microformats 0.18.2

//! Picture parsing module for microformats.
//!
//! This module handles parsing of picture-related elements and properties
//! in microformats documents.

use crate::parse::{Error, element::Node};
use microformats_types::{Image, PropertyValue, UrlValue};
use serde::{Deserialize, Serialize};
use std::collections::BTreeMap;
use url::Url;

/// A single source within a picture element with srcset and media query support.
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
#[serde(rename_all = "kebab-case")]
pub struct PictureSource {
    /// The srcset URLs with their descriptors (width, height, density).
    pub srcset: Vec<SrcSetEntry>,

    /// The media query condition for this source.
    #[serde(default, skip_serializing_if = "String::is_empty")]
    pub media: String,

    /// The MIME type of the image format.
    #[serde(default, skip_serializing_if = "String::is_empty")]
    pub r#type: String,

    /// Responsive sizing hints.
    #[serde(default, skip_serializing_if = "String::is_empty")]
    pub sizes: String,

    /// Additional attributes for this source.
    #[serde(default, skip_serializing_if = "BTreeMap::is_empty")]
    pub attributes: BTreeMap<String, String>,
}

/// A single entry in a srcset with URL and descriptor information.
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
#[serde(rename_all = "kebab-case")]
pub struct SrcSetEntry {
    /// The image URL.
    pub url: Url,

    /// Width descriptor (if applicable).
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub width: Option<u32>,

    /// Height descriptor (if applicable).
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub height: Option<u32>,

    /// Density descriptor (e.g., "2x" for retina displays).
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub density: Option<String>,
}

/// A picture element containing multiple sources and a fallback image.
#[derive(Debug, Clone, PartialEq, Eq, Serialize, Deserialize)]
#[serde(rename_all = "kebab-case")]
pub struct Picture {
    /// The alternative image sources for the picture.
    #[serde(default, skip_serializing_if = "Vec::is_empty")]
    pub sources: Vec<PictureSource>,

    /// The fallback image (typically an img element).
    #[serde(default, skip_serializing_if = "Option::is_none")]
    pub fallback: Option<Image>,

    /// Additional attributes found on the picture element.
    #[serde(default, skip_serializing_if = "BTreeMap::is_empty")]
    pub attributes: BTreeMap<String, String>,
}

/// A builder for picture parsing configuration.
#[derive(Debug, Clone)]
pub struct PictureParserBuilder {
    /// Whether to extract alt text from images
    extract_alt_text: bool,
    /// Whether to follow redirects for image URLs
    follow_redirects: bool,
    /// Maximum file size for images (in bytes)
    max_file_size: Option<u64>,
}

impl PictureParserBuilder {
    /// Enable extraction of alt text from images.
    pub fn with_alt_text_extraction(mut self, enable: bool) -> Self {
        self.extract_alt_text = enable;
        self
    }

    /// Enable following redirects for image URLs.
    pub fn with_redirect_following(mut self, enable: bool) -> Self {
        self.follow_redirects = enable;
        self
    }

    /// Set maximum file size limit for images.
    pub fn with_max_file_size(mut self, size: u64) -> Self {
        self.max_file_size = Some(size);
        self
    }
}

/// Picture parser for extracting image-related properties from HTML elements.
#[derive(Debug, Clone)]
pub struct PictureParser {
    config: PictureParserBuilder,
}

impl Default for PictureParser {
    fn default() -> Self {
        Self {
            config: PictureParserBuilder::default(),
        }
    }
}

impl PictureParser {
    /// Create a new picture parser with default configuration.
    pub fn new() -> Self {
        Self::default()
    }

    /// Create a picture parser with custom configuration.
    pub fn with_config(config: PictureParserBuilder) -> Self {
        Self { config }
    }

    /// Parse a picture element and extract image information.
    ///
    /// # Errors
    ///
    /// Returns an error if the picture element cannot be processed.
    pub fn parse_picture_element(
        &self,
        node: &Node,
        base_url: &Url,
    ) -> Result<Option<Picture>, Error> {
        self.parse_picture_element_with_sources(node, base_url)
    }

    /// Parse an image element (img tag or similar).
    ///
    /// # Errors
    ///
    /// Returns an error if the image element cannot be processed.
    pub fn parse_image_element(
        &self,
        node: &Node,
        base_url: &Url,
    ) -> Result<Option<PropertyValue>, Error> {
        // Placeholder implementation - will be implemented in future tasks
        let tag_name = node.tag();

        match tag_name {
            "img" => self.parse_img_element(node, base_url),
            "picture" => {
                if let Some(picture) = self.parse_picture_element_with_sources(node, base_url)? {
                    if let Some(fallback) = picture.fallback {
                        return Ok(Some(PropertyValue::Image(fallback)));
                    }
                }
                Ok(None)
            }
            "video" => self.parse_video_element(node, base_url),
            _ => Ok(None),
        }
    }

