Bad (?) news.
After spending entire two days getting all the code to Orbis (finally done it), I think I will just start again and do not use better AI.
There are several reasons. I can be wrong of course.
First, stability. I am not sure how is the current state of FfH, but I get many bugs that are hard to nail down
Second, performance. Most of the things are done in python and I am sure it does affect performance
Third, difficulty - too much for me to digest...
Fourth, elegance & customization - python again, every single unit is listed there. So for Orbis I would have to review value of every unit and personally write code
I am going to start again and probably use some things I like. Also, I am going to think of some simpler and more flexible methods of improving AI.
But I doubt I will ever include all (or even half) of Wild Mana changes
Too much trouble for too little improvement in AI performance.
python always slows the game more than C++ when a component becomes stable and works you may want to move it to C++ to speed up the game, callbacks hurt more then other python!
my 2¢