ocr 0.1.1

//! OCR output format support
//!
//! This module implements various OCR output formats compatible with Tesseract:
//! - Plain text
//! - JSON (structured output with bounding boxes)
//! - hOCR (HTML OCR format)
//! - TSV (character-level output)

use crate::core::text::{BoundingBox, TextResult};
use crate::utils::Result;
use chrono::Utc;
use serde::{Deserialize, Serialize};
use std::collections::HashMap;

/// Output format enumeration
#[derive(Debug, Clone, Copy, PartialEq, Eq)]
pub enum OutputFormat {
    /// Plain text output
    PlainText,
    /// JSON structured output
    Json,
    /// hOCR (HTML with OCR data)
    HOcr,
    /// TSV (tab-separated values)
    Tsv,
    /// ALTO XML (standard library format)
    Alto,
    /// Box file for Tesseract training
    Box,
    /// Searchable PDF with invisible text layer
    Pdf,
    /// Hierarchical structured JSON with document elements
    StructuredJson,
    /// Markdown output with heading/paragraph/list structure
    Markdown,
}

/// JSON output structure
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct OcrJsonOutput {
    /// Full text content
    pub text: String,
    /// Overall confidence score (0.0 to 1.0)
    pub confidence: f32,
    /// Detected language
    pub language: Option<String>,
    /// Processing time in milliseconds
    pub processing_time_ms: Option<u64>,
    /// Image dimensions
    pub image_size: ImageSize,
    /// Lines of text
    pub lines: Vec<JsonLine>,
    /// Engine metadata
    pub engine_info: EngineInfo,
    /// Timestamp
    pub timestamp: String,
}

/// Image size information
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct ImageSize {
    pub width: u32,
    pub height: u32,
    pub dpi: u32,
}

/// Line in JSON output
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct JsonLine {
    /// Line text
    pub text: String,
    /// Line confidence
    pub confidence: f32,
    /// Bounding box
    pub bounding_box: BoundingBox,
    /// Line number (0-indexed)
    pub line_number: usize,
    /// Words in this line
    pub words: Vec<JsonWord>,
}

/// Word in JSON output
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct JsonWord {
    /// Word text
    pub text: String,
    /// Word confidence
    pub confidence: f32,
    /// Bounding box
    pub bounding_box: BoundingBox,
    /// Word number in line (0-indexed)
    pub word_number: usize,
    /// Characters in this word
    pub characters: Vec<JsonCharacter>,
}

/// Character in JSON output
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct JsonCharacter {
    /// Character
    pub text: char,
    /// Character confidence
    pub confidence: f32,
    /// Bounding box
    pub bounding_box: BoundingBox,
}

/// Engine information
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct EngineInfo {
    /// Engine name
    pub name: String,
    /// Engine version
    pub version: String,
    /// Model type used
    pub model_type: Option<String>,
}

/// Convert TextResult to JSON output format
pub fn to_json_output(result: &TextResult) -> OcrJsonOutput {
    let image_size = ImageSize {
        width: result.bounding_box.width(),
        height: result.bounding_box.height(),
        dpi: 300, // Default DPI if not available
    };

    let lines: Vec<JsonLine> = result
        .lines
        .iter()
        .enumerate()
        .map(|(line_num, line)| JsonLine {
            text: line.text.clone(),
            confidence: line.confidence,
            bounding_box: line.bounding_box,
            line_number: line_num,
            words: line
                .words
                .iter()
                .enumerate()
                .map(|(word_num, word)| JsonWord {
                    text: word.text.clone(),
                    confidence: word.confidence,
                    bounding_box: word.bounding_box,
                    word_number: word_num,
                    characters: word
                        .characters
                        .iter()
                        .map(|ch| JsonCharacter {
                            text: ch.character,
                            confidence: ch.confidence,
                            bounding_box: ch.bounding_box,
                        })
                        .collect(),
                })
                .collect(),
        })
        .collect();

    OcrJsonOutput {
        text: result.text.clone(),
        confidence: result.confidence,
        language: result.language.clone(),
        processing_time_ms: None,
        image_size,
        lines,
        engine_info: EngineInfo {
            name: "OCR".to_string(),
            version: env!("CARGO_PKG_VERSION").to_string(),
            model_type: None,
        },
        timestamp: Utc::now().to_rfc3339(),
    }
}

/// Generate JSON output string
pub fn format_json(result: &TextResult) -> Result<String> {
    let json_output = to_json_output(result);
    serde_json::to_string_pretty(&json_output)
        .map_err(|e| crate::OcrError::Internal(format!("JSON serialization error: {}", e)))
        .into()
}

/// Generate hOCR (HTML OCR format) output
///
/// hOCR is an HTML-based format for representing OCR output with spatial information.
/// It follows the hOCR 4.1 specification from http://kba.cloud/hocr-spec/
pub fn format_hocr(result: &TextResult) -> Result<String> {
    let mut html = String::new();

    // HTML header
    html.push_str("<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n");
    html.push_str("<!DOCTYPE html>\n");
    html.push_str("<html xmlns=\"http://www.w3.org/1999/xhtml\" xml:lang=\"en\" lang=\"en\">\n");
    html.push_str("<head>\n");
    html.push_str("  <title>OCR Output</title>\n");
    html.push_str("  <meta http-equiv=\"Content-Type\" content=\"text/html;charset=utf-8\" />\n");
    html.push_str("  <meta name='ocr-system' content='OCR ");
    html.push_str(env!("CARGO_PKG_VERSION"));
    html.push_str("' />\n");
    html.push_str("  <meta name='ocr-capabilities' content='ocr_page ocr_carea ocr_par ocr_line ocrx_word ocr_cinfo' />\n");
    html.push_str("</head>\n");
    html.push_str("<body>\n");

    // OCR page
    let bbox = &result.bounding_box;
    html.push_str(&format!(
        "  <div class='ocr_page' id='page_1' title=\"image \"\"; bbox {} {} {} {}; ppageno 0\">\n",
        bbox.left, bbox.top, bbox.right, bbox.bottom
    ));

    // Content area (single carea for now)
    html.push_str(&format!(
        "    <div class='ocr_carea' id='block_1_1' title=\"bbox {} {} {} {}\">\n",
        bbox.left, bbox.top, bbox.right, bbox.bottom
    ));

    // Paragraph
    html.push_str(&format!(
        "      <p class='ocr_par' id='par_1_1' title=\"bbox {} {} {} {}\">\n",
        bbox.left, bbox.top, bbox.right, bbox.bottom
    ));

    // Lines
    for (line_idx, line) in result.lines.iter().enumerate() {
        let line_bbox = &line.bounding_box;
        html.push_str(&format!(
            "        <span class='ocr_line' id='line_1_{}' title=\"bbox {} {} {} {}; baseline {} 0; x_size {}; x_descenders {}; x_ascenders {}\">\n",
            line_idx + 1,
            line_bbox.left,
            line_bbox.top,
            line_bbox.right,
            line_bbox.bottom,
            line_bbox.bottom,
            line_bbox.height(),
            line_bbox.height() / 4,
            line_bbox.height() / 4
        ));

        // Words
        for (word_idx, word) in line.words.iter().enumerate() {
            let word_bbox = &word.bounding_box;
            html.push_str(&format!(
                "          <span class='ocrx_word' id='word_1_{}_{}' title=\"bbox {} {} {} {}; x_wconf {:.0}\">",
                line_idx + 1,
                word_idx + 1,
                word_bbox.left,
                word_bbox.top,
                word_bbox.right,
                word_bbox.bottom,
                word.confidence * 100.0
            ));

            html.push_str(&xml_escape(&word.text));

            html.push_str("</span>\n");
        }

        html.push_str("        </span>\n");
    }

    html.push_str("      </p>\n");
    html.push_str("    </div>\n");
    html.push_str("  </div>\n");
    html.push_str("</body>\n");
    html.push_str("</html>\n");

    Ok(html)
}

/// Generate TSV (tab-separated values) output
///
/// Matches Tesseract's TSV column format:
/// level page_num block_num par_num line_num word_num left top width height conf text
pub fn format_tsv(result: &TextResult) -> Result<String> {
    let mut output = String::new();

    // Header
    output.push_str("level\tpage_num\tblock_num\tpar_num\tline_num\tword_num\tleft\ttop\twidth\theight\tconf\ttext\n");

    let page_num = 1;
    let block_num = 1;
    let par_num = 1;

    // Page level
    let pb = &result.bounding_box;
    output.push_str(&format!(
        "1\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{:.2}\t\n",
        page_num,
        block_num,
        par_num,
        0,
        0,
        pb.left,
        pb.top,
        pb.width(),
        pb.height(),
        result.confidence * 100.0,
    ));

    // Block level
    output.push_str(&format!(
        "2\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{:.2}\t\n",
        page_num,
        block_num,
        par_num,
        0,
        0,
        pb.left,
        pb.top,
        pb.width(),
        pb.height(),
        result.confidence * 100.0,
    ));

    // Paragraph level
    output.push_str(&format!(
        "3\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{:.2}\t\n",
        page_num,
        block_num,
        par_num,
        0,
        0,
        pb.left,
        pb.top,
        pb.width(),
        pb.height(),
        result.confidence * 100.0,
    ));

    for (line_idx, line) in result.lines.iter().enumerate() {
        let line_num = line_idx + 1;
        let lb = &line.bounding_box;

        // Line level
        output.push_str(&format!(
            "4\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{:.2}\t\n",
            page_num,
            block_num,
            par_num,
            line_num,
            0,
            lb.left,
            lb.top,
            lb.width(),
            lb.height(),
            line.confidence * 100.0,
        ));

        for (word_idx, word) in line.words.iter().enumerate() {
            let word_num = word_idx + 1;
            let wb = &word.bounding_box;
            let conf = word.confidence * 100.0;

            // Word level
            output.push_str(&format!(
                "5\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{}\t{:.2}\t{}\n",
                page_num,
                block_num,
                par_num,
                line_num,
                word_num,
                wb.left,
                wb.top,
                wb.width(),
                wb.height(),
                conf,
                escape_tsv_text(&word.text),
            ));
        }
    }

    Ok(output)
}

/// Escape text for TSV output (tab and newline)
fn escape_tsv_text(s: &str) -> String {
    s.replace('\t', " ").replace('\n', " ")
}

/// Generate ALTO XML output
pub fn format_alto(result: &TextResult) -> Result<String> {
    let mut xml = String::new();
    let bbox = &result.bounding_box;

    xml.push_str("<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n");
    xml.push_str("<alto xmlns=\"http://www.loc.gov/standards/alto/ns-v4#\" xmlns:xsi=\"http://www.w3.org/2001/XMLSchema-instance\" xsi:schemaLocation=\"http://www.loc.gov/standards/alto/ns-v4# http://www.loc.gov/standards/alto/v4/alto-4-4.xsd\" VERSION=\"4.4\">\n");
    xml.push_str("  <Description>\n");
    xml.push_str("    <MeasurementUnit>pixel</MeasurementUnit>\n");
    xml.push_str("    <OCRProcessing>\n");
    xml.push_str("      <ocrProcessingStep>\n");
    xml.push_str("        <processingSoftware>\n");
    xml.push_str(&format!("          <softwareName>OCR</softwareName>\n"));
    xml.push_str(&format!(
        "          <softwareVersion>{}</softwareVersion>\n",
        env!("CARGO_PKG_VERSION")
    ));
    xml.push_str("        </processingSoftware>\n");
    xml.push_str("      </ocrProcessingStep>\n");
    xml.push_str("    </OCRProcessing>\n");
    xml.push_str("  </Description>\n");

    xml.push_str(&format!(
        "  <Layout>\n    <Page WIDTH=\"{}\" HEIGHT=\"{}\" PHYSICAL_IMG_NR=\"1\" ID=\"page_1\">\n",
        bbox.right, bbox.bottom
    ));
    xml.push_str(&format!(
        "      <PrintSpace HPOS=\"0\" VPOS=\"0\" WIDTH=\"{}\" HEIGHT=\"{}\">\n",
        bbox.right, bbox.bottom
    ));

    for (line_idx, line) in result.lines.iter().enumerate() {
        let lb = &line.bounding_box;
        xml.push_str(&format!(
            "        <TextLine ID=\"line_{}\" HPOS=\"{}\" VPOS=\"{}\" WIDTH=\"{}\" HEIGHT=\"{}\">\n",
            line_idx + 1,
            lb.left,
            lb.top,
            lb.width(),
            lb.height()
        ));

