2021SC@SDUSC
前言:本篇博客继续对QR码的生成代码进行分析。主要对Encoder类的代码进行详细分析,详见代码注释。
在使用字母数字模式创建二维码时,编码过程中需要使用字符的字母数字值。ALPHANUMERIC_TABLE提供了QR码在字母数字模式下支持的字符的参考。具体定义如下:
private static final int[] ALPHANUMERIC_TABLE = {
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, // 0x00-0x0f
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, // 0x10-0x1f
36, -1, -1, -1, 37, 38, -1, -1, -1, -1, 39, 40, -1, 41, 42, 43, // 0x20-0x2f
0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 44, -1, -1, -1, -1, -1, // 0x30-0x3f
-1, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, // 0x40-0x4f
25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, -1, -1, -1, -1, -1, // 0x50-0x5f
};
calculateMaskPenalty用于掩码惩罚的计算。基本上,它适用四条规则,并将所有处罚相加。
private static int calculateMaskPenalty(ByteMatrix matrix) {
return MaskUtil.applyMaskPenaltyRule1(matrix) //规则1
+ MaskUtil.applyMaskPenaltyRule2(matrix) //规则2
+ MaskUtil.applyMaskPenaltyRule3(matrix) //规则3
+ MaskUtil.applyMaskPenaltyRule4(matrix); //规则4
}
encode是Encoder类中最重要的函数,用来编码QR码,返回表示编码的QR码。
其中:content 表示要编码的内容文本;
ecLevel 表示要使用的错误更正级别;
如果编码无法成功,例如由于无效的内容或配置,则会 throws WriterException 。
public static QRCode encode(String content,
ErrorCorrectionLevel ecLevel,
Map<EncodeHintType,?> hints) throws WriterException {
// 确定调用者指定的字符编码(如果有的话)
Charset encoding = DEFAULT_BYTE_MODE_ENCODING;
boolean hasEncodingHint = hints != null && hints.containsKey(EncodeHintType.CHARACTER_SET);
if (hasEncodingHint) {
encoding = Charset.forName(hints.get(EncodeHintType.CHARACTER_SET).toString());
}
// 选择适合内容的编码模式。注意,这不会尝试使用多个模式/段。
Mode mode = chooseMode(content, encoding);
// 存储头信息,如模式和长度,以及“头”段(如ECI段)。
BitArray headerBits = new BitArray();
// 如果适用,附加ECI段
if (mode == Mode.BYTE && hasEncodingHint) {
CharacterSetECI eci = CharacterSetECI.getCharacterSetECI(encoding);
if (eci != null) {
appendECI(eci, headerBits);
}
}
// 如果适用,为GS1格式的数据附加FNC1模式标头
boolean hasGS1FormatHint = hints != null && hints.containsKey(EncodeHintType.GS1_FORMAT);
if (hasGS1FormatHint && Boolean.parseBoolean(hints.get(EncodeHintType.GS1_FORMAT).toString())) {
// GS1格式的代码在第一位置模式标题中以FNC1作为前缀
appendModeInfo(Mode.FNC1_FIRST_POSITION, headerBits);
}
// (ECI就位后)写入模式标记
appendModeInfo(mode, headerBits);
//单独收集主数据段内的数据,以计算其大小(如果需要)。不要将其添加到主有效载荷中。
BitArray dataBits = new BitArray();
appendBytes(content, mode, dataBits, encoding);
Version version;
if (hints != null && hints.containsKey(EncodeHintType.QR_VERSION)) {
int versionNumber = Integer.parseInt(hints.get(EncodeHintType.QR_VERSION).toString());
version = Version.getVersionForNumber(versionNumber);
int bitsNeeded = calculateBitsNeeded(mode, headerBits, dataBits, version);
if (!willFit(bitsNeeded, version, ecLevel)) {
throw new WriterException("Data too big for requested version");
}
} else {
version = recommendVersion(ecLevel, mode, headerBits, dataBits);
}
BitArray headerAndDataBits = new BitArray();
headerAndDataBits.appendBitArray(headerBits);
// 找到主段的“长度”并写下来
int numLetters = mode == Mode.BYTE ? dataBits.getSizeInBytes() : content.length();
