vtcode-tui 0.98.2

//! URL and path-preserving text wrapping.
//!
//! Wraps text while keeping URLs and file paths as atomic units to preserve
//! terminal link detection and transcript file hit-testing.

use ratatui::prelude::*;
use ratatui::widgets::Paragraph;
use regex::Regex;
use std::sync::LazyLock;
use unicode_width::UnicodeWidthStr;

/// URL/file token detection pattern - matches common URL formats and path-like tokens.
static PRESERVED_TOKEN_PATTERN: LazyLock<Regex> = LazyLock::new(|| {
    Regex::new(
        r#"[a-zA-Z][a-zA-Z0-9+.-]*://[^\s<>\[\]{}|^]+|[a-zA-Z0-9][-a-zA-Z0-9]*\.[a-zA-Z]{2,}(/[^\s<>\[\]{}|^]*)?|localhost:\d+|\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}(:\d+)?|`(?:file://|~/|/|\./|\.\./|[A-Za-z]:[\\/]|[A-Za-z0-9._-]+[\\/])[^`]+`|"(?:file://|~/|/|\./|\.\./|[A-Za-z]:[\\/]|[A-Za-z0-9._-]+[\\/])[^"]+"|'(?:file://|~/|/|\./|\.\./|[A-Za-z]:[\\/]|[A-Za-z0-9._-]+[\\/])[^']+'|(?:\./|\../|~/|/)[^\s<>\[\]{}|^]+|(?:[A-Za-z]:[\\/][^\s<>\[\]{}|^]+)|(?:[A-Za-z0-9._-]+[\\/][^\s<>\[\]{}|^]+)"#,
    )
    .unwrap()
});

/// Check if text contains a preserved URL/path token.
pub fn contains_preserved_token(text: &str) -> bool {
    PRESERVED_TOKEN_PATTERN.is_match(text)
}

/// Wrap a line, preserving URLs as atomic units.
///
/// - Lines without URLs: delegated to standard wrapping
/// - URL-only lines: returned unwrapped if they fit
/// - Mixed lines: URLs kept intact, surrounding text wrapped normally
pub fn wrap_line_preserving_urls(line: Line<'static>, max_width: usize) -> Vec<Line<'static>> {
    if max_width == 0 {
        return vec![Line::default()];
    }

    let text: String = line.spans.iter().map(|s| s.content.as_ref()).collect();

    // No URLs - use standard wrapping (delegates to text_utils)
    if !contains_preserved_token(&text) {
        return super::text_utils::wrap_line(line, max_width);
    }

    // Find all preserved tokens in the text
    let urls: Vec<_> = PRESERVED_TOKEN_PATTERN
        .find_iter(&text)
        .map(|m| (m.start(), m.end(), m.as_str()))
        .collect();

    // Single URL that fits - return unwrapped for terminal link detection
    if urls.len() == 1 && urls[0].0 == 0 && urls[0].1 == text.len() && text.width() <= max_width {
        return vec![line];
    }
    // URL too wide - fall through to wrap it

    // Mixed content - split around URLs and wrap each segment
    wrap_mixed_content(line, &text, max_width, &urls)
}

/// Wrap text that contains URLs, keeping URLs intact.
fn wrap_mixed_content(
    line: Line<'static>,
    text: &str,
    max_width: usize,
    urls: &[(usize, usize, &str)],
) -> Vec<Line<'static>> {
    use unicode_segmentation::UnicodeSegmentation;
    use unicode_width::UnicodeWidthStr;

    let mut result = Vec::with_capacity(urls.len() + 1);
    let mut current_line: Vec<Span<'static>> = Vec::new();
    let mut current_width = 0usize;
    let mut text_pos = 0usize;

    fn trim_trailing_wrap_whitespace(spans: &mut Vec<Span<'static>>) {
        while let Some(last) = spans.last_mut() {
            let trimmed_len = last.content.trim_end_matches(char::is_whitespace).len();
            if trimmed_len == last.content.len() {
                break;
            }
            if trimmed_len == 0 {
                spans.pop();
                continue;
            }
            last.content.to_mut().truncate(trimmed_len);
            break;
        }
    }

    let flush_line = |spans: &mut Vec<Span<'static>>, result: &mut Vec<Line<'static>>| {
        if spans.is_empty() {
            result.push(Line::default());
        } else {
            trim_trailing_wrap_whitespace(spans);
            result.push(Line::from(std::mem::take(spans)));
        }
    };

