wyrm | Autodifferentiation package in Rust | Machine Learning library

Basically you can think of boxes. Let's say we want to count boxes.

1. You have a big box Mobs. You have one box count = 1 Start of your recursion.

2. You open the box. You see three small boxes (NPC, Enemies, Heroes). for i in node[1]

3. You take the first box (NPC) and ask how many boxes do I have in NPC. count = count + how_many(i).

4. You have 1 Box (NPC) count = 1

5. You open the box and see a smaller box (Andrew) for i in node[1].

6. You take the box again and ask how many boxes do I have in Andrew. count = count + how_many(i)

7. You have 1 Box (Andrew). count = 1

8. You open the box and don't see any smaller box. You don't enter the for loop for i in node[1]

9. You return count which is 1.

10 You come back to the Box before (NPC) you get 1 back from how_many(i) (from Andrew) and have 1 form NPC. 1 + 1 = 2

11 There are no other boxes in NPC therefore you exit the for loop and return count which is 2.

1. You are back in your first box (Mob). You get 2 back and have 1 from the Mob box. 1 + 2 = 3.

2. You are still in the for loop and check the next box (Enemies).

3. You do the same in short: Enemies will return 5 (Enemies, Slime, Goblin, Dragon, Wyrm). You count 3 + 5 = 8.

4. The last box is still there (Heroes)

5. Do the same in short: Heroes will return 2 (Heroes, Human) and you count 8 + 2 = 10.

6. The last return is count = 10

result = 10