Find code blocks in markdown
This script was borne out of a discussion about adding syntax highlighting to code snippets embedded inside of markdown.
Because code snippets are, by their very nature, probably not compilable, this script takes a fairly brute-force approach in determining what language a code snippet is most likely to be.
Currently, the script iterates line-by-line tokenizing input and comparing against programming-language keywords. When keywords are discovered, they are scored. The score is used to determine the likelihood that a code snippet is one language vs another.