    // Merge spans into a single style for simplicity when dealing with URLs
    let default_style = line.spans.first().map(|s| s.style).unwrap_or_default();

    let push_wrapped_token = |token: &str,
                              current_line: &mut Vec<Span<'static>>,
                              current_width: &mut usize,
                              result: &mut Vec<Line<'static>>| {
        for grapheme in UnicodeSegmentation::graphemes(token, true) {
            let grapheme_width = grapheme.width();
            if grapheme_width == 0 {
                current_line.push(Span::styled(grapheme.to_string(), default_style));
                continue;
            }
            if *current_width + grapheme_width > max_width && *current_width > 0 {
                flush_line(current_line, result);
                *current_width = 0;
            }
            current_line.push(Span::styled(grapheme.to_string(), default_style));
            *current_width += grapheme_width;
        }
    };

    let push_wrapped_text = |segment: &str,
                             current_line: &mut Vec<Span<'static>>,
                             current_width: &mut usize,
                             result: &mut Vec<Line<'static>>| {
        for piece in segment.split_inclusive('\n') {
            let mut text = piece;
            let mut had_newline = false;
            if let Some(stripped) = text.strip_suffix('\n') {
                text = stripped;
                had_newline = true;
                if let Some(without_carriage) = text.strip_suffix('\r') {
                    text = without_carriage;
                }
            }

            for token in UnicodeSegmentation::split_word_bounds(text) {
                if token.is_empty() {
                    continue;
                }

                let token_width = token.width();
                if token_width == 0 {
                    current_line.push(Span::styled(token.to_string(), default_style));
                    continue;
                }

                let token_is_whitespace = token.chars().all(char::is_whitespace);
                let has_content = *current_width > 0;

                if token_is_whitespace && !result.is_empty() && !has_content {
                    continue;
                }

                if *current_width + token_width <= max_width {
                    current_line.push(Span::styled(token.to_string(), default_style));
                    *current_width += token_width;
                    continue;
                }

                if token_is_whitespace {
                    if has_content {
                        flush_line(current_line, result);
                        *current_width = 0;
                    }
                    continue;
                }

                if token_width <= max_width {
                    if has_content {
                        flush_line(current_line, result);
                        *current_width = 0;
                    }
                    current_line.push(Span::styled(token.to_string(), default_style));
                    *current_width += token_width;
                    continue;
                }

                push_wrapped_token(token, current_line, current_width, result);
            }

            if had_newline {
                flush_line(current_line, result);
                *current_width = 0;
            }
        }
    };

    for (url_start, url_end, url_text) in urls {
        // Process text before this URL
        if *url_start > text_pos {
            push_wrapped_text(
                &text[text_pos..*url_start],
                &mut current_line,
                &mut current_width,
                &mut result,
            );
        }

        // Add URL — keep atomic if it fits, otherwise break it across lines
        let url_width = url_text.width();
        if url_width <= max_width {
            if current_width > 0 && current_width + url_width > max_width {
                flush_line(&mut current_line, &mut result);
                current_width = 0;
            }
            current_line.push(Span::styled(url_text.to_string(), default_style));
            current_width += url_width;
        } else {
            // URL is wider than max_width — break it grapheme-by-grapheme
            if current_width > 0 {
                flush_line(&mut current_line, &mut result);
                current_width = 0;
            }
            push_wrapped_token(url_text, &mut current_line, &mut current_width, &mut result);
        }

        text_pos = *url_end;
    }

    // Process remaining text after last URL
    if text_pos < text.len() {
        push_wrapped_text(
            &text[text_pos..],
            &mut current_line,
            &mut current_width,
            &mut result,
        );
    }

    flush_line(&mut current_line, &mut result);
    if result.is_empty() {
        result.push(Line::default());
    }
    result
}

