Upgrade to Tesseract 4.00alpha (a011b15)

pom.xml

@@ -3,7 +3,7 @@
     <modelVersion>4.0.0</modelVersion>
     <groupId>net.sourceforge.tess4j</groupId>
     <artifactId>tess4j</artifactId>
-    <version>3.3.1-SNAPSHOT</version>
+    <version>4.0.0-SNAPSHOT</version>
     <packaging>jar</packaging>
 
     <name>Tess4J - Tesseract for Java</name>

src/main/java/net/sourceforge/tess4j/ITessAPI.java

@@ -46,13 +46,13 @@ public static interface TessOcrEngineMode {
          */
         public static final int OEM_TESSERACT_ONLY = 0;
         /**
-         * Run Cube only - better accuracy, but slower
+         * Run just the LSTM line recognizer
          */
-        public static final int OEM_CUBE_ONLY = 1;
+        public static final int OEM_LSTM_ONLY = 1;
         /**
-         * Run both and combine results - best accuracy
+         * Run the LSTM recognizer, but allow fallback to Tesseract when things get difficult
          */
-        public static final int OEM_TESSERACT_CUBE_COMBINED = 2;
+        public static final int OEM_TESSERACT_LSTM_COMBINED = 2;
         /**
          * Specify this mode when calling <code>init_*()</code>, to indicate
          * that any of the above modes should be automatically inferred from the
@@ -61,6 +61,14 @@ public static interface TessOcrEngineMode {
          * <code>OEM_TESSERACT_ONLY</code>.
          */
         public static final int OEM_DEFAULT = 3;
+//        /**
+//         * Run Cube only - better accuracy, but slower
+//         */
+//        public static final int OEM_CUBE_ONLY = 4;
+//        /**
+//         * Run both and combine results - best accuracy
+//         */
+//        public static final int OEM_TESSERACT_CUBE_COMBINED = 5;
     };
 
     /**

src/main/java/net/sourceforge/tess4j/TessAPI.java

@@ -74,7 +74,7 @@ public interface TessAPI extends Library, ITessAPI {
 
     TessResultRenderer TessHOcrRendererCreate(String outputbase);
 
-    TessResultRenderer TessPDFRendererCreate(String outputbase, String datadir);
+    TessResultRenderer TessPDFRendererCreate(String outputbase, String datadir, int textonly);
 
     TessResultRenderer TessUnlvRendererCreate(String outputbase);
 

src/main/java/net/sourceforge/tess4j/TessAPI1.java

@@ -74,7 +74,7 @@ public class TessAPI1 implements Library, ITessAPI {
 
     public static native TessResultRenderer TessHOcrRendererCreate(String outputbase);
 
-    public static native TessResultRenderer TessPDFRendererCreate(String outputbase, String datadir);
+    public static native TessResultRenderer TessPDFRendererCreate(String outputbase, String datadir, int textonly);
 
     public static native TessResultRenderer TessUnlvRendererCreate(String outputbase);
 

src/main/java/net/sourceforge/tess4j/Tesseract.java

@@ -468,9 +468,9 @@ private TessResultRenderer createRenderers(String outputbase, List<RenderedForma
                 case PDF:
                     String dataPath = api.TessBaseAPIGetDatapath(handle);
                     if (renderer == null) {
-                        renderer = api.TessPDFRendererCreate(outputbase, dataPath);
+                        renderer = api.TessPDFRendererCreate(outputbase, dataPath, TRUE);
                     } else {
-                        api.TessResultRendererInsert(renderer, api.TessPDFRendererCreate(outputbase, dataPath));
+                        api.TessResultRendererInsert(renderer, api.TessPDFRendererCreate(outputbase, dataPath, TRUE));
                     }
                     break;
                 case BOX:

