its not svg, but seems to be very close to it.
I still dont understand the encoding. deleting parts does not really seem to work, though I wonder if I could do it.
using pdf-parser by dider stevens from stackoverflow
python '.\format troubleshooting\pdf-parser.py' -o 4 -f -d '.\format troubleshooting\obj4.dump' '.\original files\WCM0001.pdf'
output to obj4.dump
https://github.com/itext/rups/releases/tag/5.5.9
the software also does some nice stuff