/// Wrap multiple lines with URL preservation.
pub fn wrap_lines_preserving_urls(
    lines: Vec<Line<'static>>,
    max_width: usize,
) -> Vec<Line<'static>> {
    if max_width == 0 {
        return vec![Line::default()];
    }
    lines
        .into_iter()
        .flat_map(|line| wrap_line_preserving_urls(line, max_width))
        .collect()
}

/// Calculate wrapped height using Paragraph::line_count.
pub fn calculate_wrapped_height(text: &str, width: u16) -> usize {
    if width == 0 {
        return text.lines().count().max(1);
    }
    Paragraph::new(text).line_count(width)
}

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

    #[test]
    fn test_url_detection() {
        assert!(contains_preserved_token("https://example.com"));
        assert!(contains_preserved_token("example.com/path"));
        assert!(contains_preserved_token("localhost:8080"));
        assert!(contains_preserved_token("192.168.1.1:8080"));
        assert!(!contains_preserved_token("not a url"));
    }

    #[test]
    fn test_file_path_detection() {
        assert!(contains_preserved_token("src/main.rs"));
        assert!(contains_preserved_token("./src/main.rs"));
        assert!(contains_preserved_token("/tmp/example.txt"));
        assert!(contains_preserved_token("\"./docs/My Notes.md\""));
        assert!(contains_preserved_token(
            "`/Users/example/Library/Application Support/Code/User/settings.json`"
        ));
    }

    #[test]
    fn test_url_only_preserved() {
        let line = Line::from(Span::raw("https://example.com"));
        let wrapped = wrap_line_preserving_urls(line, 80);
        assert_eq!(wrapped.len(), 1);
        assert!(
            wrapped[0]
                .spans
                .iter()
                .any(|s| s.content.contains("https://"))
        );
    }

    #[test]
    fn test_mixed_content() {
        let line = Line::from(Span::raw("See https://example.com for info"));
        let wrapped = wrap_line_preserving_urls(line, 25);
        let all_text: String = wrapped
            .iter()
            .flat_map(|l| l.spans.iter().map(|s| s.content.as_ref()))
            .collect();
        assert!(all_text.contains("https://example.com"));
        assert!(all_text.contains("See"));
    }

    #[test]
    fn test_quoted_path_with_spaces_is_preserved() {
        let line = Line::from(Span::raw("Open \"./docs/My Notes.md\" for details"));
        let wrapped = wrap_line_preserving_urls(line, 18);
        let all_text: String = wrapped
            .iter()
            .flat_map(|line| line.spans.iter().map(|span| span.content.as_ref()))
            .collect();

        assert!(all_text.contains("\"./docs/My Notes.md\""));
    }

    #[test]
    fn test_long_url_breaks_across_lines() {
        let long_url = "https://auth.openai.com/oauth/authorize?response_type=code&client_id=app_EMoamEEZ73f0CkXaXp7hrann&redirect_uri=http%3A%2F%2Flocalhost%3A1455%2Fauth%2Fcallback&scope=openid";
        let line = Line::from(Span::raw(long_url.to_string()));
        let wrapped = wrap_line_preserving_urls(line, 80);
        assert!(
            wrapped.len() > 1,
            "Long URL should wrap across multiple lines"
        );
        let all_text: String = wrapped
            .iter()
            .flat_map(|l| l.spans.iter().map(|s| s.content.as_ref()))
            .collect();
        assert_eq!(all_text, long_url, "All characters should be preserved");
    }

    #[test]
    fn test_no_url_delegates() {
        let line = Line::from(Span::raw("Regular text without URLs"));
        let wrapped = wrap_line_preserving_urls(line, 10);
        assert!(!wrapped.is_empty());
    }

    #[test]
    fn test_mixed_content_prefers_word_boundaries_around_urls() {
        let line = Line::from(Span::raw("alpha https://x.io beta gamma"));
        let wrapped = wrap_line_preserving_urls(line, 12);
        let rendered: Vec<String> = wrapped
            .iter()
            .map(|line| {
                line.spans
                    .iter()
                    .map(|span| span.content.as_ref())
                    .collect()
            })
            .collect();

        assert_eq!(
            rendered,
            vec![
                "alpha".to_string(),
                "https://x.io".to_string(),
                "beta gamma".to_string()
            ]
        );
    }
}