From 9d15b5f5719a2d75d744b4ea2d828f17957f6711 Mon Sep 17 00:00:00 2001 From: "U-BASIS\\dsmyda" Date: Mon, 10 Dec 2018 16:37:55 -0500 Subject: [PATCH] Fixed language pack look up to happen once at class loading --- .../sleuthkit/autopsy/textextractors/TikaTextExtractor.java | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/Core/src/org/sleuthkit/autopsy/textextractors/TikaTextExtractor.java b/Core/src/org/sleuthkit/autopsy/textextractors/TikaTextExtractor.java index 8ba64e4f0d..9b766a9e9e 100644 --- a/Core/src/org/sleuthkit/autopsy/textextractors/TikaTextExtractor.java +++ b/Core/src/org/sleuthkit/autopsy/textextractors/TikaTextExtractor.java @@ -120,6 +120,7 @@ final class TikaTextExtractor extends TextExtractor { private static final String TESSERACT_DIR_NAME = "Tesseract-OCR"; //NON-NLS 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 List TIKA_SUPPORTED_TYPES = new Tika().getParser().getSupportedTypes(new ParseContext()) @@ -177,8 +178,7 @@ final class TikaTextExtractor extends TextExtractor { 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 - System.out.println(getLanguagePacks()); - ocrConfig.setLanguage(getLanguagePacks()); + ocrConfig.setLanguage(LANGUAGE_PACKS); parseContext.set(TesseractOCRConfig.class, ocrConfig); } @@ -279,7 +279,7 @@ final class TikaTextExtractor extends TextExtractor { * * @return String of all language packs available for Tesseract to use */ - private String getLanguagePacks() { + 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