Merge pull request #18 from DarwinAwardWinner/master

Avoid leaking file descriptors and convert unrecognized colorspaces to RGB
This commit is contained in:
josch 2014-11-06 09:56:41 +01:00
commit f881a00c4b

View file

@ -204,51 +204,53 @@ def convert(images, dpi, x, y, title=None, author=None, creator=None, producer=N
pdf = pdfdoc(3, title, author, creator, producer, creationdate, pdf = pdfdoc(3, title, author, creator, producer, creationdate,
moddate, subject, keywords) moddate, subject, keywords)
for im in images: for imfilename in images:
rawdata = im.read() debug_out("Reading %s"%imfilename, verbose)
im.seek(0) with open(imfilename, "rb") as im:
try: rawdata = im.read()
imgdata = Image.open(im) im.seek(0)
except IOError as e: try:
# test if it is a jpeg2000 image imgdata = Image.open(im)
if rawdata[:12] != "\x00\x00\x00\x0C\x6A\x50\x20\x20\x0D\x0A\x87\x0A": except IOError as e:
error_out("cannot read input image (not jpeg2000)") # test if it is a jpeg2000 image
error_out("PIL: %s"%e) if rawdata[:12] != "\x00\x00\x00\x0C\x6A\x50\x20\x20\x0D\x0A\x87\x0A":
exit(1) error_out("cannot read input image (not jpeg2000)")
# image is jpeg2000 error_out("PIL: %s"%e)
width, height, ics = parsejp2(rawdata) exit(1)
imgformat = "JPEG2000" # image is jpeg2000
width, height, ics = parsejp2(rawdata)
imgformat = "JPEG2000"
if dpi: if dpi:
ndpi = dpi, dpi ndpi = dpi, dpi
debug_out("input dpi (forced) = %d x %d"%ndpi, verbose) debug_out("input dpi (forced) = %d x %d"%ndpi, verbose)
else: else:
ndpi = (96, 96) # TODO: read real dpi ndpi = (96, 96) # TODO: read real dpi
debug_out("input dpi = %d x %d"%ndpi, verbose) debug_out("input dpi = %d x %d"%ndpi, verbose)
if colorspace: if colorspace:
color = colorspace color = colorspace
debug_out("input colorspace (forced) = %s"%(ics)) debug_out("input colorspace (forced) = %s"%(ics))
else:
color = ics
debug_out("input colorspace = %s"%(ics), verbose)
else: else:
color = ics width, height = imgdata.size
debug_out("input colorspace = %s"%(ics), verbose) imgformat = imgdata.format
else:
width, height = imgdata.size
imgformat = imgdata.format
if dpi: if dpi:
ndpi = dpi, dpi ndpi = dpi, dpi
debug_out("input dpi (forced) = %d x %d"%ndpi, verbose) debug_out("input dpi (forced) = %d x %d"%ndpi, verbose)
else: else:
ndpi = imgdata.info.get("dpi", (96, 96)) ndpi = imgdata.info.get("dpi", (96, 96))
debug_out("input dpi = %d x %d"%ndpi, verbose) debug_out("input dpi = %d x %d"%ndpi, verbose)
if colorspace: if colorspace:
color = colorspace color = colorspace
debug_out("input colorspace (forced) = %s"%(color), verbose) debug_out("input colorspace (forced) = %s"%(color), verbose)
else: else:
color = imgdata.mode color = imgdata.mode
debug_out("input colorspace = %s"%(color), verbose) debug_out("input colorspace = %s"%(color), verbose)
debug_out("width x height = %d x %d"%(width,height), verbose) debug_out("width x height = %d x %d"%(width,height), verbose)
debug_out("imgformat = %s"%imgformat, verbose) debug_out("imgformat = %s"%imgformat, verbose)
@ -263,8 +265,15 @@ def convert(images, dpi, x, y, title=None, author=None, creator=None, producer=N
else: else:
# because we do not support /CCITTFaxDecode # because we do not support /CCITTFaxDecode
if color == '1': if color == '1':
debug_out("Converting colorspace 1 to L", verbose)
imgdata = imgdata.convert('L') imgdata = imgdata.convert('L')
color = 'L' color = 'L'
elif color in ("RGB", "L"):
debug_out("Colorspace is OK: %s"%color, verbose)
else:
debug_out("Converting colorspace %s to RGB"%color, verbose)
imgdata = imgdata.convert('RGB')
color = imgdata.mode
imgdata = zlib.compress(imgdata.tostring()) imgdata = zlib.compress(imgdata.tostring())
# pdf units = 1/72 inch # pdf units = 1/72 inch
@ -277,8 +286,6 @@ def convert(images, dpi, x, y, title=None, author=None, creator=None, producer=N
pdf.addimage(color, width, height, imgformat, imgdata, pdf_x, pdf_y) pdf.addimage(color, width, height, imgformat, imgdata, pdf_x, pdf_y)
im.close()
return pdf.tostring() return pdf.tostring()
@ -295,7 +302,7 @@ def valid_date(string):
parser = argparse.ArgumentParser( parser = argparse.ArgumentParser(
description='Lossless conversion/embedding of images (in)to pdf') description='Lossless conversion/embedding of images (in)to pdf')
parser.add_argument( parser.add_argument(
'images', metavar='infile', type=argparse.FileType('rb'), 'images', metavar='infile', type=str,
nargs='+', help='input file(s)') nargs='+', help='input file(s)')
parser.add_argument( parser.add_argument(
'-o', '--output', metavar='out', type=argparse.FileType('wb'), '-o', '--output', metavar='out', type=argparse.FileType('wb'),