    /// Parse an img element specifically.
    fn parse_img_element(
        &self,
        node: &Node,
        base_url: &Url,
    ) -> Result<Option<PropertyValue>, Error> {
        let src_attr_value = node.attr("src").filter(|s| !s.is_empty());
        let alt_attr_value = node.attr("alt");

        let src_url = match src_attr_value
            .as_ref()
            .and_then(|attr_value| base_url.join(attr_value).ok())
        {
            Some(url) => url,
            None => {
                tracing::debug!(
                    "Failed to resolve URL from src attribute '{}' with base_url '{}'",
                    src_attr_value.unwrap_or_default(),
                    base_url
                );
                return Ok(None);
            }
        };

        // TODO: Implement file size checking when fetching images
        if let Some(ref max_size) = self.config.max_file_size {
            tracing::debug!(
                "File size limit of {} bytes set, but checking not yet implemented",
                max_size
            );
        }

        if alt_attr_value.is_some() {
            Ok(Some(PropertyValue::Image(Image {
                value: src_url,
                alt: alt_attr_value,
            })))
        } else {
            Ok(Some(PropertyValue::Url(UrlValue::new(src_url))))
        }
    }

    /// Parse a picture element with multiple source elements.
    fn parse_picture_element_with_sources(
        &self,
        node: &Node,
        base_url: &Url,
    ) -> Result<Option<Picture>, Error> {
        // Only process picture elements
        if node.tag() != "picture" {
            return Ok(None);
        }

        let mut sources = Vec::new();
        let mut fallback = None;
        let mut attributes = BTreeMap::new();

        // Collect all attributes from the picture element
        attributes.extend(node.full_attribute_map());

        // Iterate through child elements to find source and img tags
        for child_element in node.elements() {
            match child_element.tag() {
                "source" => {
                    // Parse source element attributes
                    let srcset_attr = child_element.attr("srcset").filter(|s| !s.is_empty());
                    let media_attr = child_element.attr("media").unwrap_or_default();
                    let type_attr = child_element.attr("type").unwrap_or_default();
                    let sizes_attr = child_element.attr("sizes").unwrap_or_default();

                    // Only create a source if we have a srcset
                    if let Some(srcset) = srcset_attr {
                        let srcset_entries = self.parse_srcset_with_base(&srcset, base_url)?;

                        let source = PictureSource {
                            srcset: srcset_entries,
                            media: media_attr,
                            r#type: type_attr,
                            sizes: sizes_attr,
                            attributes: child_element.full_attribute_map(),
                        };

                        sources.push(source);
                    }
                }
                "img" => {
                    // Found the fallback img element
                    if let Some(img_property) = self.parse_img_element(&child_element, base_url)? {
                        match img_property {
                            PropertyValue::Image(image) => {
                                fallback = Some(image);
                            }
                            PropertyValue::Url(url_value) => {
                                // Convert URL to Image with no alt text for consistency
                                fallback = Some(Image {
                                    value: (*url_value).clone(),
                                    alt: None,
                                });
                            }
                            _ => {
                                // Ignore other property value types
                            }
                        }
                    }
                }
                _ => {
                    // Ignore other element types within picture
                    continue;
                }
            }
        }

        // Create the Picture struct
        let picture = Picture {
            sources,
            fallback,
            attributes,
        };

        // Always return Some(Picture) even if empty - this represents a valid picture element
        Ok(Some(picture))
    }

    /// Parse a video element that might contain a poster image.
    fn parse_video_element(
        &self,
        _node: &Node,
        _base_url: &Url,
    ) -> Result<Option<PropertyValue>, Error> {
        // Check for poster attribute first
        if let Some(poster_url) = _node.attr("poster").filter(|s| !s.is_empty()) {
            if let Ok(resolved_url) = _base_url.join(&poster_url) {
                return Ok(Some(PropertyValue::Url(UrlValue::new(resolved_url))));
            }
        }

        // Fallback: look for img elements within video
        // This would require access to children, which is not currently in Node API

        Ok(None)
    }

    /// Extract image dimensions from an element.
    ///
    /// # Returns
    ///
    /// Returns (width, height) tuple if dimensions are available.
    pub fn extract_dimensions(&self, node: &Node) -> Option<(u32, u32)> {
        let width = node.attr("width").and_then(|w| w.parse().ok());
        let height = node.attr("height").and_then(|h| h.parse().ok());

        match (width, height) {
            (Some(w), Some(h)) => Some((w, h)),
            _ => None,
        }
    }

    /// Check if an image element represents a representative image.
    ///
    /// Representative images are typically used to identify or represent
    /// a person, organization, or other entity.
    pub fn is_representative_image(&self, node: &Node) -> bool {
        // TODO: Implement more sophisticated representative image detection
        // based on microformats specification

        // For now, check for common patterns:
        // - Avatar-related class names
        // - Profile or logo-related attributes
        let class = node.attr("class").unwrap_or_default();
        let id = node.attr("id").unwrap_or_default();

        let representative_indicators =
            ["avatar", "profile", "logo", "photo", "picture", "portrait"];

        representative_indicators
            .iter()
            .any(|indicator| class.contains(indicator) || id.contains(indicator))
    }

