Upgraded to Tesseract 4.0

Core/src/org/sleuthkit/autopsy/textreaders/TikaTextExtractor.java

@@ -29,8 +29,8 @@ import java.io.InputStream;
 import java.io.PushbackReader;
 import java.io.Reader;
 import java.nio.file.Paths;
+import java.util.ArrayList;
 import java.util.Arrays;
-import java.util.HashSet;
 import java.util.List;
 import java.util.Objects;
 import java.util.concurrent.Callable;
@@ -60,6 +60,7 @@ import org.sleuthkit.autopsy.casemodule.Case;
 import org.sleuthkit.autopsy.casemodule.NoCurrentCaseException;
 import org.sleuthkit.autopsy.coreutils.ExecUtil;
 import org.sleuthkit.autopsy.coreutils.ExecUtil.ProcessTerminator;
+import org.sleuthkit.autopsy.coreutils.FileUtil;
 import org.sleuthkit.autopsy.coreutils.PlatformUtil;
 import org.sleuthkit.autopsy.textreaders.textreaderconfigs.ImageConfig;
 import org.sleuthkit.autopsy.datamodel.ContentUtils;
@@ -122,8 +123,8 @@ final class TikaTextExtractor extends TextExtractor {
 
     private static final java.util.logging.Logger tikaLogger = java.util.logging.Logger.getLogger("Tika"); //NON-NLS
 
-    private final ThreadFactory tikaThreadFactory = 
-            new ThreadFactoryBuilder().setNameFormat("tika-reader-%d").build();
+    private final ThreadFactory tikaThreadFactory
+            = new ThreadFactoryBuilder().setNameFormat("tika-reader-%d").build();
     private final ExecutorService executorService = Executors.newSingleThreadExecutor(tikaThreadFactory);
     private static final String SQLITE_MIMETYPE = "application/x-sqlite3";
 
@@ -135,8 +136,10 @@ final class TikaTextExtractor extends TextExtractor {
     private static final String TESSERACT_EXECUTABLE = "tesseract.exe"; //NON-NLS
     private static final File TESSERACT_PATH = locateTesseractExecutable();
     private static final String LANGUAGE_PACKS = getLanguagePacks();
+    private static final String TESSERACT_LANGUAGE_PACK_EXT = "traineddata"; //NON-NLS
+    private static final String TESSERACT_OUTPUT_FILE_NAME = "tess_output"; //NON-NLS
+    
     private ProcessTerminator processTerminator;
-    private static final String TESSERACT_OUTPUT_FILE_NAME = "output";
 