        for (word_idx, word) in line.words.iter().enumerate() {
            let wb = &word.bounding_box;
            let wc = (word.confidence * 100.0).round() as u32;
            xml.push_str(&format!(
                "          <String ID=\"word_{}_{}\" HPOS=\"{}\" VPOS=\"{}\" WIDTH=\"{}\" HEIGHT=\"{}\" CONTENT=\"{}\" WC=\"0.{}\"/>\n",
                line_idx + 1,
                word_idx + 1,
                wb.left,
                wb.top,
                wb.width(),
                wb.height(),
                xml_escape(&word.text),
                wc
            ));
        }

        xml.push_str("        </TextLine>\n");
    }

    xml.push_str("      </PrintSpace>\n");
    xml.push_str("    </Page>\n");
    xml.push_str("  </Layout>\n");
    xml.push_str("</alto>\n");

    Ok(xml)
}

/// Generate box file output for Tesseract training
///
/// Format per character: char left top right bottom page
/// Spaces are represented as a special box with page -1
pub fn format_box(result: &TextResult) -> Result<String> {
    let mut output = String::new();
    let page = 0;

    for line in &result.lines {
        for word in &line.words {
            for ch in &word.characters {
                let bbox = &ch.bounding_box;
                output.push_str(&format!(
                    "{} {} {} {} {} {}\n",
                    ch.character, bbox.left, bbox.top, bbox.right, bbox.bottom, page
                ));
            }
            // Add space between words
            if !word.characters.is_empty() {
                let last_ch = word.characters.last().unwrap();
                let space_left = last_ch.bounding_box.right;
                let space_right = space_left + 5;
                output.push_str(&format!(
                    "{} {} {} {} {} {}\n",
                    ' ',
                    space_left,
                    last_ch.bounding_box.top,
                    space_right,
                    last_ch.bounding_box.bottom,
                    page
                ));
            }
        }
        // Add newline between lines
        if !line.words.is_empty() {
            let last_word = line.words.last().unwrap();
            if !last_word.characters.is_empty() {
                let last_ch = last_word.characters.last().unwrap();
                output.push_str(&format!(
                    "\t {} {} {} {}\n",
                    last_ch.bounding_box.right + 5,
                    last_ch.bounding_box.top,
                    last_ch.bounding_box.right + 10,
                    last_ch.bounding_box.bottom,
                ));
            }
        }
    }

    Ok(output)
}

/// Escape text for XML attribute
fn xml_escape(s: &str) -> String {
    s.replace('&', "&amp;")
        .replace('<', "&lt;")
        .replace('>', "&gt;")
        .replace('"', "&quot;")
}

/// Escape a single character for HTML content
fn html_escape_char(c: char) -> String {
    match c {
        '<' => "&lt;".to_string(),
        '>' => "&gt;".to_string(),
        '&' => "&amp;".to_string(),
        '"' => "&quot;".to_string(),
        '\'' => "&#x27;".to_string(),
        _ => c.to_string(),
    }
}

/// Document element type for structured output
#[derive(Debug, Clone, Serialize, Deserialize)]
pub enum DocElement {
    Heading { level: u8, text: String },
    Paragraph { text: String },
    ListItem { text: String, ordered: bool, number: Option<usize> },
    Table { rows: Vec<Vec<String>> },
    Figure { caption: String },
    CodeBlock { text: String },
}

/// Structured document output
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct StructuredDocument {
    pub title: Option<String>,
    pub elements: Vec<DocElement>,
    pub metadata: HashMap<String, String>,
}

impl StructuredDocument {
    pub fn from_text(text: &str) -> Self {
        let mut elements = Vec::new();
        let mut current_paragraph = String::new();

        for line in text.lines() {
            let trimmed = line.trim();
            if trimmed.is_empty() {
                if !current_paragraph.is_empty() {
                    elements.push(DocElement::Paragraph {
                        text: current_paragraph.trim().to_string(),
                    });
                    current_paragraph.clear();
                }
                continue;
            }

            // Check for list items
            if let Some(list_item) = Self::parse_list_item(trimmed) {
                if !current_paragraph.is_empty() {
                    elements.push(DocElement::Paragraph {
                        text: current_paragraph.trim().to_string(),
                    });
                    current_paragraph.clear();
                }
                elements.push(list_item);
                continue;
            }

