pdfrs 0.1.0

/// Structured document elements parsed from Markdown.
/// These carry formatting intent so the PDF generator can render
/// headers at different sizes, indent lists, etc.

#[derive(Debug, Clone, Copy, PartialEq)]
pub enum TableAlignment {
    Left,
    Center,
    Right,
}

/// Text segment with inline formatting
#[derive(Debug, Clone, PartialEq)]
pub enum TextSegment {
    Plain(String),
    Bold(String),
    Italic(String),
    BoldItalic(String),
    Code(String),
    Link { text: String, url: String },
}

#[derive(Debug, Clone, PartialEq)]
pub enum Element {
    Heading { level: u8, text: String },
    Paragraph { text: String },
    /// Rich paragraph with multiple styled segments
    RichParagraph { segments: Vec<TextSegment> },
    UnorderedListItem { text: String, depth: u8 },
    OrderedListItem { number: u32, text: String, depth: u8 },
    TaskListItem { checked: bool, text: String },
    CodeBlock { language: String, code: String },
    InlineCode { code: String },
    TableRow { cells: Vec<String>, is_separator: bool, alignments: Vec<TableAlignment> },
    BlockQuote { text: String, depth: u8 },
    DefinitionItem { term: String, definition: String },
    Footnote { label: String, text: String },
    Link { text: String, url: String },
    Image { alt: String, path: String },
    StyledText { text: String, bold: bool, italic: bool },
    MathBlock { expression: String },
    MathInline { expression: String },
    PageBreak,
    HorizontalRule,
    EmptyLine,
}

/// Parse alignment from a table separator cell like `:---`, `:---:`, `---:`
fn parse_cell_alignment(cell: &str) -> TableAlignment {
    let t = cell.trim();
    let starts = t.starts_with(':');
    let ends = t.ends_with(':');
    if starts && ends {
        TableAlignment::Center
    } else if ends {
        TableAlignment::Right
    } else {
        TableAlignment::Left
    }
}

/// Strip inline markdown formatting from text (bold, italic, code, links, strikethrough)
pub fn strip_inline_formatting(text: &str) -> String {
    let mut s = text.to_string();

    // Strikethrough ~~text~~
    let strike_re = regex::Regex::new(r"~~(.*?)~~").unwrap();
    s = strike_re.replace_all(&s, "$1").to_string();

    // Bold+italic (***text***)
    let bold_italic_re = regex::Regex::new(r"\*\*\*(.*?)\*\*\*").unwrap();
    s = bold_italic_re.replace_all(&s, "$1").to_string();

    // Bold (**text**)
    let bold_re = regex::Regex::new(r"\*\*(.*?)\*\*").unwrap();
    s = bold_re.replace_all(&s, "$1").to_string();

    // Bold (__text__)
    let bold_re2 = regex::Regex::new(r"__(.*?)__").unwrap();
    s = bold_re2.replace_all(&s, "$1").to_string();

    // Italic (*text*)
    let italic_re = regex::Regex::new(r"\*(.*?)\*").unwrap();
    s = italic_re.replace_all(&s, "$1").to_string();

    // Italic (_text_)
    let italic_re2 = regex::Regex::new(r"_(.*?)_").unwrap();
    s = italic_re2.replace_all(&s, "$1").to_string();

    // Links [text](url)
    let link_re = regex::Regex::new(r"\[([^\]]+)\]\([^\)]+\)").unwrap();
    s = link_re.replace_all(&s, "$1").to_string();

    // Inline code `code`
    let code_re = regex::Regex::new(r"`([^`]+)`").unwrap();
    s = code_re.replace_all(&s, "$1").to_string();

    s
}

/// Parse inline markdown formatting into styled text segments
pub fn parse_inline_formatting(text: &str) -> Vec<TextSegment> {
    let mut segments = Vec::new();
    let mut remaining = text.to_string();

