better jp2 parsing based on jpylyzer

img2pdf.py

@@ -21,6 +21,7 @@ import zlib
 import argparse
 import struct
 from datetime import datetime
+from jp2 import parsejp2
 
 def parse(cont, indent=1):
     if type(cont) is dict:
@@ -95,23 +96,21 @@ def main(images, dpi, title=None, author=None, creator=None, producer=None,
     })
 
     for im in images:
+        rawdata = im.read()
         try:
             imgdata = Image.open(im)
         except IOError:
             # test if it is a jpeg2000 image
-            im.seek(0)
-            if im.read(12) != "\x00\x00\x00\x0C\x6A\x50\x20\x20\x0D\x0A\x87\x0A":
+            if rawdata[:12] != "\x00\x00\x00\x0C\x6A\x50\x20\x20\x0D\x0A\x87\x0A":
                 print "cannot read input image"
                 exit(1)
             # image is jpeg2000
+            width, height, ics = parsejp2(rawdata)
             imgformat = "JP2"
-            offset, = struct.unpack(">I", im.read(4))
-            im.seek(28+offset)
-            height, width = struct.unpack(">II", im.read(8))
             if colorspace:
                 color = colorspace
             else:
-                color = "RGB" # TODO: read real colorspace
+                color = ics
             if dpi:
                 dpi_x, dpi_y = dpi, dpi
             else:
@@ -147,12 +146,10 @@ def main(images, dpi, title=None, author=None, creator=None, producer=None,
         # either embed the whole jpeg or deflate the bitmap representation
         if imgformat is "JPEG":
             ofilter = [ "/DCTDecode" ]
-            im.seek(0)
-            imgdata = im.read()
+            imgdata = rawdata
         elif imgformat is "JP2":
             ofilter = [ "/JPXDecode" ]
-            im.seek(0)
-            imgdata = im.read()
+            imgdata = rawdata
             version = 5 # jpeg2000 needs pdf 1.5
         else:
             ofilter = [ "/FlateDecode" ]

jp2.py

@@ -0,0 +1,90 @@
+#!/usr/bin/env python
+#
+# Copyright (C) 2013 Johannes 'josch' Schauer <j.schauer at email.de>
+#
+# this module is heavily based upon jpylyzer which is
+# KB / National Library of the Netherlands, Open Planets Foundation
+# and released under the same license conditions
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU Lesser General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU Lesser General Public License for more details.
+#
+# You should have received a copy of the GNU Lesser General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+
+import struct
+
+def getBox(data, byteStart, noBytes):
+    boxLengthValue = struct.unpack(">I", data[byteStart:byteStart+4])[0]
+    boxType = data[byteStart+4:byteStart+8]
+    contentsStartOffset = 8
+    if boxLengthValue == 1:
+        boxLengthValue = struct.unpack(">Q", data[byteStart+8:byteStart+16])[0]
+        contentsStartOffset = 16
+    if boxLengthValue == 0:
+        boxLengthValue = noBytes-byteStart
+    byteEnd = byteStart + boxLengthValue
+    boxContents = data[byteStart+contentsStartOffset:byteEnd]
+    return (boxLengthValue, boxType, byteEnd, boxContents)
+
+def parse_ihdr(data):
+    height = struct.unpack(">I", data[0:4])[0]
+    width = struct.unpack(">I", data[4:8])[0]
+    return width, height
+
+def parse_colr(data):
+    meth = struct.unpack(">B", data[0:1])[0]
+    if meth != 1:
+        raise Exception("only enumerated color method supported")
+    enumCS = struct.unpack(">I", data[3:])[0]
+    if enumCS == 16:
+        return "RGB"
+    elif enumCS == 17:
+        return "L"
+    else:
+        raise Exception("only sRGB and greyscale color space is supported, got %d"%enumCS)
+
+def parse_jp2h(data):
+    width, height, colorspace = None, None, None
+    noBytes=len(data)
+    byteStart=0
+    boxLengthValue=1 # dummy value for while loop condition
+    while byteStart < noBytes and boxLengthValue != 0:
+        boxLengthValue, boxType, byteEnd, boxContents = getBox(data, byteStart, noBytes)
+        if boxType == 'ihdr':
+            width, height = parse_ihdr(boxContents)
+        elif boxType == 'colr':
+            colorspace = parse_colr(boxContents)
+        byteStart = byteEnd
+    return (width, height, colorspace)
+
+def parsejp2(data):
+    noBytes=len(data)
+    byteStart=0
+    boxLengthValue=1 # dummy value for while loop condition
+    while byteStart < noBytes and boxLengthValue != 0:
+        boxLengthValue, boxType, byteEnd, boxContents = getBox(data, byteStart, noBytes)
+        if boxType == 'jp2h':
+            width, height, colorspace = parse_jp2h(boxContents)
+        byteStart = byteEnd
+    if not width:
+        raise Exception("no width in jp2 header")
+    if not height:
+        raise Exception("no height in jp2 header")
+    if not colorspace:
+        raise Exception("no colorspace in jp2 header")
+    return (width, height, colorspace)
+
+if __name__ == "__main__":
+    import sys
+    width, height, colorspace = parsejp2(open(sys.argv[1]).read())
+    print "width = %d"%width
+    print "height = %d"%height
+    print "colorspace = %s"%colorspace