import org.apache.pdfbox.io.IOUtils;
 import org.junit.Test;
 
 import com.google.common.io.Resources;
@@ -18,19 +20,25 @@ import com.sismics.util.mime.MimeType;
 public class TestFileUtil {
     @Test
     public void extractContentOpenDocumentTextTest() throws Exception {
-        try (InputStream inputStream = Resources.getResource("file/document.odt").openStream()) {
+        try (InputStream inputStream = Resources.getResource("file/document.odt").openStream();
+                InputStream bytesInputStream = new ByteArrayInputStream(IOUtils.toByteArray(inputStream))) {
             File file = new File();
             file.setMimeType(MimeType.OPEN_DOCUMENT_TEXT);
-            Assert.assertEquals("Lorem ipsum dolor sit amen.\r\n", FileUtil.extractContent(null, file, inputStream));
+            try (InputStream pdfInputStream = FileUtil.convertToPdf(bytesInputStream, file)) {
+                Assert.assertEquals("Lorem ipsum dolor sit amen.\r\n", FileUtil.extractContent(null, file, inputStream, pdfInputStream));
+            }
         }
     }
     
     @Test
     public void extractContentOfficeDocumentTest() throws Exception {
-        try (InputStream inputStream = Resources.getResource("file/document.docx").openStream()) {
+        try (InputStream inputStream = Resources.getResource("file/document.docx").openStream();
+                InputStream bytesInputStream = new ByteArrayInputStream(IOUtils.toByteArray(inputStream))) {
             File file = new File();
             file.setMimeType(MimeType.OFFICE_DOCUMENT);
-            Assert.assertEquals("Lorem ipsum dolor sit amen.\r\n", FileUtil.extractContent(null, file, inputStream));
+            try (InputStream pdfInputStream = FileUtil.convertToPdf(bytesInputStream, file)) {
+                Assert.assertEquals("Lorem ipsum dolor sit amen.\r\n", FileUtil.extractContent(null, file, inputStream, pdfInputStream));
+            }
         }
     }
 }

docs-web/src/main/java/com/sismics/docs/rest/resource/FileResource.java

@@ -146,8 +146,14 @@ public class FileResource extends BaseResource {
             file.setUserId(principal.getId());
             String fileId = fileDao.create(file);
             
+            // Guess the mime type a second time, for open document format (first detected as simple ZIP file)
+            file.setMimeType(MimeTypeUtil.guessOpenDocumentFormat(file, fileInputStream));
+            
+            // Convert to PDF if necessary (for thumbnail and text extraction)
+            InputStream pdfIntputStream = FileUtil.convertToPdf(fileInputStream, file);
+            
             // Save the file
-            FileUtil.save(fileInputStream, file, user.getPrivateKey());
+            FileUtil.save(fileInputStream, pdfIntputStream, file, user.getPrivateKey());
             
             // Update the user quota
             user.setStorageCurrent(user.getStorageCurrent() + fileData.length);
@@ -159,6 +165,7 @@ public class FileResource extends BaseResource {
                 fileCreatedAsyncEvent.setDocument(document);
                 fileCreatedAsyncEvent.setFile(file);
                 fileCreatedAsyncEvent.setInputStream(fileInputStream);
+                fileCreatedAsyncEvent.setPdfInputStream(pdfIntputStream);
                 AppContext.getInstance().getAsyncEventBus().post(fileCreatedAsyncEvent);
             }
 

docs-web/src/main/webapp/src/partial/docs/document.edit.html

@@ -39,7 +39,8 @@
       <label class="col-sm-2 control-label" for="inputFiles">New files</label>
       <div class="col-sm-6">
         <file class="form-control" id="inputFiles" multiple="multiple" ng-model="newFiles"
-          accept="image/png,image/jpg,image/jpeg,image/gif,application/pdf" ng-disabled="fileIsUploading"></file>
+          accept="image/png,image/jpg,image/jpeg,image/gif,application/pdf,application/vnd.oasis.opendocument.text,application/vnd.openxmlformats-officedocument.wordprocessingml.document"
+          ng-disabled="fileIsUploading"></file>
       </div>
       <div class="col-sm-4" ng-if="orphanFiles.length > 0">
         + {{ orphanFiles.length }} file{{ orphanFiles.length > 1 ? 's' : '' }}

docs-web/src/test/java/com/sismics/docs/rest/TestDocumentResource.java

@@ -267,6 +267,124 @@ public class TestDocumentResource extends BaseJerseyTest {
         return json.getJsonArray("documents").size();
     }
     