appendLengthInfo(numLetters, version, mode, headerAndDataBits);
// 将数据合并到整个有效负载中
headerAndDataBits.appendBitArray(dataBits);
Version.ECBlocks ecBlocks = version.getECBlocksForLevel(ecLevel);
int numDataBytes = version.getTotalCodewords() - ecBlocks.getTotalECCodewords();
// 正确终止位。
terminateBits(numDataBytes, headerAndDataBits);
// 用纠错码交织数据位。
BitArray finalBits = interleaveWithECBytes(headerAndDataBits,
version.getTotalCodewords(),
numDataBytes,
ecBlocks.getNumBlocks());
QRCode qrCode = new QRCode();
qrCode.setECLevel(ecLevel);
qrCode.setMode(mode);
qrCode.setVersion(version);
// 选择遮罩图案并设置为“qrCode”。
int dimension = version.getDimensionForVersion();
ByteMatrix matrix = new ByteMatrix(dimension, dimension);
// 允许通过提示手动选择要使用的模式
int maskPattern = -1;
if (hints != null && hints.containsKey(EncodeHintType.QR_MASK_PATTERN)) {
int hintMaskPattern = Integer.parseInt(hints.get(EncodeHintType.QR_MASK_PATTERN).toString());
maskPattern = QRCode.isValidMaskPattern(hintMaskPattern) ? hintMaskPattern : -1;
}
if (maskPattern == -1) {
maskPattern = chooseMaskPattern(finalBits, ecLevel, version, matrix);
}
qrCode.setMaskPattern(maskPattern);
// 构建矩阵并将其设置为“qrCode”。
MatrixUtil.buildMatrix(finalBits, ecLevel, version, maskPattern, matrix);
qrCode.setMatrix(matrix);
return qrCode;
}
private static void appendECI(CharacterSetECI eci, BitArray bits) {
bits.appendBits(Mode.ECI.getBits(), 4);
// 这对于高达127的值是正确的,这就是我们现在所需要的。
bits.appendBits(eci.getValue(), 8);
}
决定包含所有提供数据的二维码的最小版本,返回QR码的版本。
难点:需要知道版本,才能知道长度需要多少位。但需要知道需要多少位才能知道版本。因此先假设版本为最小值,然后再计算出正确的版本。
private static Version recommendVersion(ErrorCorrectionLevel ecLevel,
Mode mode,
BitArray headerBits,
BitArray dataBits) throws WriterException {
// 首先,我们假设版本为最小值(1),对版本进行猜测:
int provisionalBitsNeeded = calculateBitsNeeded(mode, headerBits, dataBits, Version.getVersionForNumber(1));
Version provisionalVersion = chooseVersion(provisionalBitsNeeded, ecLevel);
// 用这个猜测来计算正确的版本。
int bitsNeeded = calculateBitsNeeded(mode, headerBits, dataBits, provisionalVersion);
return chooseVersion(bitsNeeded, ecLevel);
}
private static int calculateBitsNeeded(Mode mode,
BitArray headerBits,
BitArray dataBits,
Version version) {
return headerBits.getSize() + mode.getCharacterCountBits(version) + dataBits.getSize();
}
在字母数字模式下使用的表格的代码点,如果表格中没有相应的代码,则为-1。
static int getAlphanumericCode(int code) {
if (code < ALPHANUMERIC_TABLE.length) {
return ALPHANUMERIC_TABLE[code];
}
return -1;
}
通过检查内容选择最佳模式。请注意,“编码”用作提示;如果它是Shift_JIS,并且输入仅为双字节汉字,那么我们返回Mode。
private static Mode chooseMode(String content, Charset encoding) {
if (StringUtils.SHIFT_JIS_CHARSET.equals(encoding) && isOnlyDoubleByteKanji(content)) {
// 如果所有输入都是双字节字符,请选择汉字模式
return Mode.KANJI;
}
boolean hasNumeric = false;
boolean hasAlphanumeric = false;
for (int i = 0; i < content.length(); ++i) {
char c = content.charAt(i);
if (c >= '0' && c <= '9') {
hasNumeric = true;
} else if (getAlphanumericCode(c) != -1) {
hasAlphanumeric = true;
} else {
return Mode.BYTE;
}
}