     private static final List<String> TIKA_SUPPORTED_TYPES
             = new Tika().getParser().getSupportedTypes(new ParseContext())
@@ -182,7 +185,7 @@ final class TikaTextExtractor extends TextExtractor {
             AbstractFile file = ((AbstractFile) content);
             //Run OCR on images with Tesseract directly. 
             if (file.getMIMEType().toLowerCase().startsWith("image/")) {
-                stream = runOcrAndGetOutputStream(file);
+                stream = performOCR(file);
             } else {
                 //Otherwise, go through Tika for PDFs so that it can
                 //extract images and run Tesseract on them.     
@@ -201,11 +204,11 @@ final class TikaTextExtractor extends TextExtractor {
                 String tesseractFolder = TESSERACT_PATH.getParent();
                 ocrConfig.setTesseractPath(tesseractFolder);
                 /*
-                 * Tesseract expects language data packs to be in a
-                 * subdirectory of tesseractFolder, in a folder called
-                 * "tessdata". If they are stored somewhere else, use
-                 * ocrConfig.setTessdataPath(String tessdataPath) to point
-                 * to them
+                 * Tesseract expects language data packs to be in a subdirectory
+                 * of tesseractFolder, in a folder called "tessdata". If they
+                 * are stored somewhere else, use
+                 * ocrConfig.setTessdataPath(String tessdataPath) to point to
+                 * them
                  */
                 ocrConfig.setLanguage(LANGUAGE_PACKS);
                 parseContext.set(TesseractOCRConfig.class, ocrConfig);
@@ -228,8 +231,7 @@ final class TikaTextExtractor extends TextExtractor {
         Future<Reader> future = executorService.submit(
                 new GetTikaReader(parser, stream, metadata, parseContext));
         try {
-            final Reader tikaReader = future.get(getTimeout(content.getSize()),
-                    TimeUnit.SECONDS);
+            final Reader tikaReader = future.get(getTimeout(content.getSize()), TimeUnit.SECONDS);
             //check if the reader is empty
             PushbackReader pushbackReader = new PushbackReader(tikaReader);
             int read = pushbackReader.read();
@@ -238,11 +240,9 @@ final class TikaTextExtractor extends TextExtractor {
                         + "Tika returned empty reader for " + content);
             }
             pushbackReader.unread(read);
-
             //concatenate parsed content and meta data into a single reader.
             CharSource metaDataCharSource = getMetaDataCharSource(metadata);
-            return CharSource.concat(new ReaderCharSource(pushbackReader),
-                    metaDataCharSource).openStream();
+            return CharSource.concat(new ReaderCharSource(pushbackReader), metaDataCharSource).openStream();
         } catch (TimeoutException te) {
             final String msg = NbBundle.getMessage(this.getClass(),
                     "AbstractFileTikaTextExtract.index.tikaParseTimeout.text",
@@ -273,19 +273,19 @@ final class TikaTextExtractor extends TextExtractor {
      * @throws
      * org.sleuthkit.autopsy.textextractors.TextExtractor.ExtractionException
      */
-    private InputStream runOcrAndGetOutputStream(AbstractFile file) throws ExtractionException {
+    private InputStream performOCR(AbstractFile file) throws ExtractionException {
         File inputFile = null;
         File outputFile = null;
         try {
+            String tempDirectory = Case.getCurrentCaseThrows().getTempDirectory();
+            
             //Appending file id makes the name unique
-            String tempFileName = file.getId() + file.getName();
-            inputFile = Paths.get(Case.getCurrentCaseThrows().getTempDirectory(),
-                    tempFileName).toFile();
+            String tempFileName = FileUtil.escapeFileName(file.getId() + file.getName());
+            inputFile = Paths.get(tempDirectory, tempFileName).toFile();
             ContentUtils.writeToFile(content, inputFile);
 
-            String tempOutputName = file.getId() + TESSERACT_OUTPUT_FILE_NAME;
-            String outputFilePath = Paths.get(Case.getCurrentCaseThrows().getTempDirectory(),
-                    tempOutputName).toString();
+            String tempOutputName = FileUtil.escapeFileName(file.getId() + TESSERACT_OUTPUT_FILE_NAME);
+            String outputFilePath = Paths.get(tempDirectory, tempOutputName).toString();
             String executeablePath = TESSERACT_PATH.toString();
 
             //Build tesseract commands
@@ -324,6 +324,7 @@ final class TikaTextExtractor extends TextExtractor {
      * cancelled.
      */
     private class GetTikaReader implements Callable<Reader> {
+
         private final AutoDetectParser parser;
         private final InputStream stream;
         private final Metadata metadata;
@@ -356,6 +357,7 @@ final class TikaTextExtractor extends TextExtractor {
          * Store a reference to file on construction
          *
          * @param file
+         *
          * @throws FileNotFoundException
          */
         public CleanUpStream(File file) throws FileNotFoundException {
@@ -450,25 +452,19 @@ final class TikaTextExtractor extends TextExtractor {
      */
     private static String getLanguagePacks() {
         File languagePackRootDir = new File(TESSERACT_PATH.getParent(), "tessdata");
-        //Acceptable extensions for Tesseract-OCR version 3.05 language packs.
-        //All extensions other than traineddata are associated with cube files that
-        //have been made obsolete since version 4.0.
-        List<String> acceptableExtensions = Arrays.asList("traineddata", "params",
-                "lm", "fold", "bigrams", "nn", "word-freq", "size",
-                "user-patterns", "user-words");
-        //Pull out only unique languagePacks
-        HashSet<String> languagePacks = new HashSet<>();
-        if (languagePackRootDir.exists()) {
+        if (!languagePackRootDir.exists()) {
+            return "";
+        }
+
+        List<String> languagePacks = new ArrayList<>();
         for (File languagePack : languagePackRootDir.listFiles()) {
-                if (languagePack.isDirectory() || !acceptableExtensions.contains(
-                        FilenameUtils.getExtension(languagePack.getName()))) {
-                    continue;
-                }
-                String threeLetterPackageName = languagePack.getName().substring(0, 3);
-                //Ignore the eng language pack if accidentally added
-                languagePacks.add(threeLetterPackageName);
+            String fileExt = FilenameUtils.getExtension(languagePack.getName()); 
+            if (!languagePack.isDirectory() && TESSERACT_LANGUAGE_PACK_EXT.equals(fileExt)) {
+                String packageName = FilenameUtils.getBaseName(languagePack.getName());
+                languagePacks.add(packageName);
             }
         }
+
         return String.join("+", languagePacks);
     }
 
