Class: PNGReader

• Inheritance
• Description
• Related information
• Class protocol
  • accessing
  • initialization
  • queries
  • testing
• Instance protocol
  • accessing
  • hooks
  • reading
  • reading-private
  • reading-private chunks
  • reading-private filtering
  • reading-private pixel copy
  • writing
  • writing-private
  • writing-private chunks
• Examples

Inheritance:

   Object
   |
   +--ImageReader
      |
      +--PNGReader

Package:

stx:libview2

Category:

Graphics-Images-Readers

Version:

rev: 1.80 date: 2019/06/06 21:21:49

user: cg

file: PNGReader.st directory: libview2

module: stx stc-classLibrary: libview2

Author:

Claus Gittinger

Description:

This class provides methods for loading and saving PNG pictures.
It is currenty unfinished (some interlaced image formats are unsupported, for example: grayscale+alpha)
In the meantime, use a pngtoXXX converter for interlaced images, and read XXX.

[caveats:]
    writer can only store mask with depth24 images (for now).
    writer only generates unfiltered non-interlaced data.
    
[Special:]
    the EnforcedImageTypeQuery is asked for;
    if #rgb is returned AND the image is rgba, then the alpha channel is ignored
    and an rgb (Depth24Image) is returned instead.

    Proprietary chunks can be extracted by defining a specialChunkHandler,
    which is invoked (indexed by chunkType) with the chunk data when such a
    chunk is encountered. This allows for private data to be extracted.
    (labview does this, and expecco can do this)
    Proprietary chunks can be written by setting the specialChunks collection
    before writing.
    

Related information:

    Image
    Form
    Icon
    BlitImageReader
    FaceReader
    GIFReader
    JPEGReader
    MacOSXIconReader
    PBMReader
    PCXReader
    ST80FormReader
    SunRasterReader
    TargaReader
    TIFFReader
    WindowsIconReader
    XBMReader
    XPMReader
    XWDReader

Class protocol:

 pngHeader

 initialize

install myself in the Image classes fileFormat table for the `.png' extension.

 canRepresent: anImage

return true, if anImage can be represented in my file format.
Any image without mask is supported;
only depth24 images with mask are (currently).

 file: filename hasSpecialChunk: chunkTag

true if there is a special chunk with chunkTag in filename

 isValidImageFile: aFileName

return true, if aFileName contains a PNG image

usage example(s):

self isValidImageFile:'/home/cg/AudioExplorer_51_files/use_small.png' self isValidImageFile:'C:\Users\cg\Desktop\croquet\cobalt-base-current-build-20090210\cobalt-base-current-build-20090210\content\models\textures\checkerboard.png' self isValidImageFile:'C:\Dokumente und Einstellungen\cg\Desktop\misc\PNGs\Delete.png'

Instance protocol:

 extractSpecialChunksOnly: aBoolean

if set to true, only special chunks will be processed via
corresponding specialChunkHandlers.
No image will be extracted.
The default is (of course) false.

 makeImage

(comment from inherited method)
return the image as represented by the values found in my
instvars; these have been typically set as a side effect of
the fromStream:-method, which reads images.

 pngHeader

** This is an obsolete interface - do not use it (it may vanish in future versions) **

 specialChunks: aDictionaryOfChunkTypesAndBlobs

Modified (format): / 27-10-2018 / 16:09:56 / Claus Gittinger

 specialChunkHandlerAt: chunkType put: aHandlerBlock

define a handler action, which processes unknown chunks.
The handler block will be called with one or two arguments:
the binary data of the chunk, and optionally the reader itself

 fromStream: aStream

read a stream containing a PNG image.
Leave the image description in instance variables
(i.e. to get the image, ask with image).

 generateMaskFromPaletteAlphaEntries

if there is only one entry to care for...

 getChunk

ignored - for now

 getIHDRChunk

for now - might be changed by setColorType.

 handleText: text keyword: keyword

things like exif data etc.

 setColorType: colorType

0

 doPass: pass

Certain interlace passes are skipped with certain small image dimensions

 processBKGDChunkLen: len

 processCHRMChunkLen: len

cHRM Primary chromaticities chunk - currently unhandled

 processChunk: type len: len

---since the compressed data can span multiple
chunks, stitch them all together first. later,
if memory is an issue, we need to figure out how
to do this on the fly---

 processEXEXChunkLen: len

this cunk contains an expecco sample network.

 processGAMAChunkLen: len

 processGlobalIDATChunk

 processHISTChunkLen: len

ignored

 processICCPChunkLen: len

ignored

 processIDATChunkLen: len

 processIHDRChunkLen: len

ignored

 processITXTChunkLen: len

international (i.e.utf8) textual data.

 processInterlacedDATA: len

 processInterlacedGlobalDATA

adam7 interlace method

 processNIVIChunkLen: len

this cunk contains a labView vi.

 processNonInterlacedDATA: len

 processNonInterlacedGlobalDATA

data n

 processPHYSChunkLen: len

physical pixel chunk - currently unhandled

 processPLTEChunkLen: len

read a color palette

 processSBITChunkLen: len

 processSPALChunkLen: len

ignored

 processSRGBChunkLen: len

ignored

 processTEXTChunkLen: len

textual data in iso8859 coding.

 processTIMEChunkLen: len

 processTRNSChunkLen: len

Modified (format): / 17-02-2017 / 16:57:27 / cg

 processZTXTChunkLen: len

compressed text

 filterAverage: count

Use the average of the pixel to the left and the pixel above as a predictor

 filterHorizontal: count

use the pixel to the left as a predictor

 filterNone: count

no filter - scanline is as is

 filterPaeth: count

Select one of (the pixel to the left, the pixel above and the pixel to above left) to
predict the value of this pixel

 filterScanline: filterType count: count

 filterVertical: count

Use the pixel above as a predictor

 paethPredictLeft: l above: a aboveLeft: al

Predicts the value of a pixel based on nearby pixels, based on Paeth (GG II, 1991)

 copyPixels: y at: startX by: incX

Handle interlaced pixels of supported colorTypes

 copyPixelsGray: y at: startX by: incX

Handle interlaced pixels of supported colorTypes.
Untested code - please verify

 copyPixelsGrayAlpha: y at: startX by: incX

Handle interlaced pixels of supported colorTypes.

 copyPixelsIndexed: y at: startX by: incX

 copyPixelsRGB: y at: startX by: incX

Handle interlaced pixels of supported colorTypes.
Untested code - please verify

 copyPixelsRGBA: y at: startX by: incX

Handle interlaced pixels of supported colorTypes.
Untested code - please verify

 save: image onStream: aStream

save image in PNG-file-format onto aStream

 determinePaletteIndexForMaskedPixels

if all masked pixels are 0, and 0 is not used elsewhere in the image,

 determineColorTypeAndHasMask

sets colorType as side effect;
returns boolean if a mask is present

 writeChunk: chunkTypeChars size: len with: aBlock

 writeEndChunk

 writeIHDRChunk

make it grayAlpha / rgba

 writeImageDataChunk

for now - only support depth24 + mask

 writePaletteChunk

 writeSpecialChunks

 writeTRNSChunk

Examples: