We use the Pygments libraries to detect lexable text and report it as the programming language. But Pygments can lex many other types beyond programming langauges (e.g. INI files, etc).
We should instead:
- return the Pygments lexer name as a new, extra filetype attribute
- use a stricter, subset list of such types as real, actual programming languages AND confront that with the actual file extensions for improved accuracy.