v1.2.6

dependency-reduced-pom.xml

@@ -54,6 +54,32 @@
           </archive>
         </configuration>
       </plugin>
+      <plugin>
+        <artifactId>maven-assembly-plugin</artifactId>
+        <version>2.4</version>
+        <executions>
+          <execution>
+            <id>make-assembly</id>
+            <phase>package</phase>
+            <goals>
+              <goal>single</goal>
+            </goals>
+          </execution>
+        </executions>
+        <configuration>
+          <finalName>ocrevaluationfull</finalName>
+          <descriptorRefs>
+            <descriptorRef>jar-with-dependencies</descriptorRef>
+          </descriptorRefs>
+          <appendAssemblyId>false</appendAssemblyId>
+          <archive>
+            <manifest>
+              <addClasspath>true</addClasspath>
+              <mainClass>eu.digitisation.MainGUI</mainClass>
+            </manifest>
+          </archive>
+        </configuration>
+      </plugin>
       <plugin>
         <artifactId>maven-enforcer-plugin</artifactId>
         <version>1.3.1</version>

pom.xml

@@ -4,7 +4,7 @@
     <groupId>eu.digitisation</groupId>
     <artifactId>ocrevalUAtion</artifactId>
     <name>ocrevalUAtion</name>
-    <version>1.2.5</version>
+    <version>1.2.6</version>
     <packaging>jar</packaging>
     <description>OCR Evaluation Tool</description>
     <organization>
@@ -29,7 +29,7 @@
         <jdk.version>1.7</jdk.version>
     </properties>
     <build> 
-        <finalName>ocrevaluation</finalName>
+        <!--finalName>ocrevaluation</finalName-->
         <plugins>
             <plugin>
                 <groupId>org.apache.maven.plugins</groupId>
@@ -63,10 +63,11 @@
                     </archive>
                 </configuration>
             </plugin>
-            <!--plugin>
+            <plugin>
                 <artifactId>maven-assembly-plugin</artifactId>
                 <version>2.4</version>
                 <configuration>
+                    <finalName>ocrevaluation</finalName>
                     <descriptorRefs>
                         <descriptorRef>jar-with-dependencies</descriptorRef>
                     </descriptorRefs>
@@ -87,8 +88,8 @@
                         </goals>
                     </execution>
                 </executions>
-            </plugin-->
-            <plugin>
+            </plugin>
+            <!--plugin>
                 <inherited>true</inherited>
                 <groupId>org.apache.maven.plugins</groupId>
                 <artifactId>maven-enforcer-plugin</artifactId>
@@ -109,8 +110,8 @@
                         </configuration>
                     </execution>
                 </executions>
-            </plugin>
-            <plugin>
+            </plugin-->
+            <!--plugin>
                 <groupId>org.apache.maven.plugins</groupId>
                 <artifactId>maven-shade-plugin</artifactId>
                 <version>2.1</version>
@@ -147,7 +148,7 @@
                         </configuration>
                     </execution>
                 </executions>
-            </plugin>
+            </plugin-->
         </plugins>
     </build>
     <dependencies>

src/main/java/eu/digitisation/Main.java

@@ -15,7 +15,7 @@ public class Main {
     static final String helpMsg = "Usage:\t"
             + "ocrevalUAtion -gt file1 [encoding] "
             + "-ocr file2 [encoding] "
-            + "-d output_dir [-r replacements_file]";
+            + "-d output_dir [-r equivalences_file]";
 
     private static void exit_gracefully() {
         System.err.println(helpMsg);

src/main/java/eu/digitisation/MainGUI.java

@@ -74,16 +74,21 @@ public MainGUI() {
         repaint();
     }
 
+    /**
+     * 
+     * @return true if all required files have been selected
+     */
     private boolean checkInputFiles() {
         boolean ready = true;
         Component[] components = pane.getComponents();
-
-        for (int n = 0; n < 2; ++n) {
+        boolean[] required = {true, true, false};
+
+        for (int n = 0; n < 3; ++n) {
             InputFileSelector ifs = (InputFileSelector) components[n];
             if (ifs.accepted()) {
                 files[n] = ifs.getFile();
             }
-            if (!(ifs.accepted() && files[n].exists())) {
+            if (required[n] && !(ifs.accepted() && files[n].exists())) {
                 ifs.shade(Color.decode("#fffacd"));
                 ifs.repaint();
                 ready = false;
@@ -92,6 +97,18 @@ private boolean checkInputFiles() {
         return ready;
     }
 
