Issue 581, Hanzi mode support from Shiyuan Guo

git-svn-id: https://zxing.googlecode.com/svn/trunk@1623 59b500cc-1b3d-0410-9834-0bbf25fbcc57

AUTHORS

@@ -51,6 +51,7 @@ Roman Nurik (Google)
 Ryan Alford
 Sanford Squires
 Sean Owen (Google)
+Shiyuan Guo / 郭世元
 Simon Flannery (Ericsson)
 Steven Parkes
 Suraj Supekar

core/src/com/google/zxing/common/StringUtils.java

@@ -30,6 +30,7 @@ public final class StringUtils {
   private static final String PLATFORM_DEFAULT_ENCODING =
       System.getProperty("file.encoding");
   public static final String SHIFT_JIS = "SJIS";
+  public static final String GB2312 = "GB2312";
   private static final String EUC_JP = "EUC_JP";
   private static final String UTF8 = "UTF8";
   private static final String ISO88591 = "ISO8859_1";

core/src/com/google/zxing/qrcode/decoder/DecodedBitStreamParser.java

@@ -45,6 +45,7 @@ final class DecodedBitStreamParser {
       'O', 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z',
       ' ', '$', '%', '*', '+', '-', '.', '/', ':'
   };
+  private static final int GB2312_SUBSET = 1;
 
   private DecodedBitStreamParser() {
   }
@@ -85,18 +86,29 @@ final class DecodedBitStreamParser {
             throw FormatException.getFormatInstance();
           }
         } else {
-          // How many characters will follow, encoded in this mode?
-          int count = bits.readBits(mode.getCharacterCountBits(version));
-          if (mode.equals(Mode.NUMERIC)) {
-            decodeNumericSegment(bits, result, count);
-          } else if (mode.equals(Mode.ALPHANUMERIC)) {
-            decodeAlphanumericSegment(bits, result, count, fc1InEffect);
-          } else if (mode.equals(Mode.BYTE)) {
-            decodeByteSegment(bits, result, count, currentCharacterSetECI, byteSegments, hints);
-          } else if (mode.equals(Mode.KANJI)) {
-            decodeKanjiSegment(bits, result, count);
+          // First handle Hanzi mode which does not start with character count
+          if (mode.equals(Mode.HANZI)) {
+        		//chinese mode contains a sub set indicator right after mode indicator
+        		int subset = bits.readBits(4);
+        		int countHanzi = bits.readBits(mode.getCharacterCountBits(version));
+        		if (subset == GB2312_SUBSET) {
+        			decodeHanziSegment(bits, result, countHanzi);
+            }
           } else {
-            throw FormatException.getFormatInstance();
+            // "Normal" QR code modes:
+            // How many characters will follow, encoded in this mode?
+            int count = bits.readBits(mode.getCharacterCountBits(version));
+            if (mode.equals(Mode.NUMERIC)) {
+              decodeNumericSegment(bits, result, count);
+            } else if (mode.equals(Mode.ALPHANUMERIC)) {
+              decodeAlphanumericSegment(bits, result, count, fc1InEffect);
+            } else if (mode.equals(Mode.BYTE)) {
+              decodeByteSegment(bits, result, count, currentCharacterSetECI, byteSegments, hints);
+            } else if (mode.equals(Mode.KANJI)) {
+              decodeKanjiSegment(bits, result, count);
+            } else {
+              throw FormatException.getFormatInstance();
+            }
           }
         }
       }
@@ -105,6 +117,40 @@ final class DecodedBitStreamParser {
     return new DecoderResult(bytes, result.toString(), byteSegments.isEmpty() ? null : byteSegments, ecLevel);
   }
 
