Field Detail

• imgInfo

  public final ImageInfo imgInfo

• rowNum

  protected int rowNum

  last writen row number, starting from 0

• currentChunkGroup

  protected int currentChunkGroup

  Current chunk grounp, (0-6) already read or reading

  see ChunksList

• filterStrat

  protected IFilterWriteStrategy filterStrat

  PNG filter strategy

• filterPreserve

  protected boolean filterPreserve

  If the ImageLine has a valid filterType (and origin image is not interlaced) that file type will be used.

  Default: false

Constructor Detail

• PngWriter

  public PngWriter(File file,
           ImageInfo imgInfo,
           boolean allowoverwrite)

  Opens a file for writing.

  Sets shouldCloseStream=true. For more info see PngWriter(OutputStream, ImageInfo)

  Parameters:

  file -

  imgInfo -

  allowoverwrite - If false and file exists, an PngjOutputException is thrown

• PngWriter

  public PngWriter(File file,
           ImageInfo imgInfo)

  See Also:

  (overwrite=true)

• PngWriter

  public PngWriter(OutputStream outputStream,
           ImageInfo imgInfo)

  Constructs a new PngWriter from a output stream. After construction nothing is writen yet. You still can set some parameters (compression, filters) and queue chunks before start writing the pixels.

  Parameters:

  outputStream - Open stream for binary writing

  imgInfo - Basic image parameters

Method Detail

• queueChunk

  public void queueChunk(PngChunk chunk)

• copyChunksFrom

  public void copyChunksFrom(ChunksList chunks,
                    int copyMask)

  Sets an origin (typically from a PngReader) of Chunks to be copied. This should be called only once, before starting writing the rows. It doesn't matter the current state of the PngReader reading, this is a live object and what matters is that when the writer writes the pixels (IDAT) the reader has already read them, and that when the writer ends, the reader is already ended (all this is very natural).

  Apart from the copyMask, there is some addional heuristics:

  - The chunks will be queued, but will be written as late as possible (unless you explicitly set priority=true)

  - The chunk will not be queued if an "equivalent" chunk was already queued explicitly. And it will be overwriten another is queued explicitly.

  Parameters:

  chunks -

  copyMask - Some bitmask from ChunkCopyBehaviour

  See Also:

  copyChunksFrom(ChunksList, ChunkPredicate)

• copyChunksFrom

  public void copyChunksFrom(ChunksList chunks)

  Copy all chunks from origin. See copyChunksFrom(ChunksList, int) for more info

• copyChunksFrom

  public void copyChunksFrom(ChunksList chunks,
                    ChunkPredicate predicate)

  Copy chunks from origin depending on some ChunkPredicate

  Parameters:

  chunks -

  predicate - The chunks (ancillary or PLTE) will be copied if and only if predicate matches

  See Also:

  for more info

• computeCompressionRatio

  public double computeCompressionRatio()

  Computes compressed size/raw size, approximate.

  Actually: compressed size = total size of IDAT data , raw size = uncompressed pixel bytes = rows * (bytesPerRow + 1). This must be called after pngw.end()

• end

  public void end()

  Finalizes all the steps and closes the stream. This must be called after writing the lines.

• close

  public void close()

  Closes and releases resources

  This is normally called internally from end(), you should only call this for aborting the writing and release resources (close the stream).

  Idempotent and secure - never throws exceptions

• getChunksList

  public ChunksListForWrite getChunksList()

  returns the chunks list (queued and writen chunks)

• getMetadata

  public PngMetadata getMetadata()

  Retruns a high level wrapper over for metadata handling

• setCompLevel

  public void setCompLevel(int compLevel)

  Sets compression level of ZIP algorithm.

  This must be called just after constructor, before starting writing.

  See also setFilterType()

  Parameters:

  compLevel - between 0 and 9 (default:6 , recommended: 6 or more)

• setFilterType

  public void setFilterType(FilterType filterType)

  Sets internal prediction filter type, or strategy to choose it.

  This must be called just after constructor, before starting writing.

  See also setFilterStrategy setCompLevel()

  Parameters:

  filterType - One of the five prediction types or strategy to choose it (see PngFilterType) Recommended values: DEFAULT (default) or AGGRESIVE

• setFilterStrategy

  public void setFilterStrategy(IFilterWriteStrategy filterS)

  Set an alternative strategy for selecting the prediction filter

  Parameters:

  filterS -

• isFilterPreserve

  public boolean isFilterPreserve()

  See Also:

  filterPreserve

• setFilterPreserve

  public void setFilterPreserve(boolean filterPreserve)

  See Also:

  filterPreserve

• setIdatMaxSize

  public void setIdatMaxSize(int idatMaxSize)

  Sets maximum size of IDAT fragments. This has little effect on performance you should rarely call this

  Parameters:

  idatMaxSize - default=0 : use defaultSize (64K)

• setShouldCloseStream

  public void setShouldCloseStream(boolean shouldCloseStream)

  If true, output stream will be closed after ending write

  default=true

• setDeflaterStrategy

  public void setDeflaterStrategy(int deflaterStrategy)

  Deflater strategy: one of Deflater.FILTERED Deflater.HUFFMAN_ONLY Deflater.DEFAULT_STRATEGY

  Default: Deflater.FILTERED (for filter NONE usually DEFAULT_STRATEGY is a little better

• writeRow

  public void writeRow(IImageLine imgline)

  Writes next row, does not check row number.

  Parameters:

  imgline -

• writeRows

  public void writeRows(IImageLineSet<? extends IImageLine> imglines)

  Writes the full set of row. The ImageLineSet should contain (allow to acces) imgInfo.rows

• writeRow

  public void writeRow(IImageLine imgline,
              int rownumber)

• writeRowInt

  public void writeRowInt(int[] buf)

  Utility method, uses internaly a ImageLineInt