+    /**
+     * Show a warning  message
+     * @param text the text to be displayed
+     */
+    private void warning(String text) {
+        InputFileSelector ifs = (InputFileSelector)pane.getComponent(0);
+        ifs.setForeground(Color.RED);
+        ifs.shade(Color.decode("#fffacd"));
+        ifs.setText(text);
+        ifs.repaint();
+    }
+
     @Override
     public void actionPerformed(ActionEvent e) {
         JButton pressed = (JButton) e.getSource();
@@ -121,7 +138,11 @@ public void actionPerformed(ActionEvent e) {
                                 System.out.println(uri);
                                 Desktop.getDesktop().browse(uri);                          
                         }
-                    } catch (Exception ex) {
+                    } catch (InvalidObjectException ex) {
+                        warning(ex.getMessage());
+                    } catch (URISyntaxException ex) {
+                        Logger.getLogger(MainGUI.class.getName()).log(Level.SEVERE, null, ex);
+                    } catch (IOException ex) {
                         Logger.getLogger(MainGUI.class.getName()).log(Level.SEVERE, null, ex);
                     }
                 }

src/main/java/eu/digitisation/distance/Aligner.java

@@ -31,7 +31,7 @@ public class Aligner {
     // style for unaligned segments
     final static String uStyle = "background-color:aquamarine";
     // style for  highlight replacement in parallel text
-    final static String twinStyle = ""; 
+    final static String twinStyle = "";
 
     /**
      * @return 3-wise minimum.
@@ -109,11 +109,14 @@ private static EditTable align(String first, String second) {
      * Shows text alignment based on a pseudo-Levenshtein distance where
      * white-spaces are not allowed to be replaced with text or vice-versa
      *
-     * @param first
-     * @param second
+     * @param header1 first text title for table head
+     * @param header2 second text title for table head
+     * @param first the first text
+     * @param second the second text
      * @return a table in XHTML format showing the alignments
      */
-    public static Element alignmentMap(String first, String second) {
+    public static Element alignmentMap(String header1, String header2,
+            String first, String second) {
         EditTable B = align(first, second);
         DocumentBuilder builder = new DocumentBuilder("table");
         Element table = builder.root();
@@ -134,10 +137,12 @@ public static Element alignmentMap(String first, String second) {
         row = builder.addElement("tr");
         cell1 = builder.addElement(row, "td");
         cell2 = builder.addElement(row, "td");
+        cell1.setAttribute("width", "50%");
+        cell2.setAttribute("width", "50%");
         cell1.setAttribute("align", "center");
         cell2.setAttribute("align", "center");
-        builder.addTextElement(cell1, "h3", "Reference");
-        builder.addTextElement(cell2, "h3", "OCR");
+        builder.addTextElement(cell1, "h3", header1);
+        builder.addTextElement(cell2, "h3", header2);
         row = builder.addElement("tr");
         cell1 = builder.addElement(row, "td");
         cell2 = builder.addElement(row, "td");

src/main/java/eu/digitisation/gui/InputFileSelector.java

@@ -87,6 +87,14 @@ public void shade(Color color) {
         area.setForeground(Color.DARK_GRAY);
     }
 
+    /**
+     * Change descriptive text
+     * @param text the text to be displayed
+     */
+    public void setText(String text) {
+        area.setText(text);
+    }
+
     /**
      *
      * @return the selected input file