            // Check for heading (simple heuristic: short line, all caps, or ends without period)
            if Self::is_heading(trimmed, &current_paragraph) {
                if !current_paragraph.is_empty() {
                    elements.push(DocElement::Paragraph {
                        text: current_paragraph.trim().to_string(),
                    });
                    current_paragraph.clear();
                }
                let level = Self::heading_level(trimmed);
                elements.push(DocElement::Heading {
                    level,
                    text: trimmed.to_string(),
                });
                continue;
            }

            if !current_paragraph.is_empty() {
                current_paragraph.push(' ');
            }
            current_paragraph.push_str(trimmed);
        }

        if !current_paragraph.is_empty() {
            elements.push(DocElement::Paragraph {
                text: current_paragraph.trim().to_string(),
            });
        }

        StructuredDocument {
            title: None,
            elements,
            metadata: HashMap::new(),
        }
    }

    fn parse_list_item(text: &str) -> Option<DocElement> {
        let trimmed = text.trim_start();
        for prefix in ["• ", "- ", "* "] {
            if let Some(rest) = trimmed.strip_prefix(prefix) {
                return Some(DocElement::ListItem {
                    text: rest.to_string(),
                    ordered: false,
                    number: None,
                });
            }
        }
        // Numbered list: "1. text" or "1) text"
        let mut chars = trimmed.chars();
        let first = chars.next()?;
        if first.is_numeric() {
            let mut num_str = first.to_string();
            for ch in chars {
                if ch.is_numeric() {
                    num_str.push(ch);
                } else if ch == '.' || ch == ')' {
                    let rest = trimmed[num_str.len() + 1..].trim_start().to_string();
                    return Some(DocElement::ListItem {
                        text: rest,
                        ordered: true,
                        number: num_str.parse().ok(),
                    });
                } else {
                    break;
                }
            }
        }
        None
    }

    fn is_heading(text: &str, prev_context: &str) -> bool {
        if text.len() < 3 || text.len() > 120 {
            return false;
        }
        // All caps heading
        if text.chars().all(|c| c.is_uppercase() || c.is_numeric() || c.is_whitespace() || c == '-' || c == '_')
            && text.chars().any(|c| c.is_alphabetic())
            && text.len() < 80
        {
            return true;
        }
        // Short line without ending punctuation (likely heading)
        let words = text.split_whitespace().count();
        if words <= 8 && !text.ends_with('.') && !text.ends_with(',') && !text.ends_with(';') {
            // If previous context exists, this is likely a new section
            if !prev_context.is_empty() && text.len() > 3 {
                return true;
            }
        }
        false
    }

    fn heading_level(text: &str) -> u8 {
        if text.chars().all(|c| c.is_uppercase() || c.is_numeric() || c.is_whitespace() || c == '-' || c == '_') {
            1
        } else if text.len() < 40 {
            2
        } else {
            3
        }
    }

    pub fn to_markdown(&self) -> String {
        let mut md = String::new();
        for element in &self.elements {
            match element {
                DocElement::Heading { level, text } => {
                    let hashes = "#".repeat(*level as usize);
                    md.push_str(&format!("{} {}\n\n", hashes, text));
                }
                DocElement::Paragraph { text } => {
                    md.push_str(&format!("{}\n\n", text));
                }
                DocElement::ListItem { text, ordered: false, .. } => {
                    md.push_str(&format!("- {}\n", text));
                }
                DocElement::ListItem { text, ordered: true, number: Some(n) } => {
                    md.push_str(&format!("{}. {}\n", n, text));
                }
                DocElement::ListItem { text, ordered: true, number: None } => {
                    md.push_str(&format!("1. {}\n", text));
                }
                DocElement::Table { rows } => {
                    if !rows.is_empty() {
                        for (i, row) in rows.iter().enumerate() {
                            let cells: Vec<String> = row.iter().map(|c| format!("| {} ", c)).collect();
                            md.push_str(&cells.join(""));
                            md.push_str("|\n");
                            if i == 0 {
                                let sep = "|---".repeat(row.len());
                                md.push_str(&format!("{}|\n", sep));
                            }
                        }
                        md.push('\n');
                    }
                }
                DocElement::Figure { caption } => {
                    md.push_str(&format!("*Figure: {}*\n\n", caption));
                }
                DocElement::CodeBlock { text } => {
                    md.push_str(&format!("```\n{}\n```\n\n", text));
                }
            }
        }
        md
    }
}

