Hi,
I had to replace .FromBufferRGBA with .FromBuffer in line 518 of
wx/lib/pdfviewer/viewer.py to get the demo working . It seems that
the PyMuPDF (PyMuPDF-1.10.0) object fitz.fitz.Pixmap does not
provide the alpha channel information. The len of pix.samples was
pix.widthpix.height3 . pix.alpha was 0.
Environment: ubuntu 17.04, python_3.6.1, mupdf 1.10a .
regards
Uwe
The version of wx/lib/pdfviewer/viewer.py in wxPython-4.0.0a1 was tested using PyMuPDF 1.92 and Python 3.4 on both Windows 10 and OS X in July 2016. The Windows machine has since been rebuilt with the March 2017 release of Windows 10 and has Python 3.6. I have just installed the pre-built Windows version of PyMuPDF 1.10.0.0 there from https://github.com/JorjMcKie/PyMuPDF-optional-material which has the associated comment “Defaulting alpha=True b/o backward compatibility” and the pdf viewer works fine with it. On the line you refer to, in the call to wx.Bitmap.FromBufferRGBA(pix.width, pix.height, pix.samples) the len of pix.samples is width * height * 4.
Maybe your version of PyMuPDF is not compatible with the MuPDF you have built. The dialogue going on at https://github.com/rk700/PyMuPDF/issues/90 might be relevant.
Regards
David Hughes
I then think the problem is the following change ( https://mupdf.com/news.html ) :
MuPDF 1.10-rc1 (2016-11-09) : Alpha channel is now optional in pixmaps.
Am Dienstag, 25. April 2017 14:05:30 UTC+2 schrieb David Hughes:
The version of wx/lib/pdfviewer/viewer.py in wxPython-4.0.0a1 was tested using PyMuPDF 1.92 and Python 3.4 on both Windows 10 and OS X in July 2016. The Windows machine has since been rebuilt with the March 2017 release of Windows 10 and has Python 3.6. I have just installed the pre-built Windows version of PyMuPDF 1.10.0.0 there from https://github.com/JorjMcKie/PyMuPDF-optional-material which has the associated comment “Defaulting alpha=True b/o backward compatibility” and the pdf viewer works fine with it. On the line you refer to, in the call to wx.Bitmap.FromBufferRGBA(pix.width, pix.height, pix.samples) the len of pix.samples is width * height * 4.
Maybe your version of PyMuPDF is not compatible with the MuPDF you have built. The dialogue going on at https://github.com/rk700/PyMuPDF/issues/90 might be relevant.
Regards
David Hughes
I don’t know where your version of pyMuPDF was obtained from or when it was created, but the version available now from GitHub - pymupdf/PyMuPDF: PyMuPDF is a high performance Python library for data extraction, analysis, conversion & manipulation of PDF (and other) documents. was modified on April 3 and the default keyword argument in the definition of getPixmap in fitz/utils.py was changed from alpha=False to alpha=True. So if you download and install this version, wx.lib.pdfviewer should work without requiring any changes.
Regards, David
On Tuesday, April 25, 2017 at 2:35:11 PM UTC+1, uwe wrote:
I then think the problem is the following change ( https://mupdf.com/news.html ) :
MuPDF 1.10-rc1 (2016-11-09) : Alpha channel is now optional in pixmaps.