Skip to content

Commit

Permalink
Create javadoc
Browse files Browse the repository at this point in the history
  • Loading branch information
@flanglet
flanglet committed Feb 6, 2025
1 parent 9511081 commit dd5c032
Show file tree
Hide file tree
Showing 13 changed files with 4,095 additions and 3,991 deletions.
735 changes: 399 additions & 336 deletions java/src/main/java/io/github/flanglet/kanzi/transform/AliasCodec.java
View file

Large diffs are not rendered by default.

1,236 changes: 618 additions & 618 deletions java/src/main/java/io/github/flanglet/kanzi/transform/BWT.java
View file

Large diffs are not rendered by default.

335 changes: 173 additions & 162 deletions java/src/main/java/io/github/flanglet/kanzi/transform/BWTBlockCodec.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -21,177 +21,188 @@
import io.github.flanglet.kanzi.SliceByteArray;


// Utility class to en/de-code a BWT data block and its associated primary index(es)

// BWT stream format: Header (mode + primary index(es)) | Data (n bytes)
// mode (8 bits): xxxyyyzz
// xxx: ignored
// yyy: log(chunks)
// zz: primary index size - 1 (in bytes)
// primary indexes (chunks * (8|16|24|32 bits))

public class BWTBlockCodec implements ByteTransform
{
private static final int BWT_MAX_HEADER_SIZE = 8 * 4;

private final BWT bwt;
private final int bsVersion;


public BWTBlockCodec()
{
this.bwt = new BWT();
this.bsVersion = 6;
}


public BWTBlockCodec(Map<String, Object> ctx)
{
this.bwt = new BWT(ctx);
this.bsVersion = (ctx == null) ? 6 : (int) ctx.getOrDefault("bsVersion", 6);
}


// Return true if the compression chain succeeded. In this case, the input data
// may be modified. If the compression failed, the input data is returned unmodified.
@Override
public boolean forward(SliceByteArray input, SliceByteArray output)
{
if (input.length == 0)
return true;

if (input.array == output.array)
return false;

final int blockSize = input.length;

if (output.length - output.index < getMaxEncodedLength(blockSize))
return false;

int logBlockSize = Global.log2(blockSize);

if ((blockSize & (blockSize - 1)) != 0)
logBlockSize++;

final int pIndexSize = (logBlockSize + 7) >> 3;

if ((pIndexSize <= 0) || (pIndexSize >= 5))
return false;

final int chunks = BWT.getBWTChunks(blockSize);
final int logNbChunks = Global.log2(chunks);

if (logNbChunks > 7)
return false;

int idx0 = output.index;
output.index += (1 + chunks * pIndexSize);

// Apply forward transform
if (this.bwt.forward(input, output) == false)
return false;

final byte mode = (byte) ((logNbChunks << 2) | (pIndexSize - 1));

// Emit header
for (int i=0, idx=idx0+1; i<chunks; i++) {
final int primaryIndex = this.bwt.getPrimaryIndex(i) - 1;
int shift = (pIndexSize - 1) << 3;

while (shift >= 0)
{
output.array[idx++] = (byte) (primaryIndex>>shift);
shift -= 8;
}
}

output.array[idx0] = mode;
return true;
}


@Override
public boolean inverse(SliceByteArray input, SliceByteArray output)
{
if (input.length == 0)
return true;

if (input.array == output.array)
return false;

final int blockSize = input.length;

if (this.bsVersion > 5)
{
// Number of chunks and primary index size in bitstream since bsVersion 6
byte mode = input.array[input.index++];
final int logNbChunks = (mode >> 2) & 0x07;
final int pIndexSize = (mode & 0x03) + 1;
final int chunks = 1 << logNbChunks;
/**
* Utility class to encode and decode a BWT data block and its associated primary index(es).
* <p>
* BWT stream format: Header (mode + primary index(es)) | Data (n bytes)
* <ul>
* <li>mode (8 bits): xxxyyyzz</li>
* <li>xxx: ignored</li>
* <li>yyy: log(chunks)</li>
* <li>zz: primary index size - 1 (in bytes)</li>
* <li>primary indexes (chunks * (8|16|24|32 bits))</li>
* </ul>
*/
public class BWTBlockCodec implements ByteTransform {
private static final int BWT_MAX_HEADER_SIZE = 8 * 4;

private final BWT bwt;
private final int bsVersion;

/**
* Default constructor.
*/
public BWTBlockCodec() {
this.bwt = new BWT();
this.bsVersion = 6;
}

/**
* Constructor with a context map.
*
* @param ctx the context map
*/
public BWTBlockCodec(Map<String, Object> ctx) {
this.bwt = new BWT(ctx);
this.bsVersion = (ctx == null) ? 6 : (int) ctx.getOrDefault("bsVersion", 6);
}

