thanks for sharing this project. I am the creator of this tool. I do it on free time many years ago and I forget it.
But it had a good performand and saved a lot of time. So I decided to resurrect it.
To answer @davidsrsb, it is based on text boxes in pdf parsing, using poppler. So it does not work for old scanned datashet with images. It is very fast : about 6s for a 700 pages datasheet with 7 footprints.
For simple testing, I will propose builds for linux and windows.
Do not hesitate to send me ideas or bugs.