    /// Validate that an image URL is accessible and valid.
    ///
    /// # Errors
    ///
    /// Returns an error if the URL is invalid or fails validation.
    pub async fn validate_image_url(&self, url: &Url) -> Result<bool, Error> {
        // Placeholder for future implementation
        // This would involve making HTTP requests to validate image accessibility
        tracing::debug!("Image URL validation not yet implemented for: {}", url);
        Ok(true)
    }

    /// Extract alt text from an image element with fallback logic.
    pub fn extract_alt_text(&self, node: &Node) -> Option<String> {
        // First priority: explicit alt attribute
        if let Some(alt) = node.attr("alt") {
            if !alt.is_empty() {
                return Some(alt);
            }
        }

        // TODO: Implement fallback strategies:
        // - aria-label attribute
        // - title attribute
        // - Contextual text near the image
        // - Filename analysis

        None
    }

    /// Get the configuration used by this parser.
    pub fn config(&self) -> &PictureParserBuilder {
        &self.config
    }

    /// Parse a srcset attribute value according to HTML5 specification.
    ///
    /// This method parses srcset attribute strings and extracts source set entries
    /// with their URLs and descriptors (width, height, density).
    ///
    /// # Errors
    ///
    /// Returns an error if the srcset string cannot be parsed.
    pub fn parse_srcset(&self, srcset: &str) -> Result<Vec<SrcSetEntry>, Error> {
        // Use a dummy base URL that allows relative URLs to be parsed as-is
        self.parse_srcset_with_base(srcset, &Url::parse("http://dummy-base.com/").unwrap())
    }

    /// Parse a srcset attribute value with a base URL for resolving relative URLs.
    ///
    /// # Errors
    ///
    /// Returns an error if the srcset string cannot be parsed.
    pub fn parse_srcset_with_base(
        &self,
        srcset: &str,
        base_url: &Url,
    ) -> Result<Vec<SrcSetEntry>, Error> {
        if srcset.trim().is_empty() {
            return Ok(Vec::new());
        }

        let mut entries = Vec::new();
        let mut current_entry = String::new();
        let mut in_quotes = false;
        let mut quote_char = '\0';

        for c in srcset.chars() {
            match c {
                '"' | '\'' => {
                    if in_quotes && c == quote_char {
                        in_quotes = false;
                        quote_char = '\0';
                    } else if !in_quotes {
                        in_quotes = true;
                        quote_char = c;
                    }
                    current_entry.push(c);
                }
                ',' => {
                    if !in_quotes {
                        if !current_entry.trim().is_empty() {
                            let entry = self.parse_srcset_entry(current_entry.trim(), base_url)?;
                            entries.push(entry);
                        }
                        current_entry.clear();
                    } else {
                        current_entry.push(c);
                    }
                }
                _ => {
                    current_entry.push(c);
                }
            }
        }

        // Handle the last entry
        if !current_entry.trim().is_empty() {
            let entry = self.parse_srcset_entry(current_entry.trim(), base_url)?;
            entries.push(entry);
        }

        Ok(entries)
    }

    /// Parse a single srcset entry containing URL and descriptors.
    ///
    /// A srcset entry has the format: URL [descriptors]
    /// where descriptors can include width (w), height (h), and density (x) values.
    /// URLs can contain spaces if they're quoted or URL-encoded.
    fn parse_srcset_entry(&self, entry: &str, base_url: &Url) -> Result<SrcSetEntry, Error> {
        if entry.trim().is_empty() {
            return Err(Error::HtmlCodegen(format!(
                "Empty srcset entry: '{}'",
                entry
            )));
        }

        // Parse by finding the last contiguous sequence of valid descriptors at the end
        let trimmed_entry = entry.trim();
        let parts: Vec<&str> = trimmed_entry.split_whitespace().collect();

        if parts.is_empty() {
            return Err(Error::HtmlCodegen(format!(
                "Empty srcset entry: '{}'",
                entry
            )));
        }

        // Find the last valid descriptor by working backwards
        let mut last_valid_descriptor_index = None;
        for i in (1..parts.len()).rev() {
            if self.is_valid_descriptor(parts[i]) {
                last_valid_descriptor_index = Some(i);
                break;
            }
        }

        let (url_part, descriptor_part);
        if let Some(valid_idx) = last_valid_descriptor_index {
            // We found at least one valid descriptor
            // Now work backwards to find where the contiguous sequence of valid descriptors begins
            let mut descriptor_start = valid_idx;

            // Work backwards to find the start of valid descriptors
            for i in (1..valid_idx).rev() {
                if self.is_valid_descriptor(parts[i]) {
                    descriptor_start = i;
                } else {
                    // Found non-descriptor, stop
                    break;
                }
            }

            // Only include contiguous valid descriptors
            let url_parts = &parts[0..descriptor_start];
            let descriptor_parts = &parts[descriptor_start..];

            url_part = url_parts.join(" ");
            descriptor_part = descriptor_parts.join(" ");
        } else {
            // No valid descriptors found, everything is the URL
            url_part = trimmed_entry.to_string();
            descriptor_part = String::new();
        }

        if url_part.is_empty() {
            return Err(Error::HtmlCodegen(format!(
                "No URL found in srcset entry: '{}'",
                entry
            )));
        }

