[OS X TeX] OT: Tool for Comparing PDF files ?

Axel E. Retif axretif at igo.com.mx
Thu Mar 8 11:17:07 CET 2007


On Mar 2, 2007, at 09:41, Michael Sternberg wrote:

> Hello,
>
> On Mar 1, 2007, at 9:34 , Steffen Wolfrum wrote:
>> Does someone know a tool for comparing PDF documents?
>>
>> Now and then I make small changes in the source files and would  
>> feel saver if I'd had a tool that would show me when a resulting  
>> PDF has / has not differences (to a PDF that was made before I  
>> made the changes) …
>>
>
> Try the script tacked-on below.  It does a graphical diff:
>
> 	diffps -h
> 	diffps fileA.pdf fileB.pdf
>
> You need the netpbm[plus] package and Ghostscript.

Your shell script is wonderful! Thank you. I tried it with two  
identical PDFs, and it reported nothing; then I changed (with  
Acrobat) just 1 letter in one of the two 224-page long PDFs, and it  
found the difference.

Just one thing, though ---it calls pnmarith instead of the new  
pamarith. According to

http://netpbm.sourceforge.net/doc/pnmarith.html

pnmarith is obsolete. (And Gerben Wierda's Netpbm i-package comes  
with the new pamarith, not pnmarith.)