    // Links first (highest priority)
    let link_re = regex::Regex::new(r"\[([^\]]+)\]\(([^\)]+)\)").unwrap();
    while let Some(caps) = link_re.captures(&remaining) {
        let full_match = caps.get(0).unwrap();
        let before = &remaining[..full_match.start()];
        let link_text = caps.get(1).unwrap().as_str();
        let url = caps.get(2).unwrap().as_str();

        if !before.is_empty() {
            segments.extend(parse_formatting_no_links(before));
        }

        segments.push(TextSegment::Link {
            text: link_text.to_string(),
            url: url.to_string(),
        });
        remaining = remaining[full_match.end()..].to_string();
    }

    if !remaining.is_empty() {
        segments.extend(parse_formatting_no_links(&remaining));
    }

    segments
}

/// Parse formatting excluding links
fn parse_formatting_no_links(text: &str) -> Vec<TextSegment> {
    let mut segments = Vec::new();
    let mut remaining = text.to_string();

    // Code (high priority)
    let code_re = regex::Regex::new(r"`([^`]+)`").unwrap();
    while let Some(caps) = code_re.captures(&remaining) {
        let full_match = caps.get(0).unwrap();
        let before = &remaining[..full_match.start()];
        let code = caps.get(1).unwrap().as_str();

        if !before.is_empty() {
            segments.extend(parse_bold_italic(before));
        }

        segments.push(TextSegment::Code(code.to_string()));
        remaining = remaining[full_match.end()..].to_string();
    }

    if !remaining.is_empty() {
        segments.extend(parse_bold_italic(&remaining));
    }

    segments
}

/// Parse bold/italic formatting
fn parse_bold_italic(text: &str) -> Vec<TextSegment> {
    let mut segments = Vec::new();
    let mut remaining = text.to_string();

    loop {
        // Bold+italic: ***text*** or ___text___ (explicit patterns)
        let bi_stars_re = regex::Regex::new(r"\*\*\*(.+?)\*\*\*").unwrap();
        let bi_under_re = regex::Regex::new(r"___(.+?)___").unwrap();
        // Bold: **text** or __text__
        let b_stars_re = regex::Regex::new(r"\*\*(.+?)\*\*").unwrap();
        let b_under_re = regex::Regex::new(r"__(.+?)__").unwrap();
        // Italic: *text* or _text_ (simple pattern, may have false positives but that's acceptable)
        let i_stars_re = regex::Regex::new(r"\*([^*]+)\*").unwrap();
        let i_under_re = regex::Regex::new(r"_([^_]+)_").unwrap();

        let mut found = false;

        if let Some(caps) = bi_stars_re.captures(&remaining) {
            let full_match = caps.get(0).unwrap();
            let before = &remaining[..full_match.start()];
            let content = caps.get(1).unwrap().as_str();

            if !before.is_empty() {
                segments.push(TextSegment::Plain(before.to_string()));
            }
            segments.push(TextSegment::BoldItalic(content.to_string()));
            remaining = remaining[full_match.end()..].to_string();
            found = true;
        } else if let Some(caps) = bi_under_re.captures(&remaining) {
            let full_match = caps.get(0).unwrap();
            let before = &remaining[..full_match.start()];
            let content = caps.get(1).unwrap().as_str();

            if !before.is_empty() {
                segments.push(TextSegment::Plain(before.to_string()));
            }
            segments.push(TextSegment::BoldItalic(content.to_string()));
            remaining = remaining[full_match.end()..].to_string();
            found = true;
        } else if let Some(caps) = b_stars_re.captures(&remaining) {
            let full_match = caps.get(0).unwrap();
            let before = &remaining[..full_match.start()];
            let content = caps.get(1).unwrap().as_str();

            if !before.is_empty() {
                segments.push(TextSegment::Plain(before.to_string()));
            }
            segments.push(TextSegment::Bold(content.to_string()));
            remaining = remaining[full_match.end()..].to_string();
            found = true;
        } else if let Some(caps) = b_under_re.captures(&remaining) {
            let full_match = caps.get(0).unwrap();
            let before = &remaining[..full_match.start()];
            let content = caps.get(1).unwrap().as_str();

            if !before.is_empty() {
                segments.push(TextSegment::Plain(before.to_string()));
            }
            segments.push(TextSegment::Bold(content.to_string()));
            remaining = remaining[full_match.end()..].to_string();
            found = true;
        } else if let Some(caps) = i_stars_re.captures(&remaining) {
            let full_match = caps.get(0).unwrap();
            let before = &remaining[..full_match.start()];
            let content = caps.get(1).unwrap().as_str();

