Popup warning for training with LSTM or WordStr box files

Author: nguyenq

src/net/sourceforge/tessboxeditor/Gui.java

@@ -34,8 +34,6 @@
 import java.awt.image.BufferedImage;
 import java.io.*;
 import java.nio.charset.StandardCharsets;
-import java.nio.file.Files;
-import java.nio.file.Paths;
 import java.text.*;
 import java.util.*;
 import java.util.logging.Level;
@@ -1583,7 +1581,7 @@ void loadBoxes(File boxFile) {
             try {
                 boxPages.clear();
 
-                String str = readBoxFile(boxFile);
+                String str = Utils.readTextFile(boxFile);
                 // load into textarea
                 this.jTextAreaBoxData.setText(str);
                 boxPages = parseBoxString(str, imageList);
@@ -1607,10 +1605,6 @@ void loadBoxes(File boxFile) {
         }
     }
 
-    String readBoxFile(File boxFile) throws IOException {
-        return new String(Files.readAllBytes(Paths.get(boxFile.getPath())), StandardCharsets.UTF_8);
-    }
-
     List<TessBoxCollection> parseBoxString(String boxStr, List<BufferedImage> imageList) throws IOException {
         List<TessBoxCollection> allBoxPages = new ArrayList<TessBoxCollection>();
         isWordStrFormat = false;

src/net/sourceforge/tessboxeditor/GuiWithEdit.java

@@ -43,6 +43,7 @@
 import static net.sourceforge.tessboxeditor.Gui.prefs;
 import net.sourceforge.tessboxeditor.datamodel.TessBox;
 import net.sourceforge.tessboxeditor.datamodel.TessBoxCollection;
+import net.sourceforge.vietocr.util.Utils;
 
 public class GuiWithEdit extends GuiWithMRU implements PropertyChangeListener {
 
@@ -361,7 +362,7 @@ protected Void doInBackground() throws Exception {
                 }
 
                 List<BufferedImage> imageList = ImageIOHelper.getImageList(imageFile);
-                String str = readBoxFile(boxFile);
+                String str = Utils.readTextFile(boxFile);
                 List<TessBoxCollection> boxPages = parseBoxString(str, imageList);
                 performSegment(imageList, boxPages, instance);
 

src/net/sourceforge/tessboxeditor/TessTrainer.java

@@ -426,11 +426,16 @@ public boolean accept(File dir, String filename) {
      *
      * @return
      */
-    String[] getImageFilesWithBox() {
+    String[] getImageFilesWithBox() throws IOException {
         List<String> filesWithBox = new ArrayList<String>();
         for (String file : getImageFiles()) {
             String withoutExt = TextUtilities.stripExtension(file);
-            if (new File(inputDataDir, withoutExt + ".box").exists()) {
+            File boxFile = new File(inputDataDir, withoutExt + ".box");
+            if (boxFile.exists()) {
+                String str = Utils.readTextFile(boxFile);
+                if (str.contains("WordStr") || str.contains("\t")) {
+                    throw new RuntimeException("Cannot train with LSTM or WordStr box files.\nTraining for Tesseract 4.0x is not supported.");
+                }
                 filesWithBox.add(file);
             }
         }

src/net/sourceforge/tessboxeditor/config.xml

@@ -2,6 +2,6 @@
 <!DOCTYPE properties SYSTEM "http://java.sun.com/dtd/properties.dtd">
 <properties>
     <comment>Configuration File</comment>
-    <entry key="ReleaseDate">2020/3/6</entry>
+    <entry key="ReleaseDate">2020/3/21</entry>
     <entry key="Version">v2.3.0</entry>
 </properties>

src/net/sourceforge/vietocr/util/Utils.java

@@ -137,9 +137,9 @@ public static String join(Collection<?> s, String delimiter) {
      *
      * @param textFile
      * @return
-     * @throws Exception
+     * @throws IOException
      */
-    public static String readTextFile(File textFile) throws Exception {
+    public static String readTextFile(File textFile) throws IOException {
         return new String(Files.readAllBytes(textFile.toPath()), StandardCharsets.UTF_8); // Java 7 API
     }
 

versionchanges.txt

@@ -135,7 +135,7 @@ Version 2.2.1 (25 August 2019)
 Version 2.2.2 (8 October 2019)
 - Update dependencies
 
-Version 2.3.0 (6 March 2020)
+Version 2.3.0 (21 March 2020)
 - Support LSTM & WordStr box format
 - Support reordering boxes through table row drag-and-drop
 - Fix column alignment