        // Parse the URL - handle relative and absolute URLs properly
        let resolved_url = if url_part.starts_with("http://")
            || url_part.starts_with("https://")
            || url_part.starts_with("data:")
        {
            // Absolute URL - parse directly
            Url::parse(&url_part).map_err(|e| {
                Error::HtmlCodegen(format!(
                    "Invalid absolute URL in srcset '{}': {}",
                    url_part, e
                ))
            })?
        } else {
            // Relative URL - resolve against base
            base_url.join(&url_part).map_err(|e| {
                Error::HtmlCodegen(format!(
                    "Invalid relative URL in srcset '{}': {}",
                    url_part, e
                ))
            })?
        };

        let mut width = None;
        let mut height = None;
        let mut density = None;

        // Parse descriptors
        if !descriptor_part.is_empty() {
            for part in descriptor_part.split_whitespace() {
                if let Some(w) = part.strip_suffix('w').and_then(|w| w.parse().ok()) {
                    width = Some(w);
                } else if let Some(h) = part.strip_suffix('h').and_then(|h| h.parse().ok()) {
                    height = Some(h);
                } else if part == "x" || part.ends_with('x') {
                    let density_str = if part == "x" {
                        "1x".to_string()
                    } else {
                        part.to_string()
                    };
                    density = Some(density_str);
                } else {
                    // Unknown descriptor - according to HTML5 spec, unknown descriptors
                    // should be ignored but the entry should still be valid
                    // However, if we encounter unknown descriptors, we should stop parsing
                    // as they indicate malformed input
                    tracing::warn!(
                        "Unknown descriptor in srcset entry: '{}' in '{}'",
                        part,
                        entry
                    );
                    break;
                }
            }
        }

        Ok(SrcSetEntry {
            url: resolved_url,
            width,
            height,
            density,
        })
    }

    /// Check if a string is a valid srcset descriptor
    fn is_valid_descriptor(&self, s: &str) -> bool {
        match s {
            "x" => true, // Density descriptor without number
            s if s.ends_with('w') => {
                // Width descriptor: number + w
                s.strip_suffix('w')
                    .map_or(false, |num| !num.is_empty() && num.parse::<u32>().is_ok())
            }
            s if s.ends_with('h') => {
                // Height descriptor: number + h
                s.strip_suffix('h')
                    .map_or(false, |num| !num.is_empty() && num.parse::<u32>().is_ok())
            }
            s if s.ends_with('x') => {
                // Density descriptor: number + x
                s.strip_suffix('x')
                    .map_or(true, |num| num.is_empty() || num.parse::<f64>().is_ok())
            }
            _ => false, // Not a valid descriptor
        }
    }
}

impl Default for PictureParserBuilder {
    fn default() -> Self {
        Self {
            extract_alt_text: true,
            follow_redirects: false,
            max_file_size: None,
        }
    }
}

#[cfg(test)]
mod tests {
    use super::*;
    use crate::parse::element::test::{from_html_str, grab_element_from_document};
    use microformats_types::{Class, PropertyValue};
    use std::str::FromStr;
    use url::Url;

    #[test]
    fn test_picture_source_creation() {
        let source = PictureSource {
            srcset: vec![SrcSetEntry {
                url: "https://example.com/image.jpg".parse().unwrap(),
                width: Some(800),
                height: None,
                density: Some("1x".to_string()),
            }],
            media: "screen and (min-width: 800px)".to_string(),
            r#type: "image/jpeg".to_string(),
            sizes: "100vw".to_string(),
            attributes: BTreeMap::new(),
        };

        assert_eq!(source.srcset.len(), 1);
        assert_eq!(source.media, "screen and (min-width: 800px)");
        assert_eq!(source.r#type, "image/jpeg");
    }

    #[test]
    fn test_picture_creation() {
        let fallback = Image {
            value: "https://example.com/fallback.jpg".parse().unwrap(),
            alt: Some("Fallback image".to_string()),
        };

        let source = PictureSource {
            srcset: vec![SrcSetEntry {
                url: "https://example.com/responsive.jpg".parse().unwrap(),
                width: Some(1200),
                height: None,
                density: None,
            }],
            media: "screen and (min-width: 1200px)".to_string(),
            r#type: "image/webp".to_string(),
            sizes: "50vw".to_string(),
            attributes: BTreeMap::new(),
        };

        let picture = Picture {
            sources: vec![source],
            fallback: Some(fallback),
            attributes: BTreeMap::new(),
        };

        assert_eq!(picture.sources.len(), 1);
        assert!(picture.fallback.is_some());
        assert_eq!(
            picture.fallback.as_ref().unwrap().alt,
            Some("Fallback image".to_string())
        );
    }

    #[test]
    fn test_srcset_entry_serialization() {
        let entry = SrcSetEntry {
            url: "https://example.com/image-2x.jpg".parse().unwrap(),
            width: None,
            height: None,
            density: Some("2x".to_string()),
        };

        let json = serde_json::to_string(&entry).unwrap();
        assert!(json.contains("image-2x.jpg"));
        assert!(json.contains("2x"));
    }