            if !before.is_empty() {
                segments.push(TextSegment::Plain(before.to_string()));
            }
            segments.push(TextSegment::Italic(content.to_string()));
            remaining = remaining[full_match.end()..].to_string();
            found = true;
        } else if let Some(caps) = i_under_re.captures(&remaining) {
            let full_match = caps.get(0).unwrap();
            let before = &remaining[..full_match.start()];
            let content = caps.get(1).unwrap().as_str();

            if !before.is_empty() {
                segments.push(TextSegment::Plain(before.to_string()));
            }
            segments.push(TextSegment::Italic(content.to_string()));
            remaining = remaining[full_match.end()..].to_string();
            found = true;
        }

        if !found {
            break;
        }
    }

    if !remaining.is_empty() {
        segments.push(TextSegment::Plain(remaining));
    }

    segments
}

/// Check if text contains any inline markdown formatting
pub fn has_inline_formatting(text: &str) -> bool {
    text.contains("**") || text.contains("__") || text.contains("***") || text.contains("___") || text.contains("`") || text.contains("[")
}

/// Parse markdown text into structured elements
pub fn parse_markdown(markdown: &str) -> Vec<Element> {
    let mut elements = Vec::new();
    let mut in_code_block = false;
    let mut code_lang = String::new();
    let mut code_buf = String::new();
    let mut in_math_block = false;
    let mut math_buf = String::new();
    let lines: Vec<&str> = markdown.lines().collect();
    let mut i = 0;

    while i < lines.len() {
        let line = lines[i];
        let trimmed = line.trim();

        // Math block toggle ($$...$$)
        if trimmed.starts_with("$$") && !in_code_block && !in_math_block {
            // Check if it's a single-line math block: $$ expr $$
            let rest = trimmed[2..].trim();
            if rest.ends_with("$$") && rest.len() > 2 {
                let expr = rest[..rest.len() - 2].trim().to_string();
                elements.push(Element::MathBlock { expression: expr });
                i += 1;
                continue;
            }
            in_math_block = true;
            math_buf.clear();
            i += 1;
            continue;
        }

        if in_math_block {
            if trimmed == "$$" {
                elements.push(Element::MathBlock { expression: math_buf.clone() });
                math_buf.clear();
                in_math_block = false;
            } else {
                if !math_buf.is_empty() {
                    math_buf.push('\n');
                }
                math_buf.push_str(trimmed);
            }
            i += 1;
            continue;
        }

        // Code block toggle
        if trimmed.starts_with("```") {
            if in_code_block {
                elements.push(Element::CodeBlock {
                    language: code_lang.clone(),
                    code: code_buf.clone(),
                });
                code_buf.clear();
                code_lang.clear();
                in_code_block = false;
            } else {
                in_code_block = true;
                code_lang = trimmed[3..].trim().to_string();
            }
            i += 1;
            continue;
        }

        if in_code_block {
            if !code_buf.is_empty() {
                code_buf.push('\n');
            }
            code_buf.push_str(line);
            i += 1;
            continue;
        }

        // Empty line
        if trimmed.is_empty() {
            elements.push(Element::EmptyLine);
            i += 1;
            continue;
        }

        // Horizontal rule
        if (trimmed == "---" || trimmed == "***" || trimmed == "___")
            && trimmed.len() >= 3
        {
            elements.push(Element::HorizontalRule);
            i += 1;
            continue;
        }

        // Headings
        if trimmed.starts_with('#') {
            let level = trimmed.chars().take_while(|&c| c == '#').count().min(6) as u8;
            let text = trimmed[level as usize..].trim().to_string();
            elements.push(Element::Heading { level, text });
            i += 1;
            continue;
        }

        // Page break: <!-- pagebreak --> or \pagebreak
        if trimmed == "<!-- pagebreak -->" || trimmed == "\\pagebreak" {
            elements.push(Element::PageBreak);
            i += 1;
            continue;
        }

        // Image: ![alt](path)
        if trimmed.starts_with("![") {
            let img_re = regex::Regex::new(r"^!\[([^\]]*)\]\(([^\)]+)\)$").unwrap();
            if let Some(caps) = img_re.captures(trimmed) {
                let alt = caps[1].to_string();
                let path = caps[2].to_string();
                elements.push(Element::Image { alt, path });
                i += 1;
                continue;
            }
        }

        // Standalone link line: [text](url) — only if the entire line is a link
        if trimmed.starts_with('[') && !trimmed.starts_with("[^") {
            let link_re = regex::Regex::new(r"^\[([^\]]+)\]\(([^\)]+)\)$").unwrap();
            if let Some(caps) = link_re.captures(trimmed) {
                let text = caps[1].to_string();
                let url = caps[2].to_string();
                elements.push(Element::Link { text, url });
                i += 1;
                continue;
            }
        }

        // Blockquote
        if trimmed.starts_with('>') {
            let mut depth: u8 = 0;
            let mut rest = trimmed;
            while rest.starts_with('>') {
                depth += 1;
                rest = rest[1..].trim_start();
            }
            let text = strip_inline_formatting(rest);
            elements.push(Element::BlockQuote { text, depth });
            i += 1;
            continue;
        }

        // Task list items: - [ ] or - [x]
        if trimmed.starts_with("- [ ] ") || trimmed.starts_with("- [x] ") || trimmed.starts_with("- [X] ") {
            let checked = !trimmed.starts_with("- [ ] ");
            let text = strip_inline_formatting(&trimmed[6..]);
            elements.push(Element::TaskListItem { checked, text });
            i += 1;
            continue;
        }

        // Table rows (contains |)
        if trimmed.starts_with('|') && trimmed.ends_with('|') {
            let inner = &trimmed[1..trimmed.len() - 1];
            let cells: Vec<String> = inner.split('|').map(|c| c.trim().to_string()).collect();
            let is_separator = cells.iter().all(|c| {
                let t = c.trim_matches(':').trim();
                !t.is_empty() && t.chars().all(|ch| ch == '-')
            });
            if is_separator {
                let alignments: Vec<TableAlignment> = cells.iter().map(|c| parse_cell_alignment(c)).collect();
                elements.push(Element::TableRow { cells, is_separator: true, alignments });
            } else {
                let cells: Vec<String> = cells.into_iter().map(|c| strip_inline_formatting(&c)).collect();
                let alignments = vec![TableAlignment::Left; cells.len()];
                elements.push(Element::TableRow { cells, is_separator: false, alignments });
            }
            i += 1;
            continue;
        }

        // Footnote definition: [^label]: text
        if trimmed.starts_with("[^") {
            if let Some(close) = trimmed.find("]:") {
                let label = trimmed[2..close].to_string();
                let text = strip_inline_formatting(trimmed[close + 2..].trim());
                elements.push(Element::Footnote { label, text });
                i += 1;
                continue;
            }
        }

        // Definition list: line starting with ": " after a paragraph
        if trimmed.starts_with(": ") {
            let definition = strip_inline_formatting(&trimmed[2..]);
            // The term is the previous paragraph element
            let term = match elements.last() {
                Some(Element::Paragraph { text }) => text.clone(),
                _ => String::new(),
            };
            // Remove the paragraph that was actually the term
            if !term.is_empty() {
                elements.pop();
            }
            elements.push(Element::DefinitionItem { term, definition });
            i += 1;
            continue;
        }

        // Unordered list items (detect indentation depth)
        if trimmed.starts_with("- ") || trimmed.starts_with("* ") {
            let indent = line.len() - line.trim_start().len();
            let depth = (indent / 2) as u8;
            let text = strip_inline_formatting(&trimmed[2..]);
            elements.push(Element::UnorderedListItem { text, depth });
            i += 1;
            continue;
        }

        // Ordered list items
        if let Some(dot_pos) = trimmed.find(". ") {
            let num_part = &trimmed[..dot_pos];
            if !num_part.is_empty() && num_part.chars().all(|c| c.is_ascii_digit()) {
                let number: u32 = num_part.parse().unwrap_or(1);
                let indent = line.len() - line.trim_start().len();
                let depth = (indent / 3) as u8;
                let text = strip_inline_formatting(&trimmed[dot_pos + 2..]);
                elements.push(Element::OrderedListItem { number, text, depth });
                i += 1;
                continue;
            }
        }

