GZIP文件格式解析和Inflate静态Huffman解压缩

本文主要是介绍GZIP文件格式解析和Inflate静态Huffman解压缩，希望对大家解决编程问题提供一定的参考价值，需要的开发者们随着小编来一起学习吧！

GZIP是封装了Deflate压缩的格式文件；Deflate使用了无压缩、Huffman+LZ77进行压缩；解压是Inflate，Huffman包括静态Huffman压缩和动态Huffman压缩两种模式。

Java语言实现了GZIP格式解析、Inflate的静态Huffman解压缩、CRC32校验算法。

gzip文件格式解析代码

	    BinaryInputStream bis = new BinaryInputStream(bytes); // 二进制字节流读取类Header header = new Header();  // 读取GZIP headerheader.ID1 = bis.ReadUInt8();  // GZIP ID1header.ID2 = bis.ReadUInt8();  // GzIP ID2header.CM = bis.ReadUInt8();  header.Flag = bis.ReadUInt8();header.Time = bis.ReadUnix32TimeStamp(); header.Xfl = bis.ReadUInt8();header.OS = bis.ReadUInt8();gzip.header = header;Console.printlnf("ID=%xb%xb", header.ID1, header.ID2); // 打印Console.printlnf("CM=%xb", header.CM);Console.printlnf("Flag=%xb", header.Flag);Console.printlnf("MTime=%S", header.Time);Console.printlnf("XFL=%xb", header.Xfl);Console.printlnf("OS=%xb", header.OS);if ( (header.Flag & 0b00001000) != 0) {  // 如果标志位是文件名String filename = bis.ReadCString(); // 读取字节流until 0->char[]->stringConsole.printlnf("filename=%S", filename);}else // 其他待实现throw new java.lang.IllegalArgumentException(header.Flag +"");byte[] result ; // 解压后的字节int data_len ;  // 解压前数据长度int crc;        // 程序计算解压后的CRC32（见上篇文章）if (header.CM == 8) { // deflate data_len = bytes.length - bis.GetPosition() - 8;			byte[] data = bis.ReadBytes(data_len); // 解压缩前数据	result = Deflate.uncompress(data);  // 解压缩，返回解压后字节流crc = CRC.CRC32(result); // 计算循环冗余码}else // 待实现throw new java.lang.IllegalArgumentException(header.CM +""); gzip.crc = bis.ReadUInt32(); // gzip文件自身存储的crc32值gzip.isize = bis.ReadInt32(); System.out.println("gzip crc="+ Long.toHexString(gzip.crc) +",calc-crc=" + Integer.toHexString( crc) );

解析结果如下：

显示了GZIP标志、压缩方法、压缩时间、原始文件名、操作系统类型、CRC校验值

GZIP中存储的CRC32值(gzip crc)==解压后计算的CRC32值(calc crc)。(0xa93145a2)

Inflate -静态Huffman解压缩：

	// 读取扩展Codeprivate static int ReadExtCode(BitsInputStream bis, int len) {bis.setOrder(BitOrder.LeftIsHigh);return bis.ReadBits(len);}// 读取距离private static int ReadDistance(BitsInputStream bis) {bis.setOrder(BitOrder.RightIsHigh);return bis.ReadBits(5);}// 读取Codeprivate static int ReadCode(BitsInputStream bis) {bis.setOrder(BitOrder.RightIsHigh);int code = bis.ReadBits(7);Integer value = FixHuffmanTable_7.get(code); // 7位长查表if (value == null) {int ext = bis.ReadBit();code = (code << 1 | ext);value = FixHuffmanTable_8.get(code); // 8位长查表if (value == null) {ext = bis.ReadBit();code = (code << 1 | ext);value = FixHuffmanTable_9.get(code); // 9位长查表if (value == -1)throw new java.lang.IllegalArgumentException(code + "");}}return value;}// Deflate解压缩public static byte[] uncompress(final byte[] _input) throws IOException {IntArrayBuffer baos = new IntArrayBuffer(); // 输出窗口// 位流读取类BitsInputStream bis = new BitsInputStream(_input);while (true) {int bFinal = bis.ReadBits(1); // 读取Deflate头，0 – 还有后续子块；1 – 该子块是最后一块。int bType = bis.ReadBits(2);  // 读取Deflate头，00 – 不压缩；01 – 静态Huffman编码压缩；10 – 动态Huffman编码压缩；11 – 保留if (bType == 0) { // 无压缩,未实现int len = bis.ReadBits(16);int nlen = bis.ReadBits(16);assert len + nlen == 65535;throw new java.lang.UnsupportedOperationException(bType + ""); }else if (bType == 1) { // fixed Huffmanwhile (true) {int value = ReadCode(bis); // 读取Huffman code// 根据literal范围判断if (value >= 0 && value <= 255) // literalbaos.Write(value);else if (value == 256) // 结束标志break ;else if (value >= 257 && value <= 285) { // 距离或长度int length = LengthExtraCodeLengthsTable.get(value);int bits = LengthExtraCodeBitsTable.get(value); if (bits != 0) {int ext =  ReadExtCode(bis, bits); // 读取长度length = length + ext; }value = ReadDistance(bis); // 读取距离int distance = DistanceExtraCodeLengthsTable.get(value);bits = DistanceExtraCodeBitsTable.get(value);if (bits != 0) {int ext =ReadExtCode(bis , bits); // 读取距离distance = distance + ext; }// LZ77滑动窗口计算获取量int[] arr = baos.GetInts();int d = arr.length - distance;if (d < 0) {d = 0;length = length + distance - arr.length;}// 读取滑动窗口，写入到结果for (int i=0; i<length; i++) {	int m = arr[ d + i];baos.Write(m);arr = baos.GetInts();}}else throw new java.lang.IllegalArgumentException(value+ "");}}else if (bType == 2) { // dynamic huffman 动态Huffman待实现throw new java.lang.UnsupportedOperationException(bType + "");}else throw new java.lang.IllegalArgumentException(bType + "");if (bFinal == 1) // 最后一个块break ;}// 结果int[] b = baos.GetInts();byte[] result = new byte[b.length];for (int i=0; i<b.length; i++) result[i] = (byte) b[i];return result;

测试结果如下：