/// Generate structured JSON output with document elements
pub fn format_structured_json(result: &TextResult) -> Result<String> {
    let doc = StructuredDocument::from_text(&result.text);
    serde_json::to_string_pretty(&doc)
        .map_err(|e| crate::OcrError::Internal(format!("JSON serialization error: {}", e)).into())
}

/// Generate Markdown output with document structure
pub fn format_markdown(result: &TextResult) -> Result<String> {
    let doc = StructuredDocument::from_text(&result.text);
    Ok(doc.to_markdown())
}

/// Format output in the specified format
pub fn format_output(result: &TextResult, format: OutputFormat) -> Result<String> {
    match format {
        OutputFormat::PlainText => Ok(result.text.clone()),
        OutputFormat::Json => format_json(result),
        OutputFormat::HOcr => format_hocr(result),
        OutputFormat::Tsv => format_tsv(result),
        OutputFormat::Alto => format_alto(result),
        OutputFormat::Box => format_box(result),
        OutputFormat::Pdf => Err(crate::OcrError::Internal(
            "PDF output requires image path. Use format_pdf_with_image instead.".to_string(),
        )
        .into()),
        OutputFormat::StructuredJson => format_structured_json(result),
        OutputFormat::Markdown => format_markdown(result),
    }
}

/// Parse output format from string
pub fn parse_output_format(s: &str) -> Result<OutputFormat> {
    match s.to_lowercase().as_str() {
        "text" | "txt" | "plain" => Ok(OutputFormat::PlainText),
        "json" => Ok(OutputFormat::Json),
        "hocr" | "html" => Ok(OutputFormat::HOcr),
        "tsv" => Ok(OutputFormat::Tsv),
        "alto" | "xml" => Ok(OutputFormat::Alto),
        "box" => Ok(OutputFormat::Box),
        "pdf" => Ok(OutputFormat::Pdf),
        "md" | "markdown" => Ok(OutputFormat::Markdown),
        "structured" | "structured-json" | "doc" => Ok(OutputFormat::StructuredJson),
        _ => Err(crate::OcrError::InvalidInput(format!(
            "Unknown output format: {}. Supported: text, json, hocr, tsv, alto, box, pdf, markdown, structured-json",
            s
        ))
        .into()),
    }
}

#[cfg(test)]
mod tests {
    use super::*;
    use crate::core::text::{BoundingBox, CharacterResult, LineResult, TextResult, WordResult};

    fn create_test_result() -> TextResult {
        let mut result = TextResult::new(
            "Hello World".to_string(),
            0.9,
            BoundingBox::new(0, 0, 200, 50),
        );

        // Add a line
        let mut line = LineResult::new(
            "Hello World".to_string(),
            0.9,
            BoundingBox::new(0, 0, 200, 20),
        );

        // Add words
        let mut word1 = WordResult::new("Hello".to_string(), 0.95, BoundingBox::new(0, 0, 60, 20));
        let mut word2 =
            WordResult::new("World".to_string(), 0.85, BoundingBox::new(70, 0, 130, 20));

        // Add characters
        for (i, ch) in "Hello".chars().enumerate() {
            let bbox = BoundingBox::new(i as u32 * 10, 0, (i + 1) as u32 * 10, 20);
            word1.characters.push(CharacterResult::new(ch, 0.95, bbox));
        }

        for (i, ch) in "World".chars().enumerate() {
            let bbox = BoundingBox::new(70 + i as u32 * 10, 0, 70 + (i + 1) as u32 * 10, 20);
            word2.characters.push(CharacterResult::new(ch, 0.85, bbox));
        }

        line.words.push(word1);
        line.words.push(word2);
        result.lines.push(line);

        result
    }

    #[test]
    fn test_format_json() {
        let result = create_test_result();
        let json = format_json(&result).unwrap();

        assert!(json.contains("Hello World"));
        assert!(json.contains("\"confidence\": 0.9"));
        assert!(json.contains("Hello"));
        assert!(json.contains("World"));
    }

    #[test]
    fn test_format_hocr() {
        let result = create_test_result();
        let hocr = format_hocr(&result).unwrap();

        assert!(hocr.contains("<html"));
        assert!(hocr.contains("ocr_page"));
        assert!(hocr.contains("ocr_line"));
        assert!(hocr.contains("Hello"));
        assert!(hocr.contains("World"));
    }