        // Inline math: line that is entirely $expression$ (single dollar)
        if trimmed.starts_with('$') && !trimmed.starts_with("$$") && trimmed.ends_with('$') && trimmed.len() > 2 {
            let expr = trimmed[1..trimmed.len() - 1].to_string();
            if !expr.is_empty() {
                elements.push(Element::MathInline { expression: expr });
                i += 1;
                continue;
            }
        }

        // Regular paragraph — also strip footnote references [^N] -> (N)
        let footnote_ref_re = regex::Regex::new(r"\[\^([^\]]+)\]").unwrap();
        let trimmed_with_refs = footnote_ref_re.replace_all(trimmed, "($1)").to_string();

        // Check for inline formatting and use RichParagraph if present
        if has_inline_formatting(&trimmed_with_refs) {
            let segments = parse_inline_formatting(&trimmed_with_refs);
            if !segments.is_empty() {
                elements.push(Element::RichParagraph { segments });
            }
        } else {
            let text = strip_inline_formatting(&trimmed_with_refs);
            if !text.is_empty() {
                elements.push(Element::Paragraph { text });
            }
        }
        i += 1;
    }

    // Close unclosed code block
    if in_code_block && !code_buf.is_empty() {
        elements.push(Element::CodeBlock {
            language: code_lang,
            code: code_buf,
        });
    }

    // Close unclosed math block
    if in_math_block && !math_buf.is_empty() {
        elements.push(Element::MathBlock { expression: math_buf });
    }

    elements
}

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

    #[test]
    fn test_parse_heading() {
        let elements = parse_markdown("# Hello\n## World");
        assert_eq!(elements.len(), 2);
        assert_eq!(elements[0], Element::Heading { level: 1, text: "Hello".into() });
        assert_eq!(elements[1], Element::Heading { level: 2, text: "World".into() });
    }

    #[test]
    fn test_parse_task_list() {
        let elements = parse_markdown("- [ ] Todo\n- [x] Done");
        assert_eq!(elements.len(), 2);
        assert_eq!(elements[0], Element::TaskListItem { checked: false, text: "Todo".into() });
        assert_eq!(elements[1], Element::TaskListItem { checked: true, text: "Done".into() });
    }

    #[test]
    fn test_parse_strikethrough() {
        assert_eq!(strip_inline_formatting("~~removed~~"), "removed");
        assert_eq!(strip_inline_formatting("keep ~~this~~ text"), "keep this text");
    }

    #[test]
    fn test_parse_blockquote() {
        let elements = parse_markdown("> quoted text\n>> nested");
        assert_eq!(elements.len(), 2);
        assert_eq!(elements[0], Element::BlockQuote { text: "quoted text".into(), depth: 1 });
        assert_eq!(elements[1], Element::BlockQuote { text: "nested".into(), depth: 2 });
    }

    #[test]
    fn test_parse_horizontal_rule() {
        let elements = parse_markdown("---");
        assert_eq!(elements.len(), 1);
        assert_eq!(elements[0], Element::HorizontalRule);
    }

    #[test]
    fn test_parse_code_block() {
        let md = "```rust\nfn main() {}\n```";
        let elements = parse_markdown(md);
        assert_eq!(elements.len(), 1);
        assert_eq!(elements[0], Element::CodeBlock {
            language: "rust".into(),
            code: "fn main() {}".into(),
        });
    }

    #[test]
    fn test_parse_table() {
        let md = "| A | B |\n|---|---|\n| 1 | 2 |";
        let elements = parse_markdown(md);
        assert_eq!(elements.len(), 3);
        match &elements[0] {
            Element::TableRow { cells, is_separator, .. } => {
                assert!(!is_separator);
                assert_eq!(cells, &["A", "B"]);
            }
            _ => panic!("Expected TableRow"),
        }
        match &elements[1] {
            Element::TableRow { is_separator, .. } => assert!(is_separator),
            _ => panic!("Expected separator"),
        }
    }