if (hasAlphanumeric) {
return Mode.ALPHANUMERIC;
}
if (hasNumeric) {
return Mode.NUMERIC;
}
return Mode.BYTE;
}
private static boolean isOnlyDoubleByteKanji(String content) {
byte[] bytes = content.getBytes(StringUtils.SHIFT_JIS_CHARSET);
int length = bytes.length;
if (length % 2 != 0) {
return false;
}
for (int i = 0; i < length; i += 2) {
int byte1 = bytes[i] & 0xFF;
if ((byte1 < 0x81 || byte1 > 0x9F) && (byte1 < 0xE0 || byte1 > 0xEB)) {
return false;
}
}
return true;
}
类似的,还有chooseMaskPattern、chooseVersion等方法,在这里不再赘述。
private static boolean willFit(int numInputBits, Version version, ErrorCorrectionLevel ecLevel) {
// numBytes = 196 (版本7-H)
int numBytes = version.getTotalCodewords();
// getNumECBytes = 130
Version.ECBlocks ecBlocks = version.getECBlocksForLevel(ecLevel);
int numEcBytes = ecBlocks.getTotalECCodewords();
// getNumDataBytes = 196 - 130 = 66
int numDataBytes = numBytes - numEcBytes;
int totalInputBytes = (numInputBits + 7) / 8;
return numDataBytes >= totalInputBytes;
}
static void terminateBits(int numDataBytes, BitArray bits) throws WriterException {
int capacity = numDataBytes * 8;
if (bits.getSize() > capacity) {
throw new WriterException("data bits cannot fit in the QR Code" + bits.getSize() + " > " +
capacity);
}
for (int i = 0; i < 4 && bits.getSize() < capacity; ++i) {
bits.appendBit(false);
}
// 如果最后一个字节不是8位对齐的,将添加填充位。
int numBitsInLastByte = bits.getSize() & 0x07;
if (numBitsInLastByte > 0) {
for (int i = numBitsInLastByte; i < 8; i++) {
bits.appendBit(false);
}
}
// 如果我们有更多的空间,我们将用定义的填充模式填充空间。
int numPaddingBytes = numDataBytes - bits.getSizeInBytes();
for (int i = 0; i < numPaddingBytes; ++i) {
bits.appendBits((i & 0x01) == 0 ? 0xEC : 0x11, 8);
}
if (bits.getSize() != capacity) {
throw new WriterException("Bits size does not equal capacity");
}
}
static void getNumDataBytesAndNumECBytesForBlockID(int numTotalBytes,
int numDataBytes,
int numRSBlocks,
int blockID,
int[] numDataBytesInBlock,
int[] numECBytesInBlock) throws WriterException {
if (blockID >= numRSBlocks) {
throw new WriterException("Block ID too large");
}
// numRsBlocksInGroup2 = 196 % 5 = 1
int numRsBlocksInGroup2 = numTotalBytes % numRSBlocks;
// numRsBlocksInGroup1 = 5 - 1 = 4
int numRsBlocksInGroup1 = numRSBlocks - numRsBlocksInGroup2;
// numTotalBytesInGroup1 = 196 / 5 = 39
int numTotalBytesInGroup1 = numTotalBytes / numRSBlocks;
// numTotalBytesInGroup2 = 39 + 1 = 40
int numTotalBytesInGroup2 = numTotalBytesInGroup1 + 1;
// numDataBytesInGroup1 = 66 / 5 = 13
int numDataBytesInGroup1 = numDataBytes / numRSBlocks;
// numDataBytesInGroup2 = 13 + 1 = 14
int numDataBytesInGroup2 = numDataBytesInGroup1 + 1;
// numEcBytesInGroup1 = 39 - 13 = 26
int numEcBytesInGroup1 = numTotalBytesInGroup1 - numDataBytesInGroup1;
// numEcBytesInGroup2 = 40 - 14 = 26
int numEcBytesInGroup2 = numTotalBytesInGroup2 - numDataBytesInGroup2;
// 合理性检验
// 26 = 26
if (numEcBytesInGroup1 != numEcBytesInGroup2) {
throw new WriterException("EC bytes mismatch");
}
// 5 = 4 + 1.