    #[test]
    fn test_format_tsv() {
        let result = create_test_result();
        let tsv = format_tsv(&result).unwrap();

        // Should have tab-separated values
        assert!(tsv.contains('\t'));
        assert!(tsv.chars().filter(|&c| c == '\n').count() > 0);
    }

    #[test]
    fn test_parse_output_format() {
        assert!(matches!(
            parse_output_format("json"),
            Ok(OutputFormat::Json)
        ));
        assert!(matches!(
            parse_output_format("TEXT"),
            Ok(OutputFormat::PlainText)
        ));
        assert!(matches!(
            parse_output_format("hocr"),
            Ok(OutputFormat::HOcr)
        ));
        assert!(parse_output_format("invalid").is_err());
    }

    #[test]
    fn test_format_output_plain() {
        let result = create_test_result();
        let output = format_output(&result, OutputFormat::PlainText).unwrap();
        assert_eq!(output, "Hello World");
    }

    #[test]
    fn test_format_alto() {
        let result = create_test_result();
        let alto = format_alto(&result).unwrap();

        assert!(alto.contains("<?xml version=\"1.0\""));
        assert!(alto.contains("<alto xmlns=\"http://www.loc.gov/standards/alto/ns-v4#\""));
        assert!(alto.contains("<String"));
        assert!(alto.contains("CONTENT=\"Hello\""));
        assert!(alto.contains("CONTENT=\"World\""));
        assert!(alto.contains("</alto>"));
    }

    #[test]
    fn test_xml_escape() {
        assert_eq!(xml_escape("A & B"), "A &amp; B");
        assert_eq!(xml_escape("<tag>"), "&lt;tag&gt;");
        assert_eq!(xml_escape("say \"hi\""), "say &quot;hi&quot;");
    }

    #[test]
    fn test_parse_alto_format() {
        assert!(matches!(
            parse_output_format("alto"),
            Ok(OutputFormat::Alto)
        ));
        assert!(matches!(parse_output_format("xml"), Ok(OutputFormat::Alto)));
    }

    #[test]
    fn test_parse_markdown_format() {
        assert!(matches!(
            parse_output_format("markdown"),
            Ok(OutputFormat::Markdown)
        ));
        assert!(matches!(parse_output_format("md"), Ok(OutputFormat::Markdown)));
    }

    #[test]
    fn test_parse_structured_json_format() {
        assert!(matches!(
            parse_output_format("structured-json"),
            Ok(OutputFormat::StructuredJson)
        ));
        assert!(matches!(parse_output_format("doc"), Ok(OutputFormat::StructuredJson)));
    }

    #[test]
    fn test_structured_document_from_text() {
        let text = "INTRODUCTION\n\nThis is a paragraph with multiple sentences. It continues on.\n\n• First bullet\n• Second bullet\n\n1. Ordered item\n2. Another ordered\n\nConclusion text here.";
        let doc = StructuredDocument::from_text(text);
        assert!(!doc.elements.is_empty());
        assert!(matches!(doc.elements[0], DocElement::Heading { level: 1, .. }));
    }

    #[test]
    fn test_markdown_output() {
        let text = "INTRODUCTION\n\nThis is a paragraph.\n\n• Bullet item\n\n1. Numbered item\n\nSUBSECTION\n\nMore text.";
        let doc = StructuredDocument::from_text(text);
        let md = doc.to_markdown();
        assert!(md.contains("# INTRODUCTION"));
        assert!(md.contains("- Bullet item"));
        assert!(md.contains("1. Numbered item"));
        assert!(md.contains("# SUBSECTION"));
    }

    #[test]
    fn test_format_markdown_output() {
        let mut result = create_test_result();
        result.text = "HEADING\n\nBody text here.".to_string();
        let md = format_markdown(&result).unwrap();
        assert!(md.contains("# HEADING"));
        assert!(md.contains("Body text here."));
    }

    #[test]
    fn test_format_structured_json_output() {
        let mut result = create_test_result();
        result.text = "INTRO\n\nParagraph.".to_string();
        let json = format_structured_json(&result).unwrap();
        assert!(json.contains("\"elements\""));
        assert!(json.contains("\"Heading\""));
    }
}