src/main/java/eu/digitisation/io/Batch.java

@@ -38,25 +38,29 @@ public class Batch {
     public Batch(File dir1, File dir2) throws InvalidObjectException {
         if (dir1.isDirectory()) {
             files1 = dir1.listFiles();
+            java.util.Arrays.sort(files1);
         } else {
             String[] a = {};
             files1 = new File[1];
             files1[0] = dir1;
         }
         if (dir2.isDirectory()) {
             files2 = dir2.listFiles();
+            java.util.Arrays.sort(files2);
         } else {
             files2 = new File[1];
             files2[0] = dir2;
         }
         if (files1.length != files2.length) {
-            throw new java.io.InvalidObjectException(dir1 + " and " + dir2
+            throw new java.io.InvalidObjectException(dir1.getName() 
+                    + " and " + dir2.getName()
                     + " contain a different number of files");
         } else {
             size = files1.length;
         }
         if (!consistent()) {
-            throw new java.io.InvalidObjectException(dir1 + " and " + dir2
+            throw new java.io.InvalidObjectException(dir1.getName() 
+                    + " and " + dir2.getName()
                     + " contain files with inconsistent names");
         }
     }

src/main/java/eu/digitisation/io/CharFilter.java

@@ -46,7 +46,7 @@ public CharFilter(File file) {
             BufferedReader reader = new BufferedReader(new FileReader(file));
             while (reader.ready()) {
                 String line = reader.readLine();
-                String[] tokens = line.split("([,;])");
+                String[] tokens = line.split("([,;\t])");
                 if (tokens.length > 1) {  // allow comments in line
                     String key = UnicodeReader.codepointsToString(tokens[0]);
                     String value = UnicodeReader.codepointsToString(tokens[1]);

src/main/java/eu/digitisation/io/TextContent.java

@@ -221,6 +221,7 @@ protected void readTextFile(File file) {
             encoding = Encoding.detect(file);
         }
         System.err.println("Text file " + file + " encoding is " + encoding);
+
         // read content
         try {
             FileInputStream fis = new FileInputStream(file);

src/main/java/eu/digitisation/io/UnicodeReader.java

@@ -43,7 +43,7 @@ protected static String codepointsToString(String codes) throws IOException {
         for (String token : tokens) {
             if (token.length() % 4 != 0) {
                 throw new IOException(token
-                        + " is not a alid Unicode hex sequence");
+                        + " is not a valid Unicode hex sequence");
             }
             for (int pos = 0; pos + 3 < token.length(); pos += 4) {
                 String sub = token.substring(pos, pos + 4);

src/main/java/eu/digitisation/ocrevaluation/CharStatTable.java

@@ -167,6 +167,7 @@ public double cer() {
                     + value(c, EdOp.SUBSTITUTE)
                     + value(c, EdOp.DELETE);
         }
-        return (spu + sub + add) / (double) tot * 100;
+
+        return (spu + sub + add) / (double) tot;
     }
 }

src/main/java/eu/digitisation/ocrevaluation/Report.java

@@ -98,7 +98,8 @@ public Report(File gtfile, String gtencoding,
         double cerDL = ErrorMeasure.cerDL(gts, ocrs);
         double wer = ErrorMeasure.wer(gts, ocrs);
         double ber = ErrorMeasure.ber(gts, ocrs);
-        Element alitab = Aligner.alignmentMap(gts, ocrs);
+        Element alitab = Aligner.alignmentMap(gtfile.getName(), 
+                ocrfile.getName(), gts, ocrs);
         CharStatTable stats = new CharStatTable(gts, ocrs);
 
         // General info
@@ -121,7 +122,7 @@ public Report(Batch batch, String gtencoding, String ocrencoding,
             File eqfile) {
         super("html");
         init();
-
+       
         CharFilter filter = (eqfile == null) ? null : new CharFilter(eqfile);
         CharStatTable stats = new CharStatTable();
         Element summaryTab;
@@ -144,8 +145,10 @@ public Report(Batch batch, String gtencoding, String ocrencoding,
             TokenArray ocrarray = factory.newTokenArray(ocrs);
             BagOfWords gtbag = new BagOfWords(gts);
             BagOfWords ocrbag = new BagOfWords(ocrs);
-            Element alitab = Aligner.alignmentMap(gts, ocrs);
+            Element alitab = Aligner.alignmentMap(input.first.getName(),
+                    input.second.getName(), gts, ocrs);
             stats.add(gts, ocrs);
+            addTextElement(body, "div", " ");
             addElement(body, alitab);
             numwords += gtarray.length();
             wdist += ArrayEditDistance.distance(gtarray.tokens(), ocrarray.tokens(),