Same here. Diffs published from the Journal Officiel aren't so easy to parse.
I am afraid it could be manually generated, which is a dead end.
I do agree that the technical details would be interesting.
It would make the author a _very_ efficient worker.