@@ -499,8 +495,8 @@ final class TikaTextExtractor extends TextExtractor {
      * Determines how the extraction process will proceed given the settings
      * stored in this context instance.
      *
-     * See the ImageConfig class in the extractionconfigs package
- for available settings.
+     * See the ImageConfig class in the extractionconfigs package for available
+     * settings.
      *
      * @param context Instance containing config classes
      */

KeywordSearch/src/org/sleuthkit/autopsy/keywordsearch/KeywordSearchIngestModule.java

@@ -31,6 +31,7 @@ import org.openide.util.NbBundle.Messages;
 import org.openide.util.lookup.Lookups;
 import org.sleuthkit.autopsy.casemodule.Case;
 import org.sleuthkit.autopsy.casemodule.NoCurrentCaseException;
+import org.sleuthkit.autopsy.coreutils.ExecUtil.ProcessTerminator;
 import org.sleuthkit.autopsy.coreutils.Logger;
 import org.sleuthkit.autopsy.coreutils.MessageNotifyUtil;
 import org.sleuthkit.autopsy.ingest.FileIngestModule;
@@ -477,10 +478,11 @@ public final class KeywordSearchIngestModule implements FileIngestModule {
         private boolean extractTextAndIndex(AbstractFile aFile, String detectedFormat) throws IngesterException {
             ImageConfig imageConfig = new ImageConfig();
             imageConfig.setOCREnabled(KeywordSearchSettings.getOcrOption());
-            Lookup extractionContext = Lookups.fixed(imageConfig);
+            ProcessTerminator terminator = () -> context.fileIngestIsCancelled();
+            Lookup extractionContext = Lookups.fixed(imageConfig, terminator);
             
             try {
-                Reader specializedReader = TextReaders.getReader(aFile,extractionContext);
+                Reader specializedReader = TextReaders.getReader(aFile, extractionContext);
                 //divide into chunks and index
                 return Ingester.getDefault().indexText(specializedReader,aFile.getId(),aFile.getName(), aFile, context);
             } catch (TextReaders.NoTextReaderFound ex) {

thirdparty/Tesseract-OCR/tessdata/configs/lstm.train

@@ -0,0 +1,13 @@
+disable_character_fragments T
+file_type                   .bl
+textord_fast_pitch_test	T
+tessedit_single_match	0
+tessedit_zero_rejection T
+tessedit_minimal_rejection F
+tessedit_write_rep_codes F
+il1_adaption_test 1
+edges_children_fix F
+edges_childarea 0.65
+edges_boxarea 0.9
+tessedit_train_line_recognizer T
+textord_no_rejects T

thirdparty/Tesseract-OCR/tessdata/eng.cube.fold

@@ -1,12 +0,0 @@
-0oO
-lI1
-cC
-kK
-pP
-sS
-uU
-vV
-wW
-xX
-yY
-zZ

thirdparty/Tesseract-OCR/tessdata/eng.cube.lm

@@ -1,7 +0,0 @@
-LeadPunc="({[`'
-TrailPunc=}:;-]!?`,.)"'
-NumLeadPunc=#({[@$
-NumTrailPunc=}):;].,%
-Operators=*+-/.:,()[]
-Digits=0123456789
-Alphas=abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ

thirdparty/Tesseract-OCR/tessdata/eng.cube.params

@@ -1,14 +0,0 @@
-RecoWgt=1.0
-SizeWgt=0.2435
-OODWgt=0.0214
-NumWgt=0.036
-CharBigramsWgt=0.1567
-MaxSegPerChar=8
-BeamWidth=10
-ConvGridSize=48
-WordUnigramsWgt=0.01
-MaxWordAspectRatio=20.0000
-MinSpaceHeightRatio=0.5000
-MaxSpaceHeightRatio=0.6000
-HistWindWid=2
-MinConCompSize=0

thirdparty/Tesseract-OCR/tessdata/eng.user-patterns

@@ -1,2 +0,0 @@
-1-\d\d\d-GOOG-411
-www.\n\\\*.com

thirdparty/Tesseract-OCR/tessdata/eng.user-words

@@ -1,5 +0,0 @@
-the
-quick
-brown
-fox
-jumped