Well, essentially yes, it is based on text analysis, but it's a lot more complex than "a script that calls subs2srs". (:
(My whole codebase is over 100k lines of code, all written by me.)
I have a unified morphological analysis engine that I use for every type of media that I have in my database, and I use it to generate stats and vocabulary lists.
Sorry, I might have gotten carried away with all of the fancy words. (: The "morphological" part comes from linguistics (see the "Morphology (linguistics)" article on wikipedia), by "unified" I meant that I apply it to every kind of texts, and an "engine" is just another word for "software" (think "game engine" - a reusable piece of software that can be used for many things in a single domain).
(My whole codebase is over 100k lines of code, all written by me.)
I have a unified morphological analysis engine that I use for every type of media that I have in my database, and I use it to generate stats and vocabulary lists.