+    /**
+     * Test ODT extraction.
+     * 
+     * @throws Exception 
+     */
+    @Test
+    public void testOdtExtraction() throws Exception {
+        // Login document_odt
+        clientUtil.createUser("document_odt");
+        String documentOdtToken = clientUtil.login("document_odt");
+
+        // Create a document
+        long create1Date = new Date().getTime();
+        JsonObject json = target().path("/document").request()
+                .cookie(TokenBasedSecurityFilter.COOKIE_NAME, documentOdtToken)
+                .put(Entity.form(new Form()
+                        .param("title", "My super title document 1")
+                        .param("description", "My super description for document 1")
+                        .param("language", "eng")
+                        .param("create_date", Long.toString(create1Date))), JsonObject.class);
+        String document1Id = json.getString("id");
+        Assert.assertNotNull(document1Id);
+        
+        // Add a PDF file
+        String file1Id = null;
+        try (InputStream is = Resources.getResource("file/document.odt").openStream()) {
+            StreamDataBodyPart streamDataBodyPart = new StreamDataBodyPart("file", is, "document.odt");
+            try (FormDataMultiPart multiPart = new FormDataMultiPart()) {
+                json = target()
+                        .register(MultiPartFeature.class)
+                        .path("/file").request()
+                        .cookie(TokenBasedSecurityFilter.COOKIE_NAME, documentOdtToken)
+                        .put(Entity.entity(multiPart.field("id", document1Id).bodyPart(streamDataBodyPart),
+                                MediaType.MULTIPART_FORM_DATA_TYPE), JsonObject.class);
+                file1Id = json.getString("id");
+                Assert.assertNotNull(file1Id);
+            }
+        }
+        
+        // Search documents by query in full content
+        json = target().path("/document/list")
+                .queryParam("search", "full:ipsum")
+                .request()
+                .cookie(TokenBasedSecurityFilter.COOKIE_NAME, documentOdtToken)
+                .get(JsonObject.class);
+        Assert.assertTrue(json.getJsonArray("documents").size() == 1);
+        
+        // Get the file thumbnail data
+        Response response = target().path("/file/" + file1Id + "/data")
+                .queryParam("size", "thumb")
+                .request()
+                .cookie(TokenBasedSecurityFilter.COOKIE_NAME, documentOdtToken)
+                .get();
+        InputStream is = (InputStream) response.getEntity();
+        byte[] fileBytes = ByteStreams.toByteArray(is);
+        Assert.assertTrue(fileBytes.length > 0); // Images rendered from PDF differ in size from OS to OS due to font issues
+        Assert.assertEquals(MimeType.IMAGE_JPEG, MimeTypeUtil.guessMimeType(fileBytes));
+    }
+    
+    /**
+     * Test DOCX extraction.
+     * 
+     * @throws Exception 
+     */
+    @Test
+    public void testDocxExtraction() throws Exception {
+        // Login document_docx
+        clientUtil.createUser("document_docx");
+        String documentDocxToken = clientUtil.login("document_docx");
+
+        // Create a document
+        long create1Date = new Date().getTime();
+        JsonObject json = target().path("/document").request()
+                .cookie(TokenBasedSecurityFilter.COOKIE_NAME, documentDocxToken)
+                .put(Entity.form(new Form()
+                        .param("title", "My super title document 1")
+                        .param("description", "My super description for document 1")
+                        .param("language", "eng")
+                        .param("create_date", Long.toString(create1Date))), JsonObject.class);
+        String document1Id = json.getString("id");
+        Assert.assertNotNull(document1Id);
+        
+        // Add a PDF file
+        String file1Id = null;
+        try (InputStream is = Resources.getResource("file/document.docx").openStream()) {
+            StreamDataBodyPart streamDataBodyPart = new StreamDataBodyPart("file", is, "document.docx");
+            try (FormDataMultiPart multiPart = new FormDataMultiPart()) {
+                json = target()
+                        .register(MultiPartFeature.class)
+                        .path("/file").request()
+                        .cookie(TokenBasedSecurityFilter.COOKIE_NAME, documentDocxToken)
+                        .put(Entity.entity(multiPart.field("id", document1Id).bodyPart(streamDataBodyPart),
+                                MediaType.MULTIPART_FORM_DATA_TYPE), JsonObject.class);
+                file1Id = json.getString("id");
+                Assert.assertNotNull(file1Id);
+            }
+        }
+        
+        // Search documents by query in full content
+        json = target().path("/document/list")
+                .queryParam("search", "full:dolor")
+                .request()
+                .cookie(TokenBasedSecurityFilter.COOKIE_NAME, documentDocxToken)
+                .get(JsonObject.class);
+        Assert.assertTrue(json.getJsonArray("documents").size() == 1);
+        
+        // Get the file thumbnail data
+        Response response = target().path("/file/" + file1Id + "/data")
+                .queryParam("size", "thumb")
+                .request()
+                .cookie(TokenBasedSecurityFilter.COOKIE_NAME, documentDocxToken)
+                .get();
+        InputStream is = (InputStream) response.getEntity();
+        byte[] fileBytes = ByteStreams.toByteArray(is);
+        Assert.assertTrue(fileBytes.length > 0); // Images rendered from PDF differ in size from OS to OS due to font issues
+        Assert.assertEquals(MimeType.IMAGE_JPEG, MimeTypeUtil.guessMimeType(fileBytes));
+    }
+    
     /**
      * Test PDF extraction.
      * 
@@ -274,14 +392,14 @@ public class TestDocumentResource extends BaseJerseyTest {
      */
     @Test
     public void testPdfExtraction() throws Exception {
-        // Login document2
+        // Login document_pdf
-        clientUtil.createUser("document2");
+        clientUtil.createUser("document_pdf");
-        String document2Token = clientUtil.login("document2");
+        String documentPdfToken = clientUtil.login("document_pdf");
 
