Memory efficient trie with incremental garbage collection.

Initialize a triehugger and add some words to it:

triehugger = Triehugger()
triehugger.add("beets")
triehugger.add("beetle")
triehugger.add("beetlejuice")

Your triehugger looks like:

-
 - b
  - be
   - bee
    - beet
     - beets
     - beetl
      - beetle
       - beetlej
        - beetleju
         - beetlejui
          - beetlejuic
           - beetlejuice

Now let's delete beetlejuice:

triehugger.delete("beetlejuice")

It now looks like:

-
 - b
  - be
   - bee
    - beet
     - beets
     - beetl
      - beetle

Note that when you deleted the word beetlejuice, the nodes beetlej, beetleju, beetlejui, beetlejuic were also removed, since they only existed to direct you to beetlejuice. Once beetlejuice is gone, they have no reason to remain! However, beetle remains because it was a word we explicitly added and one we want to store.