pub fn estimate_tokens(text: &str) -> usize
Estimate token count from string (rough approximation)
Simple estimation: 1 token ≈ 4 characters. Returns a minimum of 1 for non-empty strings.