/**
* Performs the forward transform, encoding the input data.
*
* @param input the input byte array
* @param output the output byte array
* @return true if the transform was successful, false otherwise
*/
@Override
public boolean forward(SliceByteArray input, SliceByteArray output) {
if (input.length == 0)
return true;

if (input.array == output.array)
return false;

final int blockSize = input.length;

if (output.length - output.index < getMaxEncodedLength(blockSize))
return false;

int logBlockSize = Global.log2(blockSize);

if ((blockSize & (blockSize - 1)) != 0)
logBlockSize++;

final int pIndexSize = (logBlockSize + 7) >> 3;

if ((pIndexSize <= 0) || (pIndexSize >= 5))
return false;

final int chunks = BWT.getBWTChunks(blockSize);
final int logNbChunks = Global.log2(chunks);

if (logNbChunks > 7)
return false;

int idx0 = output.index;
output.index += (1 + chunks * pIndexSize);

// Apply forward transform
if (!this.bwt.forward(input, output))
return false;

final byte mode = (byte) ((logNbChunks << 2) | (pIndexSize - 1));

// Emit header
for (int i = 0, idx = idx0 + 1; i < chunks; i++) {
final int primaryIndex = this.bwt.getPrimaryIndex(i) - 1;
int shift = (pIndexSize - 1) << 3;

while (shift >= 0) {
output.array[idx++] = (byte) (primaryIndex >> shift);
shift -= 8;
}
}

output.array[idx0] = mode;
return true;
}

/**
* Performs the inverse transform, decoding the input data.
*
* @param input the input byte array
* @param output the output byte array
* @return true if the transform was successful, false otherwise
*/
@Override
public boolean inverse(SliceByteArray input, SliceByteArray output) {
if (input.length == 0)
return true;

if (input.array == output.array)
return false;

final int blockSize = input.length;

if (this.bsVersion > 5) {
// Number of chunks and primary index size in bitstream since bsVersion 6
byte mode = input.array[input.index++];
final int logNbChunks = (mode >> 2) & 0x07;
final int pIndexSize = (mode & 0x03) + 1;
final int chunks = 1 << logNbChunks;

if (chunks != BWT.getBWTChunks(blockSize))
return false;

if (chunks != BWT.getBWTChunks(blockSize))
return false;
final int headerSize = 1 + chunks * pIndexSize;

final int headerSize = 1 + chunks*pIndexSize;
if (blockSize < headerSize)
return false;

if (blockSize < headerSize)
return false;
// Read header
for (int i = 0; i < chunks; i++) {
int shift = (pIndexSize - 1) << 3;
int primaryIndex = 0;

// Read header
for (int i=0; i<chunks; i++)
{
int shift = (pIndexSize - 1) << 3;
int primaryIndex = 0;
// Extract BWT primary index
while (shift >= 0) {
primaryIndex = (primaryIndex << 8) | (input.array[input.index++] & 0xFF);
shift -= 8;
}

// Extract BWT primary index
while (shift >= 0) {
primaryIndex = (primaryIndex << 8) | (input.array[input.index++] & 0xFF);
shift -= 8;
}
if (!this.bwt.setPrimaryIndex(i, primaryIndex + 1))
return false;
}

if (this.bwt.setPrimaryIndex(i, primaryIndex + 1) == false)
return false;
}
input.length = blockSize - headerSize;
} else {
final int chunks = BWT.getBWTChunks(blockSize);

input.length = blockSize - headerSize;
}
else
{
final int chunks = BWT.getBWTChunks(blockSize);
for (int i = 0; i < chunks; i++) {
// Read block header (mode + primary index). See top of file for format
final int blockMode = input.array[input.index++] & 0xFF;
final int pIndexSizeBytes = 1 + ((blockMode >>> 6) & 0x03);

for (int i=0; i<chunks; i++)
{
// Read block header (mode + primary index). See top of file for format
final int blockMode = input.array[input.index++] & 0xFF;
final int pIndexSizeBytes = 1 + ((blockMode >>> 6) & 0x03);
if (input.length < pIndexSizeBytes)
return false;

if (input.length < pIndexSizeBytes)
return false;
input.length -= pIndexSizeBytes;
int shift = (pIndexSizeBytes - 1) << 3;
int primaryIndex = (blockMode & 0x3F) << shift;

input.length -= pIndexSizeBytes;
int shift = (pIndexSizeBytes - 1) << 3;
int primaryIndex = (blockMode & 0x3F) << shift;
// Extract BWT primary index
for (int n = 1; n < pIndexSizeBytes; n++) {
shift -= 8;
primaryIndex |= ((input.array[input.index++] & 0xFF) << shift);
}

// Extract BWT primary index
for (int n=1; n<pIndexSizeBytes; n++)
{
shift -= 8;
primaryIndex |= ((input.array[input.index++] & 0xFF) << shift);
if (!this.bwt.setPrimaryIndex(i, primaryIndex))
return false;
}

if (this.bwt.setPrimaryIndex(i, primaryIndex) == false)
return false;
}
}

// Apply inverse Transform
return this.bwt.inverse(input, output);
}


@Override
public int getMaxEncodedLength(int srcLen)
{
return srcLen + BWT_MAX_HEADER_SIZE;
}
}
}

// Apply inverse transform
return this.bwt.inverse(input, output);
}

/**
* Returns the maximum encoded length, which includes the header size.
*
* @param srcLen the source length
* @return the maximum encoded length
*/
@Override
public int getMaxEncodedLength(int srcLen) {
return srcLen + BWT_MAX_HEADER_SIZE;
}
}
Loading

0 comments on commit dd5c032

Please sign in to comment.