    #[test]
    fn test_parse_table_alignment() {
        let md = "| L | C | R |\n|:---|:---:|---:|\n| a | b | c |";
        let elements = parse_markdown(md);
        match &elements[1] {
            Element::TableRow { is_separator, alignments, .. } => {
                assert!(is_separator);
                assert_eq!(alignments, &[TableAlignment::Left, TableAlignment::Center, TableAlignment::Right]);
            }
            _ => panic!("Expected separator with alignments"),
        }
    }

    #[test]
    fn test_parse_definition_list() {
        let md = "Term\n: Definition text";
        let elements = parse_markdown(md);
        assert_eq!(elements.len(), 1);
        assert_eq!(elements[0], Element::DefinitionItem {
            term: "Term".into(),
            definition: "Definition text".into(),
        });
    }

    #[test]
    fn test_parse_nested_list() {
        let md = "- Top\n  - Nested\n    - Deep";
        let elements = parse_markdown(md);
        assert_eq!(elements.len(), 3);
        assert_eq!(elements[0], Element::UnorderedListItem { text: "Top".into(), depth: 0 });
        assert_eq!(elements[1], Element::UnorderedListItem { text: "Nested".into(), depth: 1 });
        assert_eq!(elements[2], Element::UnorderedListItem { text: "Deep".into(), depth: 2 });
    }

    #[test]
    fn test_strip_inline_formatting() {
        assert_eq!(strip_inline_formatting("**bold**"), "bold");
        assert_eq!(strip_inline_formatting("*italic*"), "italic");
        assert_eq!(strip_inline_formatting("`code`"), "code");
        assert_eq!(strip_inline_formatting("[link](http://x.com)"), "link");
        assert_eq!(strip_inline_formatting("***both***"), "both");
    }

    #[test]
    fn test_parse_footnote_definition() {
        let md = "[^1]: This is a footnote.";
        let elements = parse_markdown(md);
        assert_eq!(elements.len(), 1);
        assert_eq!(elements[0], Element::Footnote {
            label: "1".into(),
            text: "This is a footnote.".into(),
        });
    }

    #[test]
    fn test_parse_footnote_reference_in_paragraph() {
        let md = "Some text with a reference[^1].";
        let elements = parse_markdown(md);
        assert_eq!(elements.len(), 1);
        match &elements[0] {
            Element::Paragraph { text } => {
                assert!(text.contains("(1)"), "Footnote ref should be converted to (1), got: {}", text);
                assert!(!text.contains("[^1]"), "Raw footnote ref should be stripped");
            }
            _ => panic!("Expected Paragraph"),
        }
    }

    #[test]
    fn test_parse_footnote_named_label() {
        let md = "[^note]: A named footnote.";
        let elements = parse_markdown(md);
        assert_eq!(elements.len(), 1);
        assert_eq!(elements[0], Element::Footnote {
            label: "note".into(),
            text: "A named footnote.".into(),
        });
    }

    #[test]
    fn test_parse_image() {
        let md = "![Logo](images/logo.png)";
        let elements = parse_markdown(md);
        assert_eq!(elements.len(), 1);
        assert_eq!(elements[0], Element::Image {
            alt: "Logo".into(),
            path: "images/logo.png".into(),
        });
    }

    #[test]
    fn test_parse_image_empty_alt() {
        let md = "![](photo.jpg)";
        let elements = parse_markdown(md);
        assert_eq!(elements.len(), 1);
        assert_eq!(elements[0], Element::Image {
            alt: "".into(),
            path: "photo.jpg".into(),
        });
    }

    #[test]
    fn test_parse_standalone_link() {
        let md = "[Click here](https://example.com)";
        let elements = parse_markdown(md);
        assert_eq!(elements.len(), 1);
        assert_eq!(elements[0], Element::Link {
            text: "Click here".into(),
            url: "https://example.com".into(),
        });
    }

    #[test]
    fn test_parse_pagebreak_html() {
        let md = "<!-- pagebreak -->";
        let elements = parse_markdown(md);
        assert_eq!(elements.len(), 1);
        assert_eq!(elements[0], Element::PageBreak);
    }

    #[test]
    fn test_parse_pagebreak_latex() {
        let md = "\\pagebreak";
        let elements = parse_markdown(md);
        assert_eq!(elements.len(), 1);
        assert_eq!(elements[0], Element::PageBreak);
    }