Thank you again,

Axel

> By default, it uses "xv" ("Preview" on MacOS) to display differing  
> pages.  Use "-x foo" to specify another viewer, which must read ppm  
> and png files.
>
> For repeated uses, it uses a page-cache, which you can override  
> with -f and clean with -c.
>
>
> Regards, Michael
> ------------------------------------------------------------
> #!/bin/bash
> # compare pages in two similar ps-files by highlighting their  
> differences
> # (uses grayscale pixmaps for comparing)
> #
> # Usage:  (see -h)
> #
> # Created by Michael Sternberg, 2001-2007. Use at your own risk.
>
> PROGRAM=`basename $0`
>
> CACHE=.diffps
> PAGES="*"
> RES=72
> VIEWER="xv -nolimit -24"
> PAIR_FILE=pairs
> VIEWS=1
> HIST_THRESHOLD=1
>
> case `uname` in
>   Darwin)	VIEWER="open -a Preview" ;;
> esac
>
> Usage () {
>     cat << EOT
> Compare postscript/pdf files visually.
> Usage: $PROGRAM [options] file1 [file2 | dir]
>
>   If file2 is not given, the latest  version from CVS is used.
>
> Options:
>   Page rendering:
>       -d directory
>       		directory for page cache (default: "$CACHE")
>
>       -p pages
>       		view only the given pages (quoted shell glob pattern)
> 		(default: "$PAGES")
>
>       -t threshold
>       		minimum number of pixels to differ (default: $HIST_THRESHOLD)
>
>       -r res	Resolution for pixmap rendering (default: $RES)
>
>       -f	re-do comparison (force; discard cache)
>
>   Viewing:
>       -0	report only
>       -1	view differing pages in diff-mode (red = recent; default)
>       -2	view differing pages pairwise
>       -3	both of the above
>       -x viewer	specify image viewer for above (default: xv)
>
>   General:
>       -h	This help.
>       -c	clean cache
>
> Created by Michael Sternberg, 2001-2007. Use at your own risk.
> EOT
>     exit
> }
>
> Clean_Cache () {
>     case $CACHE in
>       */*)	echo $CACHE: not a subdirectory -- please clean  
> manually. 1>&2
> 		exit ;;
>     esac
>     rm -rf $CACHE		# better know what you're doing
> }
>
>
>     # parse options
> while :
> do
>     case "$1" in
>       -d)   CACHE=$2; shift 2 ;;
>       -p)   PAGES=$2; shift 2 ;;
>       -r)   RES=$2; shift 2 ;;
>       -f)   FORCE=1; shift ;;
>       -t)   HIST_THRESHOLD=$2; shift 2 ;;
>
>       -0)   VIEWS=0; shift ;;
>       -1)   VIEWS=1; shift ;;
>       -2)   VIEWS=2; shift ;;
>       -3)   VIEWS=3; shift ;;
>       -x)   VIEWER=$2; shift 2 ;;
>
>       -c)   CLEAN=1; shift ;;
>       -h)   Usage ;;
>
>       -*)   echo $0: unknown option 1>&2
> 	    Usage
> 	    exit 1 ;;
>       *)    break ;;
>     esac
> done
>
>     # clean cache.  Exit if this is the only task.
> if [ -n "$CLEAN" ]; then
>     Clean_Cache
>     case $# in
>       0)	exit ;;
>     esac
> fi
>
>     # attempt to create cache dir
> mkdir $CACHE 2> /dev/null
>
> A_PS="$1"
> B_PS="${2-$CACHE}"
> [ -d "$B_PS" ] && B_PS="$B_PS/$A_PS"
>
> case $# in
>   2)	;;
>   1)	# get older copy from CVS
>   	cvs up -p "$A_PS" > "$B_PS" || exit
> 	# swap A and B to have named file as B, i.e., newer copy
> 	X="$B_PS"; B_PS="$A_PS"; A_PS="$X"
> 	;;
>   *)	echo Invalid input. 1>&2
> 	Usage
>   	exit 1
> 	;;
> esac
>
> A_BASE="${A_PS//\//_}"
> B_BASE="${B_PS//\//_}"
>
>     # convert to pixmap format; use cache when available and not  
> outdated
> if [ ! -f $CACHE/"$A_BASE"-001.pgm \
>     -o "$A_PS" -nt $CACHE/"$A_BASE"-001.pgm \
>     -o -n "$FORCE" \
>    ]
> then
>     gs -dNOPAUSE -sDEVICE=pgmraw -r$RES -sOutputFile= 
> $CACHE/"$A_BASE"-%03d.pgm \
>     	"$A_PS" quit.ps  || exit
> fi
>
> if [ ! -f $CACHE/"$B_BASE"-001.pgm \
>     -o "$B_PS" -nt $CACHE/"$B_BASE"-001.pgm \
>     -o -n "$FORCE" \
>    ]
> then
>     gs -dNOPAUSE -sDEVICE=pgmraw -r$RES -sOutputFile= 
> $CACHE/"$B_BASE"-%03d.pgm \
>     	"$B_PS" quit.ps  || exit
> fi
>
>     # compare pages
> OWD=`pwd`
> cd $CACHE
> rm -f $PAIR_FILE 2> /dev/null
> for A_PGM in "$A_BASE"-${PAGES}.pgm
> do
>     SUFFIX="${A_PGM//*-/}"
>     N=${SUFFIX/.pgm/}
>
>     B_PGM="$B_BASE-${SUFFIX}"
>
>     H_DAT="$A_BASE-$B_BASE-${N}-hist.dat"
>     V="$A_BASE-$B_BASE-${N}-view.png"
>     D="$A_BASE-$B_BASE-${N}-diff.png"
>
>     if [ ! -f "$H_DAT" -o -n "$FORCE" ]; then
> 	# get histogram of diffs
> 	pnmarith -diff "$A_PGM" "$B_PGM" | tee "$D".pgm | pgmhist > "$H_DAT"
>     fi
>
>     ## Sample histogram:
>     # value   count   b%      w%
>     # -----   -----   --      --
>     # 0       484690    100%    100%
>     # 255     14        100%  0.00289%
>
>     # count non-black pixels
>     H_COUNT=`awk 'NR>3 { sum += $2} END {print 1*sum}' "$H_DAT"`
>
>     # assemble views of differing pages (only)
>     if [ $H_COUNT -ge $HIST_THRESHOLD ]; then
> 	echo $N differ 1>&2
> 	if [ ! -f "$V" -o -n "$FORCE" ]; then
> 	    rgb3toppm "$A_PGM" "$B_PGM" "$B_PGM" \
> 		| pnmtopng -transparent white -background grey50 > "$V"
> 	    pnmtopng "$D".pgm > "$D"
> 	fi
> 	echo "$V" "$A_PGM" "$B_PGM" >> $PAIR_FILE
>     fi
>     rm -f "$D".pgm 2> /dev/null
>
>     ## When memory is tight -- This renders options "-2" and "-3"  
> useless.
>     #if [ -z "$VIEWS" ]; then
>     #	rm "$A_PGM" "$B_PGM
>     #fi
> done
>
> # decide which images to view
> case $VIEWS in
>   1)  COLS=1 ;;		# diff-view only
>   2)  COLS=2-3 ;;	# page pairs only
>   3)  COLS=1-3 ;;	# all
>   *)  exit ;;
> esac
>
> # see if xargs supports the flag -r --no-run-if-empty
> xargs -r < /dev/null 2> /dev/null && XARGS_ARGS="-r"
>
> if [ -f $PAIR_FILE ]; then
>     cut -f$COLS -d' ' $PAIR_FILE | xargs $XARGS_ARGS $VIEWER
> fi
>
> # EOF
>
>
> ------------------------- Helpful Info -------------------------
> Mac-TeX Website: http://www.esm.psu.edu/mac-tex/
> TeX FAQ: http://www.tex.ac.uk/faq
> List Archive: http://tug.org/pipermail/macostex-archives/
> List Reminders & Etiquette: http://www.esm.psu.edu/mac-tex/list/
>
>
>


------------------------- Helpful Info -------------------------
Mac-TeX Website: http://www.esm.psu.edu/mac-tex/
TeX FAQ: http://www.tex.ac.uk/faq
List Archive: http://tug.org/pipermail/macostex-archives/
List Reminders & Etiquette: http://www.esm.psu.edu/mac-tex/list/





More information about the macostex-archives mailing list