Yes, as Stefan has pointed out, making your own lexer/parser is not likely to be the best way to approach this project (and perhaps we should make that more explicit).
Your best bet here is to have a look at the existing Julia parser - get a feel for things, then think about improvements you could make. I can't give you a specific warm-up task, but making some improvement to the existing code would be a great start. Hope this helps!