    #[test]
    fn test_picture_serialization() {
        let picture = Picture {
            sources: vec![],
            fallback: None,
            attributes: BTreeMap::new(),
        };

        let json = serde_json::to_string(&picture).unwrap();
        println!("JSON output: {}", json);
        // When sources and fallback are empty/None, they should be skipped due to skip_serializing_if
        assert_eq!(json, "{}");
    }

    #[test]
    fn test_picture_parser_creation() {
        let parser = PictureParser::new();
        assert!(parser.config().extract_alt_text);
    }

    #[test]
    fn test_picture_parser_builder() {
        let builder = PictureParserBuilder {
            extract_alt_text: false,
            follow_redirects: true,
            max_file_size: Some(1024 * 1024),
        };

        let parser = PictureParser::with_config(builder);
        assert!(!parser.config().extract_alt_text);
        assert!(parser.config().follow_redirects);
        assert_eq!(parser.config().max_file_size, Some(1024 * 1024));
    }

    #[test]
    fn test_extract_dimensions() {
        let html = r#"<img src="test.jpg" width="100" height="200" alt="Test">"#;
        let doc = from_html_str(html);
        let element = grab_element_from_document(&doc, "img").unwrap();

        let parser = PictureParser::new();
        let node = crate::parse::element::Node { elem: element };
        let dimensions = parser.extract_dimensions(&node);

        assert_eq!(dimensions, Some((100, 200)));
    }

    #[test]
    fn test_is_representative_image() {
        let html = r#"<img src="avatar.jpg" class="u-photo" alt="Profile">"#;
        let doc = from_html_str(html);
        let element = grab_element_from_document(&doc, "img").unwrap();

        let parser = PictureParser::new();
        let node = crate::parse::element::Node { elem: element };
        assert!(parser.is_representative_image(&node));
    }

    #[test]
    fn test_extract_alt_text() {
        let html = r#"<img src="test.jpg" alt="Alternative text">"#;
        let doc = from_html_str(html);
        let element = grab_element_from_document(&doc, "img").unwrap();

        let parser = PictureParser::new();
        let node = crate::parse::element::Node { elem: element };
        let alt_text = parser.extract_alt_text(&node);

        assert_eq!(alt_text, Some("Alternative text".to_string()));
    }

    #[test]
    fn test_parse_simple_srcset() {
        let parser = PictureParser::new();
        let srcset =
            "https://example.com/image-400.jpg 400w, https://example.com/image-800.jpg 800w";
        let result = parser.parse_srcset(srcset).unwrap();

        assert_eq!(result.len(), 2);
        assert_eq!(result[0].url.as_str(), "https://example.com/image-400.jpg");
        assert_eq!(result[0].width, Some(400));
        assert_eq!(result[0].height, None);
        assert_eq!(result[0].density, None);

        assert_eq!(result[1].url.as_str(), "https://example.com/image-800.jpg");
        assert_eq!(result[1].width, Some(800));
        assert_eq!(result[1].height, None);
        assert_eq!(result[1].density, None);
    }

    #[test]
    fn test_parse_srcset_with_density_descriptors() {
        let parser = PictureParser::new();
        let srcset = "https://example.com/image-1x.jpg 1x, https://example.com/image-2x.jpg 2x";
        let result = parser.parse_srcset(srcset).unwrap();

        assert_eq!(result.len(), 2);
        assert_eq!(result[0].url.as_str(), "https://example.com/image-1x.jpg");
        assert_eq!(result[0].density, Some("1x".to_string()));
        assert_eq!(result[1].url.as_str(), "https://example.com/image-2x.jpg");
        assert_eq!(result[1].density, Some("2x".to_string()));
    }

    #[test]
    fn test_parse_srcset_with_height_descriptors() {
        let parser = PictureParser::new();
        let srcset =
            "https://example.com/image-600h.jpg 600h, https://example.com/image-900h.jpg 900h";
        let result = parser.parse_srcset(srcset).unwrap();

        assert_eq!(result.len(), 2);
        assert_eq!(result[0].url.as_str(), "https://example.com/image-600h.jpg");
        assert_eq!(result[0].height, Some(600));
        assert_eq!(result[1].url.as_str(), "https://example.com/image-900h.jpg");
        assert_eq!(result[1].height, Some(900));
    }

    #[test]
    fn test_parse_srcset_with_multiple_descriptors() {
        let parser = PictureParser::new();
        let srcset = "https://example.com/image-400x600-2x.jpg 400w 600h 2x, https://example.com/image-800x1200-1x.jpg 800w 1200h 1x";
        let result = parser.parse_srcset(srcset).unwrap();

        assert_eq!(result.len(), 2);
        assert_eq!(
            result[0].url.as_str(),
            "https://example.com/image-400x600-2x.jpg"
        );
        assert_eq!(result[0].width, Some(400));
        assert_eq!(result[0].height, Some(600));
        assert_eq!(result[0].density, Some("2x".to_string()));

        assert_eq!(
            result[1].url.as_str(),
            "https://example.com/image-800x1200-1x.jpg"
        );
        assert_eq!(result[1].width, Some(800));
        assert_eq!(result[1].height, Some(1200));
        assert_eq!(result[1].density, Some("1x".to_string()));
    }

