//! Cursor position and document statistics — pure helpers over the text.

/// (line, column) of `cursor` (byte offset), both 0-based; column counts
/// characters of the line, not bytes.
#[must_use]
pub fn line_col(text: &str, cursor: usize) -> (usize, usize) {
    let floored = if text.is_char_boundary(cursor) {
        usize::min(cursor, text.len())
    } else {
        let mut p = cursor;
        while p > 0 && !text.is_char_boundary(p) {
            p -= 1;
        }
        p
    };
    let before = text.get(..floored).unwrap_or("");
    let line = before.matches('\n').count();
    let col = before.chars().rev().take_while(|c| *c != '\n').count();
    (line, col)
}

/// Words in the text: runs of characters that are alphanumeric or `_`.
#[must_use]
pub fn word_count(text: &str) -> usize {
    let mut count = 0usize;
    let mut in_word = false;
    for c in text.chars() {
        if c.is_alphanumeric() || c == '_' {
            if !in_word {
                count += 1;
            }
            in_word = true;
        } else {
            in_word = false;
        }
    }
    count
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn line_col_counts_chars() {
        let text = "ab é\nxyz";
        let byte_x = text.find('x').unwrap_or(0);
        assert_eq!(line_col(text, byte_x), (1, 0));
        assert_eq!(line_col(text, 0), (0, 0));
        let end = text.len();
        assert_eq!(line_col(text, end), (1, 3));
    }

    #[test]
    fn line_col_at_char_start() {
        let mut text = String::from("éx");
        let byte_after_e = text.find('x').unwrap_or(0);
        assert_eq!(line_col(&text, byte_after_e), (0, 1));
        text.push_str("\nq");
        let byte_q = text.find('q').unwrap_or(0);
        assert_eq!(line_col(&text, byte_q), (1, 0));
    }

    #[test]
    fn word_count_runs() {
        assert_eq!(word_count(""), 0);
        assert_eq!(word_count("hello world"), 2);
        assert_eq!(word_count("éx_ one-two #tag"), 4);
        assert_eq!(word_count("...,,, "), 0);
    }
}