+  /**
+   * See specification GBT 18284-2000
+   */
+  private static void decodeHanziSegment(BitSource bits,
+                                         StringBuffer result,
+                                         int count) throws FormatException {
+    // Each character will require 2 bytes. Read the characters as 2-byte pairs
+    // and decode as GB2312 afterwards
+    byte[] buffer = new byte[2 * count];
+    int offset = 0;
+    while (count > 0) {
+      // Each 13 bits encodes a 2-byte character
+      int twoBytes = bits.readBits(13);
+      int assembledTwoBytes = ((twoBytes / 0x060) << 8) | (twoBytes % 0x060);
+      if (assembledTwoBytes < 0x003BF) {
+        // In the 0xA1A1 to 0xAAFE range
+        assembledTwoBytes += 0x0A1A1;
+      } else {
+        // In the 0xB0A1 to 0xFAFE range
+        assembledTwoBytes += 0x0A6A1;
+      }
+      buffer[offset] = (byte) ((assembledTwoBytes >> 8) & 0xFF);
+      buffer[offset + 1] = (byte) ((assembledTwoBytes) & 0xFF);
+      offset += 2;
+      count--;
+    }
+    
+    try {
+      result.append(new String(buffer, StringUtils.GB2312));
+    } catch (UnsupportedEncodingException uee) {
+      throw FormatException.getFormatInstance();
+    }
+  }
+
   private static void decodeKanjiSegment(BitSource bits,
                                          StringBuffer result,
                                          int count) throws FormatException {

core/src/com/google/zxing/qrcode/decoder/Mode.java

@@ -35,6 +35,8 @@ public final class Mode {
   public static final Mode KANJI = new Mode(new int[]{8, 10, 12}, 0x08, "KANJI");
   public static final Mode FNC1_FIRST_POSITION = new Mode(null, 0x05, "FNC1_FIRST_POSITION");
   public static final Mode FNC1_SECOND_POSITION = new Mode(null, 0x09, "FNC1_SECOND_POSITION");
+  /** See GBT 18284-2000; "Hanzi" is a transliteration of this mode name. */
+  public static final Mode HANZI = new Mode(new int[]{8, 10, 12}, 0x0D, "HANZI");
 
   private final int[] characterCountBitsForVersions;
   private final int bits;
@@ -71,6 +73,9 @@ public final class Mode {
         return KANJI;
       case 0x9:
         return FNC1_SECOND_POSITION;
+      case 0xD:
+    	  // 0xD is defined in GBT 18284-2000, may not be supported in foreign country
+        return HANZI;
       default:
         throw new IllegalArgumentException();
     }

core/test/data/blackbox/qrcode-2/33.txt

@@ -0,0 +1 @@
+
AD:SUB:阿;;

core/test/src/com/google/zxing/qrcode/QRCodeBlackBox2TestCase.java

@@ -27,10 +27,10 @@ public final class QRCodeBlackBox2TestCase extends AbstractBlackBoxTestCase {
 
   public QRCodeBlackBox2TestCase() {
     super("test/data/blackbox/qrcode-2", new MultiFormatReader(), BarcodeFormat.QR_CODE);
-    addTest(26, 26, 0.0f);
-    addTest(26, 26, 90.0f);
-    addTest(26, 26, 180.0f);
-    addTest(25, 25, 270.0f);
+    addTest(27, 27, 0.0f);
+    addTest(27, 27, 90.0f);
+    addTest(27, 27, 180.0f);
+    addTest(26, 26, 270.0f);
   }
 
 }

core/test/src/com/google/zxing/qrcode/decoder/DecodedBitStreamParserTestCase.java

@@ -69,6 +69,18 @@ public final class DecodedBitStreamParserTestCase extends Assert {
     assertEquals("\u00ed\u00f3\u00fa", result);
   }
 
+  @Test
+  public void testHanzi() throws Exception {
+    BitSourceBuilder builder = new BitSourceBuilder();
+    builder.write(0x0D, 4); // Hanzi mode
+    builder.write(0x01, 4); // Subset 1 = GB2312 encoding
+    builder.write(0x01, 8); // 1 characters
+    builder.write(0x03C1, 13);
+    String result = DecodedBitStreamParser.decode(builder.toByteArray(),
+        Version.getVersionForNumber(1), null, null).getText();
+    assertEquals("\u963f", result);
+  }
+
   // TODO definitely need more tests here
 
 }