用于拆解和组合PDF中各个对象的shell脚本

拆解脚本

header_start=0
header_len=15

xref_start=$(strings -a -t d $1 | grep -e "\bxref\b" | awk '{print $1}')
trailer_start=$(strings -a -t d $1 | grep -e "\btrailer\b" | awk '{print $1}')
#echo $xref_start
#echo $trailer_start

xref_len=$(echo "$trailer_start - $xref_start" | bc)
#echo $xref_len

header_dump=$(echo "$1" | sed -re 's/^(.*)\.pdf/tdis\_\1\_header\.bin/g')
dd if=$1 of=$header_dump bs=1 skip=$header_start count=$header_len

xref_dump=$(echo "$1" | sed -re 's/^(.*)\.pdf/tdis\_\1\_xref\.bin/g')
dd if=$1 of=$xref_dump bs=1 skip=$xref_start count=$xref_len

trailer_dump=$(echo "$1" | sed -re 's/^(.*)\.pdf/tdis\_\1\_trailer\.bin/g')
dd if=$1 of=$trailer_dump bs=1 skip=$trailer_start

#cat tdis_"$1"_xref.bin | awk 'NF==3' | awk 'NR!=1{printf("%d 0 obj is at offset: %d\n", NR-1, $1);}'
cat $xref_dump | awk 'NF==3' | awk 'NR!=1{printf("%08d %08d\n", $1, NR-1);}' | sort > tdis_"$xref_dump"
#echo "$xref_start 0" >> tdis_"$xref_dump"
printf "%08d %08d\n" $xref_start 0 >> tdis_"$xref_dump"

cat tdis_$xref_dump | awk 'BEGIN{loffset=0;lobjnum=0;}{printf("%3d %3d %3d\n", loffset, $1-loffset, lobjnum);loffset=$1;lobjnum=$2;}' | awk 'NR!=1' > tdis_metrics_"$xref_dump"

if [ ! -d objects ]
then
    mkdir objects
fi
cat tdis_metrics_"$xref_dump" | while read offset len objn
do
#echo $offset, $len, $objn
obj_name=$(printf "%s_%03d" $1 $objn | sed -re 's/^(.*)\.pdf/tdis\_\1\_obj/g' | awk '{printf("objects/%s.bin", $0);}')
#echo $obj_name
dd if=$1 of=$obj_name bs=1 skip=$offset count=$len
done

　　

组合脚本

target=$1
dd if=$(ls -1 | grep "header.bin") of=$target bs=1 count=15

obj_offset=15
obj_nums=0
for file in $(ls -1 objects)
do
    #echo $file
    obj_len=$(wc objects/$file | awk '{print $3}')
    dd if=objects/$file of=$target bs=1 count=$obj_len seek=$obj_offset
    printf "%010d %05d n\n" $obj_offset 0 >> "tas_generated_"$1"_xref.bin"
    obj_offset=$[ $obj_offset + $obj_len ]
    obj_nums=$[ $obj_nums + 1 ]
done
echo "xref" >> $target
printf "0 %d\n" $obj_nums >> $target
echo "0000000000 65535 f" >> $target
cat "tas_generated_"$1"_xref.bin" >> $target

awk 'NR<=2' $(ls -1 | grep "trailer.bin") >> $target
echo "startxref" >> $target
echo $obj_offset >> $target
echo "%%EOF" >> $target

这样，我们就可以对解析出来的单个pdf对象进行单独操作了。

手动找出包含graphic operators stream的对象，使用下面脚本解压stream

target=$(ls -1 objects | grep "_obj_"$1".bin")
grep -Ubo --binary-file=text stream objects/$target | sed -e 's/:/ /g' | awk 'NR==1{printf("%d ",$1+7);}NR==2{printf("%d ", $1-10);}' > tdeflate_stream.bin
read xstart xend < tdeflate_stream.bin
dd if=objects/$target of=flated.bin bs=1 skip=$xstart count=$[ $xend - $xstart ]
cat flated.bin | zlib-flate -uncompress > deflated.bin

重新编辑deflated.bin文件，再使用下面脚本压缩

printf "%d 0 obj\n" $1 > tflate_"$1".bin
printf "<</Length %d/Filter/FlateDecode>>stream\n" >> tflate_"$1".bin
cat deflated.bin | zlib-flate -compress >> tflate_"$1".bin
echo "" >> tflate_"$1".bin
echo "endstream" >> tflate_"$1".bin
echo "endobj" >> tflate_"$1".bin

target=$(ls -1 objects | grep "_obj_"$1".bin")
rm objects/$target
mv tflate_"$1".bin objects/$target