    #[test]
    fn test_parse_srcset_with_urls_containing_spaces() {
        let parser = PictureParser::new();
        let srcset = "https://example.com/path%20to%20image.jpg 400w, https://example.com/path with spaces.jpg 800w";
        let result = parser.parse_srcset(srcset).unwrap();

        assert_eq!(result.len(), 2);
        assert_eq!(
            result[0].url.as_str(),
            "https://example.com/path%20to%20image.jpg"
        );
        assert_eq!(result[0].width, Some(400));
        // Spaces in URLs should be URL-encoded when parsed
        assert_eq!(
            result[1].url.as_str(),
            "https://example.com/path%20with%20spaces.jpg"
        );
        assert_eq!(result[1].width, Some(800));
    }

    #[test]
    fn test_parse_srcset_respects_commas() {
        let parser = PictureParser::new();
        let srcset = "https://example.com/image1.jpg, https://example.com/image2.jpg 400w, https://example.com/image3.jpg, https://example.com/image4.jpg 2x";
        let result = parser.parse_srcset(srcset).unwrap();

        assert_eq!(result.len(), 4);
        assert_eq!(result[0].url.as_str(), "https://example.com/image1.jpg");
        assert_eq!(result[1].url.as_str(), "https://example.com/image2.jpg");
        assert_eq!(result[2].url.as_str(), "https://example.com/image3.jpg");
        assert_eq!(result[3].url.as_str(), "https://example.com/image4.jpg");
    }

    #[test]
    fn test_parse_srcset_handles_malformed_input() {
        let parser = PictureParser::new();
        let srcset = "https://example.com/image.jpg 400w invalid, https://example.com/another.jpg";
        let result = parser.parse_srcset(srcset).unwrap();

        assert_eq!(result.len(), 2);
        assert_eq!(result[0].url.as_str(), "https://example.com/image.jpg");
        assert_eq!(result[0].width, Some(400));
        assert_eq!(result[1].url.as_str(), "https://example.com/another.jpg");
    }

    #[test]
    fn test_parse_srcset_empty_input() {
        let parser = PictureParser::new();
        let result = parser.parse_srcset("").unwrap();
        assert_eq!(result.len(), 0);
    }

    #[test]
    fn test_parse_srcset_url_parsing() {
        let parser = PictureParser::new();
        let base_url = Url::parse("https://example.com/").unwrap();
        let srcset = "/images/hero-800w.jpg 800w, /images/hero-400w.jpg 400w";
        let result = parser.parse_srcset_with_base(srcset, &base_url).unwrap();

        assert_eq!(result.len(), 2);
        assert_eq!(
            result[0].url.as_str(),
            "https://example.com/images/hero-800w.jpg"
        );
        assert_eq!(
            result[1].url.as_str(),
            "https://example.com/images/hero-400w.jpg"
        );
    }

    #[test]
    fn test_parse_picture_element_basic() {
        let html = r#"
        <picture>
            <source srcset="/images/hero-800w.jpg 800w, /images/hero-400w.jpg 400w" media="(min-width: 800px)" type="image/jpeg">
            <source srcset="/images/hero-600h.jpg 600h" media="(min-width: 600px)" type="image/webp">
            <img src="/images/fallback.jpg" alt="Hero image">
        </picture>
        "#;

        let doc = from_html_str(html);
        let element = grab_element_from_document(&doc, "picture").unwrap();
        let base_url = Url::parse("https://example.com/").unwrap();

        let parser = PictureParser::new();
        let node = Node { elem: element };
        let result = parser.parse_picture_element(&node, &base_url).unwrap();

        assert!(result.is_some());
        let picture = result.unwrap();

        // Verify sources
        assert_eq!(picture.sources.len(), 2);

        // First source
        assert_eq!(picture.sources[0].media, "(min-width: 800px)");
        assert_eq!(picture.sources[0].r#type, "image/jpeg");
        assert_eq!(picture.sources[0].srcset.len(), 2);
        assert_eq!(
            picture.sources[0].srcset[0].url.as_str(),
            "https://example.com/images/hero-800w.jpg"
        );
        assert_eq!(picture.sources[0].srcset[0].width, Some(800));
        assert_eq!(
            picture.sources[0].srcset[1].url.as_str(),
            "https://example.com/images/hero-400w.jpg"
        );
        assert_eq!(picture.sources[0].srcset[1].width, Some(400));

        // Second source
        assert_eq!(picture.sources[1].media, "(min-width: 600px)");
        assert_eq!(picture.sources[1].r#type, "image/webp");
        assert_eq!(picture.sources[1].srcset.len(), 1);
        assert_eq!(
            picture.sources[1].srcset[0].url.as_str(),
            "https://example.com/images/hero-600h.jpg"
        );
        assert_eq!(picture.sources[1].srcset[0].height, Some(600));

        // Fallback image
        assert!(picture.fallback.is_some());
        let fallback = picture.fallback.unwrap();
        assert_eq!(
            fallback.value.as_str(),
            "https://example.com/images/fallback.jpg"
        );
        assert_eq!(fallback.alt, Some("Hero image".to_string()));
    }