if (numRSBlocks != numRsBlocksInGroup1 + numRsBlocksInGroup2) {
throw new WriterException("RS blocks mismatch");
}
// 196 = (13 + 26) * 4 + (14 + 26) * 1
if (numTotalBytes !=
((numDataBytesInGroup1 + numEcBytesInGroup1) *
numRsBlocksInGroup1) +
((numDataBytesInGroup2 + numEcBytesInGroup2) *
numRsBlocksInGroup2)) {
throw new WriterException("Total bytes mismatch");
}
if (blockID < numRsBlocksInGroup1) {
numDataBytesInBlock[0] = numDataBytesInGroup1;
numECBytesInBlock[0] = numEcBytesInGroup1;
} else {
numDataBytesInBlock[0] = numDataBytesInGroup2;
numECBytesInBlock[0] = numEcBytesInGroup2;
}
}
static BitArray interleaveWithECBytes(BitArray bits,
int numTotalBytes,
int numDataBytes,
int numRSBlocks) throws WriterException {
// “bits”必须有“getNumDataBytes”字节的数据。
if (bits.getSizeInBytes() != numDataBytes) {
throw new WriterException("Number of bits and data bytes does not match");
}
// 第一步。将数据字节划分为块,并为其生成纠错字节。我们将把分割的数据字节块和纠错字节块存储到“块”中。
int dataBytesOffset = 0;
int maxNumDataBytes = 0;
int maxNumEcBytes = 0;
// 因为我们知道reedsolmon块的数量,所以我们可以用这个数字初始化向量。
Collection<BlockPair> blocks = new ArrayList<>(numRSBlocks);
for (int i = 0; i < numRSBlocks; ++i) {
int[] numDataBytesInBlock = new int[1];
int[] numEcBytesInBlock = new int[1];
getNumDataBytesAndNumECBytesForBlockID(
numTotalBytes, numDataBytes, numRSBlocks, i,
numDataBytesInBlock, numEcBytesInBlock);
int size = numDataBytesInBlock[0];
byte[] dataBytes = new byte[size];
bits.toBytes(8 * dataBytesOffset, dataBytes, 0, size);
byte[] ecBytes = generateECBytes(dataBytes, numEcBytesInBlock[0]);
blocks.add(new BlockPair(dataBytes, ecBytes));
maxNumDataBytes = Math.max(maxNumDataBytes, size);
maxNumEcBytes = Math.max(maxNumEcBytes, ecBytes.length);
dataBytesOffset += numDataBytesInBlock[0];
}
if (numDataBytes != dataBytesOffset) {
throw new WriterException("Data bytes does not match offset");
}
BitArray result = new BitArray();
// 首先,放置数据块。
for (int i = 0; i < maxNumDataBytes; ++i) {
for (BlockPair block : blocks) {
byte[] dataBytes = block.getDataBytes();
if (i < dataBytes.length) {
result.appendBits(dataBytes[i], 8);
}
}
}
//然后,放置纠错块。
for (int i = 0; i < maxNumEcBytes; ++i) {
for (BlockPair block : blocks) {
byte[] ecBytes = block.getErrorCorrectionBytes();
if (i < ecBytes.length) {
result.appendBits(ecBytes[i], 8);
}
}
}
if (numTotalBytes != result.getSizeInBytes()) { // Should be same. 应该是一样的。
throw new WriterException("Interleaving error: " + numTotalBytes + " and " +
result.getSizeInBytes() + " differ.");
}
return result;
}
static void appendModeInfo(Mode mode, BitArray bits) {
bits.appendBits(mode.getBits(), 4);
}
static void appendLengthInfo(int numLetters, Version version, Mode mode, BitArray bits) throws WriterException {