src/main/java/net/sourceforge/tess4j/Tesseract1.java

@@ -434,9 +434,9 @@ private TessResultRenderer createRenderers(String outputbase, List<RenderedForma
                 case PDF:
                     String dataPath = TessBaseAPIGetDatapath(handle);
                     if (renderer == null) {
-                        renderer = TessPDFRendererCreate(outputbase, dataPath);
+                        renderer = TessPDFRendererCreate(outputbase, dataPath, TRUE);
                     } else {
-                        TessResultRendererInsert(renderer, TessPDFRendererCreate(outputbase, dataPath));
+                        TessResultRendererInsert(renderer, TessPDFRendererCreate(outputbase, dataPath, TRUE));
                     }
                     break;
                 case BOX:

src/main/java/net/sourceforge/tess4j/util/LoadLibs.java

@@ -47,7 +47,7 @@ public class LoadLibs {
     /**
      * Native library name.
      */
-    public static final String LIB_NAME = "libtesseract305";
+    public static final String LIB_NAME = "libtesseract400";
     public static final String LIB_NAME_NON_WIN = "tesseract";
 
     private static final org.slf4j.Logger logger = LoggerFactory.getLogger(new LoggHelper().toString());

src/test/java/net/sourceforge/tess4j/TessAPI1Test.java

@@ -209,7 +209,7 @@ public void testTessBaseAPIGetComponentImages() throws Exception {
     @Test
     public void testTessVersion() {
         logger.info("TessVersion");
-        String expResult = "3.05";
+        String expResult = "4.00";
         String result = TessAPI1.TessVersion();
         logger.info(result);
         assertTrue(result.startsWith(expResult));
@@ -636,7 +636,7 @@ public void testResultRenderer() throws Exception {
         TessAPI1.TessResultRendererInsert(renderer, TessAPI1.TessBoxTextRendererCreate(outputbase));
         TessAPI1.TessResultRendererInsert(renderer, TessAPI1.TessTextRendererCreate(outputbase));
         String dataPath = TessAPI1.TessBaseAPIGetDatapath(handle);
-        TessAPI1.TessResultRendererInsert(renderer, TessAPI1.TessPDFRendererCreate(outputbase, dataPath));
+        TessAPI1.TessResultRendererInsert(renderer, TessAPI1.TessPDFRendererCreate(outputbase, dataPath, TRUE));
         int result = TessAPI1.TessBaseAPIProcessPages(handle, image, null, 0, renderer);
 
 //        if (result == FALSE) {

src/test/java/net/sourceforge/tess4j/TessAPIImpl.java

@@ -69,7 +69,7 @@ public ITessAPI.TessResultRenderer TessHOcrRendererCreate(String outputbase) {
     }
 
     @Override
-    public ITessAPI.TessResultRenderer TessPDFRendererCreate(String outputbase, String datadir) {
+    public ITessAPI.TessResultRenderer TessPDFRendererCreate(String outputbase, String datadir, int textonly) {
         throw new UnsupportedOperationException("Not supported yet.");
     }
 

src/test/java/net/sourceforge/tess4j/TessAPITest.java

@@ -212,7 +212,7 @@ public void testTessBaseAPIGetComponentImages() throws Exception {
     @Test
     public void testTessVersion() {
         logger.info("TessVersion");
-        String expResult = "3.05";
+        String expResult = "4.00";
         String result = api.TessVersion();
         logger.info(result);
         assertTrue(result.startsWith(expResult));
@@ -642,7 +642,7 @@ public void testResultRenderer() throws Exception {
         api.TessResultRendererInsert(renderer, api.TessBoxTextRendererCreate(outputbase));
         api.TessResultRendererInsert(renderer, api.TessTextRendererCreate(outputbase));
         String dataPath = api.TessBaseAPIGetDatapath(handle);
-        api.TessResultRendererInsert(renderer, api.TessPDFRendererCreate(outputbase, dataPath));
+        api.TessResultRendererInsert(renderer, api.TessPDFRendererCreate(outputbase, dataPath, TRUE));
         int result = api.TessBaseAPIProcessPages(handle, image, null, 0, renderer);
 
         if (result == FALSE) {