package org.basex.util;
import java.util.*;
/**
* This class compresses and decompresses tokens. It is inspired by the
* Huffman coding, but was simplified to speed up processing.
*
* @author BaseX Team 2005-17, BSD License
* @author Christian Gruen
*/
public final class Compress {
/** Private constructor. */
private Compress() { }
/**
* Compresses the specified text. Returns the original text if the packed text is not shorter.
* @param text text to be packed
* @return packed or original text
*/
public static byte[] pack(final byte[] text) {
// only compress texts with more than 4 characters
final int length = text.length - 1;
if(length < 4) return text;
// store length at beginning of array
final byte[] bytes = new byte[length];
int size = Num.set(bytes, length + 1);
// find lower-case and non-ascii characters
int lc = 0, uc = 0, out = 0;
for(final byte t : text) {
lc += t >= 'A' && t <= 'Z' ? -1 : 1;
uc += t >= 0 ? 1 : -1;
}
// too many non-ascii characters: skip compression
if(uc < 0) return text;
// first bit: packer version (0), second bit: mapping type (upper/lower case)
final byte[] map;
if(lc >= 0) {
out = 2;
map = PACK1;
} else {
map = PACK2;
}
// loop through and compress all characters
int in, off = 2;
for(final byte t : text) {
final int b = t >= 0 ? map[t] : t, s;
if(b >= 0x00 && b < 0x08) { // 1 xxx
in = 1 | b << 1;
s = 4;
} else if(b >= 0x08 && b < 0x10) { // 01 xxx
in = 2 | b << 2;
s = 5;
} else if(b >= 0x10 && b < 0x20) { // 001 xxxx
in = 4 | b << 3;
s = 7;
} else if(b >= 0x20 && b < 0x40) { // 0001 xxxxx
in = 8 | b << 4;
s = 9;
} else { // 0000 xxxxxxxx
in = b << 4;
s = 12;
}
for(int i = 0; i < s; i++) {
out |= (in & 1) << off;
in >>>= 1;
off = (off + 1) & 7;
if(off == 0) {
// skip compression if packed array gets too large
if(size == length) return text;
bytes[size++] = (byte) out;
out = 0;
}
}
}
if(off != 0) {
// skip compression if packed array gets too large
if(size == length) return text;
bytes[size++] = (byte) out;
}
return size < length ? Arrays.copyOf(bytes, size) : bytes;
}
/**
* Decompresses the specified text.
* @param text compressed text
* @return unpacked text
*/
public static byte[] unpack(final byte[] text) {
// bit position: skip stored length and packer bit
int pos = (Num.length(text, 0) << 3) + 1;
// choose mapping
final byte[] map = isSet(text, pos++) ? UNPACK1 : UNPACK2;
// decompress all characters
final int size = Num.get(text, 0);
final byte[] bytes = new byte[size];
for(int b = 0; b < size; b++) {
final int bits;
int out = 0;
if(isSet(text, pos++)) { // 1 xxx
bits = 3;
} else if(isSet(text, pos++)) { // 01 xxx
bits = 3;
out = 0x08;
} else if(isSet(text, pos++)) { // 001 xxxx
bits = 4;
out = 0x10;
} else if(isSet(text, pos++)) { // 0001 xxxxx
bits = 5;
out = 0x20;
} else { // 0000 xxxxxxxx
bits = 8;
}
for(int bit = 0; bit < bits; bit++, pos++) {
if((text[pos >>> 3] & 1 << (pos & 7)) != 0) out |= 1 << bit;
}
bytes[b] = (byte) (out >= 0x80 ? out : map[out]);
}
return bytes;
}
/**
* Checks if a specified bit is set.
* @param txt text to be unpacked
* @param pos current position
* @return result
*/
private static boolean isSet(final byte[] txt, final int pos) {
return (txt[pos >>> 3] & 1 << (pos & 7)) != 0;
}
/** First mapping for unpacking data (lower-case texts). */
private static final byte[] UNPACK1 = {
0x20, 0x61, 0x65, 0x6E, 0x69, 0x6F, 0x72, 0x73, // encode via 1 xxx
0x74, 0x6C, 0x75, 0x68, 0x64, 0x63, 0x67, 0x6D, // encode via 01 xxx
0x70, 0x79, 0x62, 0x6B, 0x66, 0x76, 0x43, 0x53, // encode via 001 xxxx
0x77, 0x4D, 0x41, 0x42, 0x50, 0x7A, 0x2E, 0x0A,
0x54, 0x52, 0x4B, 0x4C, 0x47, 0x4E, 0x48, 0x6A, // encode via 0001 xxxxx
0x45, 0x49, 0x44, 0x46, 0x4A, 0x78, 0x4F, 0x71,
0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37,
0x38, 0x39, 0x3A, 0x2D, 0x27, 0x2C, 0x22, 0x3F,
0x56, 0x57, 0x55, 0x5A, 0x59, 0x51, 0x58, 0x09,
0x40, 0x28, 0x2F, 0x29, 0x2B, 0x7E, 0x21, 0x23, // encode via 0000 xxxxxxxx
0x24, 0x25, 0x26, 0x2A, 0x3B, 0x3C, 0x3D, 0x3E,
0x5B, 0x5C, 0x5D, 0x5E, 0x5F, 0x60, 0x7B, 0x7C,
0x7D, 0x7F, 0x00, 0x01, 0x02, 0x03, 0x04, 0x05,
0x06, 0x07, 0x08, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F,
0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17,
0x18, 0x19, 0x1A, 0x1B, 0x1C, 0x1D, 0x1E, 0x1F
};
/** Second mapping for unpacking data (upper-case texts). */
private static final byte[] UNPACK2 = new byte[UNPACK1.length];
/** First mapping for packing data (lower-case texts). */
private static final byte[] PACK1 = new byte[UNPACK1.length];
/** Second mapping for packing data (upper-case texts). */
private static final byte[] PACK2 = new byte[UNPACK1.length];
// initializes the character mappings
static {
final int pl = UNPACK1.length;
for(int p = 0; p < pl; p++) {
final byte b1 = UNPACK1[p];
// swap lower and upper case in second mapping
final byte b2 = (byte) (b1 >= 'A' && b1 <= 'Z' ? b1 + 0x20 :
b1 >= 'a' && b1 <= 'z' ? b1 - 0x20 : b1);
UNPACK2[p] = b2;
PACK1[b1] = (byte) p;
PACK2[b2] = (byte) p;
}
}
}