    #[test]
    fn test_parse_picture_element_with_sizes() {
        let html = r#"
        <picture>
            <source srcset="/images/responsive-800w.jpg 800w, /images/responsive-400w.jpg 400w"
                    media="(min-width: 800px)"
                    type="image/jpeg"
                    sizes="(min-width: 800px) 800px, 400px">
            <img src="/images/fallback-400.jpg" alt="Responsive image" width="400" height="300">
        </picture>
        "#;

        let doc = from_html_str(html);
        let element = grab_element_from_document(&doc, "picture").unwrap();
        let base_url = Url::parse("https://example.com/").unwrap();

        let parser = PictureParser::new();
        let node = Node { elem: element };
        let result = parser.parse_picture_element(&node, &base_url).unwrap();

        assert!(result.is_some());
        let picture = result.unwrap();

        assert_eq!(picture.sources.len(), 1);
        let source = &picture.sources[0];
        assert_eq!(source.sizes, "(min-width: 800px) 800px, 400px");
        assert_eq!(source.srcset.len(), 2);

        assert!(picture.fallback.is_some());
        let fallback = picture.fallback.unwrap();
        assert_eq!(
            fallback.value.as_str(),
            "https://example.com/images/fallback-400.jpg"
        );
    }

    #[test]
    fn test_parse_picture_element_empty_picture() {
        let html = r#"<picture></picture>"#;

        let doc = from_html_str(html);
        let element = grab_element_from_document(&doc, "picture").unwrap();
        let base_url = Url::parse("https://example.com/").unwrap();

        let parser = PictureParser::new();
        let node = Node { elem: element };
        let result = parser.parse_picture_element(&node, &base_url).unwrap();

        assert!(result.is_some());
        let picture = result.unwrap();

        assert_eq!(picture.sources.len(), 0);
        assert!(picture.fallback.is_none());
    }

    #[test]
    fn test_parse_picture_element_only_sources() {
        let html = r#"
        <picture>
            <source srcset="/images/hero-800w.jpg 800w" media="(min-width: 800px)" type="image/webp">
            <source srcset="/images/hero-400w.jpg 400w" media="(min-width: 400px)" type="image/png">
        </picture>
        "#;

        let doc = from_html_str(html);
        let element = grab_element_from_document(&doc, "picture").unwrap();
        let base_url = Url::parse("https://example.com/").unwrap();

        let parser = PictureParser::new();
        let node = Node { elem: element };
        let result = parser.parse_picture_element(&node, &base_url).unwrap();

        assert!(result.is_some());
        let picture = result.unwrap();

        assert_eq!(picture.sources.len(), 2);
        assert!(picture.fallback.is_none());

        assert_eq!(
            picture.sources[0].srcset[0].url.as_str(),
            "https://example.com/images/hero-800w.jpg"
        );
        assert_eq!(
            picture.sources[1].srcset[0].url.as_str(),
            "https://example.com/images/hero-400w.jpg"
        );
    }

    #[test]
    fn test_parse_picture_element_only_fallback() {
        let html = r#"
        <picture>
            <img src="/images/single.jpg" alt="Single image">
        </picture>
        "#;

        let doc = from_html_str(html);
        let element = grab_element_from_document(&doc, "picture").unwrap();
        let base_url = Url::parse("https://example.com/").unwrap();

        let parser = PictureParser::new();
        let node = Node { elem: element };
        let result = parser.parse_picture_element(&node, &base_url).unwrap();

        assert!(result.is_some());
        let picture = result.unwrap();

        assert_eq!(picture.sources.len(), 0);
        assert!(picture.fallback.is_some());

        let fallback = picture.fallback.unwrap();
        assert_eq!(
            fallback.value.as_str(),
            "https://example.com/images/single.jpg"
        );
        assert_eq!(fallback.alt, Some("Single image".to_string()));
    }

    // ========== COMPREHENSIVE END-TO-END INTEGRATION TESTS ==========

    /// Helper function to parse full HTML document and return the first item
    fn parse_html_to_item(html: &str) -> Result<microformats_types::Item, crate::Error> {
        let base_url = Url::parse("https://example.com/").unwrap();
        let mut parser = crate::parse::Parser::from_html(html.to_string())?;
        let document = parser.into_document(Some(base_url))?;

        document
            .items
            .into_iter()
            .next()
            .ok_or_else(|| crate::Error::HtmlCodegen("No items found".to_string()))
    }

    #[test]
    fn test_pname_from_picture() -> Result<(), crate::Error> {
        let html = r#"
        <article class="h-entry">
            <h1 class="p-name">Blog Post Title</h1>
            <picture class="p-name">
                <source srcset="/images/hero-800w.jpg 800w, /images/hero-400w.jpg 400w"
                        media="(min-width: 800px)" type="image/jpeg">
                <img src="/images/fallback.jpg" alt="Hero image">
            </picture>
        </article>
        "#;

        let item = parse_html_to_item(html)?;

        let name_property = item
            .properties
            .get("name")
            .ok_or_else(|| crate::Error::HtmlCodegen("No name property found".to_string()))?;