         // Create a document
         long create1Date = new Date().getTime();
         JsonObject json = target().path("/document").request()
-                .cookie(TokenBasedSecurityFilter.COOKIE_NAME, document2Token)
+                .cookie(TokenBasedSecurityFilter.COOKIE_NAME, documentPdfToken)
                 .put(Entity.form(new Form()
                         .param("title", "My super title document 1")
                         .param("description", "My super description for document 1")
@@ -298,7 +416,7 @@ public class TestDocumentResource extends BaseJerseyTest {
                 json = target()
                         .register(MultiPartFeature.class)
                         .path("/file").request()
-                        .cookie(TokenBasedSecurityFilter.COOKIE_NAME, document2Token)
+                        .cookie(TokenBasedSecurityFilter.COOKIE_NAME, documentPdfToken)
                         .put(Entity.entity(multiPart.field("id", document1Id).bodyPart(streamDataBodyPart),
                                 MediaType.MULTIPART_FORM_DATA_TYPE), JsonObject.class);
                 file1Id = json.getString("id");
@@ -310,7 +428,7 @@ public class TestDocumentResource extends BaseJerseyTest {
         json = target().path("/document/list")
                 .queryParam("search", "full:vrandecic")
                 .request()
-                .cookie(TokenBasedSecurityFilter.COOKIE_NAME, document2Token)
+                .cookie(TokenBasedSecurityFilter.COOKIE_NAME, documentPdfToken)
                 .get(JsonObject.class);
         Assert.assertTrue(json.getJsonArray("documents").size() == 1);
         
@@ -318,7 +436,7 @@ public class TestDocumentResource extends BaseJerseyTest {
         Response response = target().path("/file/" + file1Id + "/data")
                 .queryParam("size", "thumb")
                 .request()
-                .cookie(TokenBasedSecurityFilter.COOKIE_NAME, document2Token)
+                .cookie(TokenBasedSecurityFilter.COOKIE_NAME, documentPdfToken)
                 .get();
         InputStream is = (InputStream) response.getEntity();
         byte[] fileBytes = ByteStreams.toByteArray(is);