[reportlab-users] Non-JPEG CMYK images

[Stevens, Ian]

Sorry ... I had thought I had included the filename.  This is patched
from r323 in trunk/reportlab/pdfgen/pdfimages.py. 

The way I understand is that when an image is added either directly on
the canvas or as a flowable, Canvas.drawInlineImage() is called.  That
in turn creates a PDFImage object, passing in the image to the
constructor.  From there, PDFImage.getImageData() is called, which
determines whether the image is a string, file or PIL object.  If it's a
PIL object, PDFImage.PIL_imagedata() is called, which converts the image
to RGB.  That's the method I've patched.

I didn't know about ImageReader.  I suppose I could try to wrap my PIL
image around that, but I suspect it will also convert the image to RGB,
given that the first thing Canvas.drawImage() does is call
ImageReader.getRGBData().

Ian.


> -----Original Message-----

> From: reportlab-users-bounces at reportlab.com 

> [mailto:reportlab-users-bounces at reportlab.com] On Behalf Of 

> Robin Becker

> Sent: Tuesday, February 26, 2008 6:34 AM

> To: Support list for users of Reportlab software

> Subject: Re: [reportlab-users] Non-JPEG CMYK images

> 

> Stevens, Ian wrote:

> .........

> >  

> > What follows is the patch I'm using.  I don't know exactly what PIL 

> > image modes of '1', 'L', 'RGBX' or 'RGBA' mean in this context, so 

> > those modes get converted to 'RGB':

> >  

> > 98c98,107

> > <         myimage = image.convert('RGB')

> > ---

> >>         # Use the colorspace in the image

> >>         mode = image.mode

> >>         if mode == 'CMYK':

> >>             myimage = image

> >>             colorSpace = 'DeviceCMYK'

> >>         else:

> >>             myimage = image.convert('RGB')

> >>             colorSpace = 'RGB'

> >>

> > 104c113

> > <         imagedata=['BI /W %d /H %d /BPC 8 /CS /RGB /F 

> [/A85 /Fl] ID' %

> > (imgwidth, imgheight)]

> > ---

> >>         imagedata=['BI /W %d /H %d /BPC 8 /CS /%s /F [/A85 

> /Fl] ID' %

> > (imgwidth, imgheight, colorSpace)]

> > 

> .......

> it's a bit hard to figure out what you're patching without 

> either of the filename and or svn version :)

> 

> I looked in the latest trunk code and the ImageReader object 

> has this if sys.platform[0:4] == 'java':

> ........

> else:

> 	im = self._image

> 	mode = self.mode = im.mode

> 	if mode=='RGBA':

> 		self._dataA = ImageReader(im.split()[3])

> 		im = im.convert('RGB')

> 		self.mode = 'RGB'

> 	elif mode not in ('L','RGB','CMYK'):

> 		im = im.convert('RGB')

> 		self.mode = 'RGB'

> self._data = im.tostring()

> 

> so in the PIL branch we're trying to preserve the existing 

> mode. Are you saying that even so the PIL images get 

> converted somewhere else to RGB? I suspect that would not be 

> intended, but this image stuff is fairly horrendous.

> --

> Robin Becker

> _______________________________________________

> reportlab-users mailing list

> reportlab-users at reportlab.com

> http://two.pairlist.net/mailman/listinfo/reportlab-users

>