    #[test]
    fn test_parse_mixed_new_elements() {
        let md = "# Title\n\n![img](a.png)\n\n[link](http://x.com)\n\n<!-- pagebreak -->\n\nParagraph after break.";
        let elements = parse_markdown(md);
        let types: Vec<&str> = elements.iter().map(|e| match e {
            Element::Heading { .. } => "heading",
            Element::Image { .. } => "image",
            Element::Link { .. } => "link",
            Element::PageBreak => "pagebreak",
            Element::Paragraph { .. } => "paragraph",
            Element::EmptyLine => "empty",
            _ => "other",
        }).collect();
        assert!(types.contains(&"heading"));
        assert!(types.contains(&"image"));
        assert!(types.contains(&"link"));
        assert!(types.contains(&"pagebreak"));
        assert!(types.contains(&"paragraph"));
    }
}

#[cfg(test)]
mod proptest_tests {
    use super::*;
    use proptest::prelude::*;

    proptest! {
        #[test]
        fn strip_inline_formatting_plain_text_idempotent(s in "[a-zA-Z0-9 ]{0,500}") {
            // For plain text (no markdown chars), stripping should be idempotent
            let stripped_once = strip_inline_formatting(&s);
            let stripped_twice = strip_inline_formatting(&stripped_once);
            prop_assert_eq!(stripped_once, stripped_twice);
        }

        #[test]
        fn strip_inline_formatting_removes_formatting(s in "\\PC{0,500}") {
            // Stripping should reduce or keep same length
            let original = s.len();
            let stripped = strip_inline_formatting(&s).len();
            prop_assert!(stripped <= original, "Stripping should not increase length");
        }

        #[test]
        fn strip_inline_formatting_doesnt_crash(s in "\\PC{0,500}") {
            // Just ensure we don't panic on any input
            let _ = strip_inline_formatting(&s);
        }

        #[test]
        fn heading_levels_valid(heading in "#[ \t]{0,10}[a-zA-Z0-9 ]{0,100}") {
            let elements = parse_markdown(&heading);
            for elem in elements {
                match elem {
                    Element::Heading { level, .. } => {
                        prop_assert!(level >= 1 && level <= 6, "Heading level must be 1-6");
                    }
                    _ => {}
                }
            }
        }

        #[test]
        fn list_depths_non_negative(list in "-([ \t]{0,10}[a-zA-Z0-9 ]{1,50}){0,5}") {
            let elements = parse_markdown(&list);
            for elem in elements {
                match elem {
                    Element::UnorderedListItem { depth, .. } |
                    Element::OrderedListItem { depth, .. } => {
                        prop_assert!(depth <= 10, "List depth should be reasonable");
                    }
                    _ => {}
                }
            }
        }

        #[test]
        fn empty_markdown_yields_empty(input in "\\PC*") {
            let truncated = if input.len() > 1000 { &input[..1000] } else { &input };
            let trimmed = truncated.trim();
            if trimmed.is_empty() {
                let elements = parse_markdown(trimmed);
                prop_assert_eq!(elements.len(), 0);
            }
        }

        #[test]
        fn double_newline_creates_empty_line(text in "[a-zA-Z0-9 ]{1,50}") {
            let md = format!("{}\n\n{}", text, text);
            let elements = parse_markdown(&md);
            // Should have at least one EmptyLine between paragraphs
            prop_assert!(elements.iter().any(|e| matches!(e, Element::EmptyLine)));
        }

        #[test]
        fn parse_then_strip_does_not_crash(s in "\\PC{0,1000}") {
            let _ = parse_markdown(&s);
            let _ = strip_inline_formatting(&s);
            // Just ensure we don't panic
        }

        #[test]
        fn footnote_definition_has_label(md in "\\[\\^[a-zA-Z0-9_]{1,20}\\]:[ \t]{0,5}[a-zA-Z0-9 ]{0,100}") {
            let elements = parse_markdown(&md);
            if !elements.is_empty() {
                match &elements[0] {
                    Element::Footnote { label, .. } => {
                        prop_assert!(!label.is_empty());
                        prop_assert!(label.len() <= 20);
                    }
                    _ => {}
                }
            }
        }
    }
}