Pre-process emoji.

[SamCB]

A problem we have is that any time the library is loaded, it goes through a process of reading the emoji-test.txt file and loading from there. On most systems, this is very quick. I chose this because it meant future updates to the emoji spec would simply require replacing the emoji-test.txt file with whatever the most up-to-date version was.

However, on slower systems... such as those which actually have use for this library... this actually takes quite a while to do. As such, we should create a script that pre-processes the emoji-test.txt file, crease some kind of serializable object out of it, and include that object in the final library. I expect this will speed up the loading process phenomanly. As a bonus, we won't need to package the emoji-test.txt file, which will save ~400kb.