int numBits = mode.getCharacterCountBits(version);
if (numLetters >= (1 << numBits)) {
throw new WriterException(numLetters + " is bigger than " + ((1 << numBits) - 1));
}
bits.appendBits(numLetters, numBits);
}
static void appendBytes(String content,
Mode mode,
BitArray bits,
Charset encoding) throws WriterException {
switch (mode) {
case NUMERIC:
appendNumericBytes(content, bits);
break;
case ALPHANUMERIC:
appendAlphanumericBytes(content, bits);
break;
case BYTE:
append8BitBytes(content, bits, encoding);
break;
case KANJI:
appendKanjiBytes(content, bits);
break;
default:
throw new WriterException("Invalid mode: " + mode);
}
}
static void appendNumericBytes(CharSequence content, BitArray bits) {
int length = content.length();
int i = 0;
while (i < length) {
int num1 = content.charAt(i) - '0';
if (i + 2 < length) {
// 将三个数字字母编码为十位。
int num2 = content.charAt(i + 1) - '0';
int num3 = content.charAt(i + 2) - '0';
bits.appendBits(num1 * 100 + num2 * 10 + num3, 10);
i += 3;
} else if (i + 1 < length) {
// 将两个数字字母编码为七位。
int num2 = content.charAt(i + 1) - '0';
bits.appendBits(num1 * 10 + num2, 7);
i += 2;
} else {
// 将一个数字字母编码为四位。
bits.appendBits(num1, 4);
i++;
}
}
}
static void appendAlphanumericBytes(CharSequence content, BitArray bits) throws WriterException {
int length = content.length();
int i = 0;
while (i < length) {
int code1 = getAlphanumericCode(content.charAt(i));
if (code1 == -1) {
throw new WriterException();
}
if (i + 1 < length) {
int code2 = getAlphanumericCode(content.charAt(i + 1));
if (code2 == -1) {
throw new WriterException();
}
// 将两个字母数字字母编码为11位。
bits.appendBits(code1 * 45 + code2, 11);
i += 2;
} else {
// 将一个字母数字字母编码为六位。
bits.appendBits(code1, 6);
i++;
}
}
}
static void append8BitBytes(String content, BitArray bits, Charset encoding) {
byte[] bytes = content.getBytes(encoding);
for (byte b : bytes) {
bits.appendBits(b, 8);
}
}
static void appendKanjiBytes(String content, BitArray bits) throws WriterException {
byte[] bytes = content.getBytes(StringUtils.SHIFT_JIS_CHARSET);
if (bytes.length % 2 != 0) {
throw new WriterException("Kanji byte size not even");
}
int maxI = bytes.length - 1; // bytes.length must be even
for (int i = 0; i < maxI; i += 2) {
int byte1 = bytes[i] & 0xFF;
int byte2 = bytes[i + 1] & 0xFF;
int code = (byte1 << 8) | byte2;
int subtracted = -1;
if (code >= 0x8140 && code <= 0x9ffc) {
subtracted = code - 0x8140;
} else if (code >= 0xe040 && code <= 0xebbf) {
subtracted = code - 0xc140;
}
if (subtracted == -1) {
throw new WriterException("Invalid byte sequence");
}
int encoded = ((subtracted >> 8) * 0xc0) + (subtracted & 0xff);
bits.appendBits(encoded, 13);
}
}