        // Should have 2 name properties: one from h1 and one from picture
        assert_eq!(name_property.len(), 2);

        // Verify we have "Blog Post Title" from h1 and "Hero image" from picture
        let mut found_blog_title = false;
        let mut found_hero_image = false;

        for val in name_property.iter() {
            if let PropertyValue::Plain(plain) = val {
                let text = plain.to_string();
                if text == "Blog Post Title" {
                    found_blog_title = true;
                } else if text == "Hero image" {
                    found_hero_image = true;
                }
            }
        }

        assert!(found_blog_title, "Blog Post Title not found");
        assert!(found_hero_image, "Hero image not found");

        Ok(())
    }

    #[test]
    fn test_multiple_pictures() -> Result<(), crate::Error> {
        let html = r#"
        <article class="h-entry">
            <h1 class="p-name">Blog Post with Images</h1>
            <picture class="u-photo">
                <source srcset="/images/hero-800w.jpg 800w, /images/hero-400w.jpg 400w"
                        media="(min-width: 800px)" type="image/jpeg">
                <img src="/images/hero-fallback.jpg" alt="Hero image">
            </picture>
            <picture class="u-photo">
                <source srcset="/images/thumb-200w.jpg 200w" media="(min-width: 200px)" type="image/png">
                <img src="/images/thumb-fallback.png" alt="Thumbnail">
            </picture>
        </article>
        "#;

        let item = parse_html_to_item(html)?;

        let photo_property = item
            .properties
            .get("photo")
            .ok_or_else(|| crate::Error::HtmlCodegen("No photo property found".to_string()))?;

        // Should have two photo values
        assert_eq!(photo_property.len(), 2);

        // Verify both images
        match &photo_property[0] {
            PropertyValue::Image(image) => {
                assert_eq!(
                    image.value.as_str(),
                    "https://example.com/images/hero-fallback.jpg"
                );
                assert_eq!(image.alt, Some("Hero image".to_string()));
            }
            _ => panic!(
                "Expected Image property value for first photo, got: {:?}",
                photo_property[0]
            ),
        }

        match &photo_property[1] {
            PropertyValue::Image(image) => {
                assert_eq!(
                    image.value.as_str(),
                    "https://example.com/images/thumb-fallback.png"
                );
                assert_eq!(image.alt, Some("Thumbnail".to_string()));
            }
            _ => panic!(
                "Expected Image property value for second photo, got: {:?}",
                photo_property[1]
            ),
        }

        Ok(())
    }

    #[test]
    fn test_picture_multiple_sources() -> Result<(), crate::Error> {
        let html = r#"
        <article class="h-entry">
            <h1 class="p-name">Responsive Image Post</h1>
            <picture class="u-photo">
                <source srcset="/images/desktop-1200w.jpg 1200w, /images/desktop-800w.jpg 800w"
                        media="(min-width: 1200px)" type="image/jpeg">
                <source srcset="/images/tablet-800w.jpg 800w, /images/tablet-400w.jpg 400w"
                        media="(min-width: 768px)" type="image/webp">
                <source srcset="/images/mobile-400w.jpg 400w"
                        media="(min-width: 320px)" type="image/png">
                <img src="/images/fallback.jpg" alt="Responsive hero image">
            </picture>
        </article>
        "#;

        let item = parse_html_to_item(html)?;

        let name_property = item
            .properties
            .get("name")
            .ok_or_else(|| crate::Error::HtmlCodegen("No name property found".to_string()))?;

        // Should have 1 name property: just from h1 (picture with u-photo doesn't contribute to p-name)
        assert_eq!(name_property.len(), 1);

        // The name should be from the h1 element
        match &name_property[0] {
            PropertyValue::Plain(plain) => {
                assert_eq!(plain.to_string(), "Responsive Image Post");
            }
            _ => panic!(
                "Expected Plain property value for p-name, got: {:?}",
                name_property[0]
            ),
        }

        Ok(())
    }

    #[test]
    fn test_picture_data_urls() -> Result<(), crate::Error> {
        let html = r#"
        <article class="h-entry">
            <h1 class="p-name">Post with Data URL</h1>
            <picture class="u-photo">
                <source srcset="data:image/jpeg;base64,/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAYEBQYFBAYGBQYHBwYIChAKCgkJChQODwwQFxQYGBcUFhYaHSUfGhsjHBYWICwgIyYnKSopGR8tMC0oMCUoKSj/2wBDAQcHBwoIChMKChMoGhYaKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCj/wAARCAABAAEDASIAAhEBAxEB/8QAFQABAQAAAAAAAAAAAAAAAAAAAAv/xAAUEAEAAAAAAAAAAAAAAAAAAAAA/8QAFQEBAQAAAAAAAAAAAAAAAAAAAAX/xAAUEQEAAAAAAAAAAAAAAAAAAAAA/9oADAMBAAIRAxEAPwCdABmX/9k= 1x"
                        type="image/jpeg">
                <img src="data:image/jpeg;base64,/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAYEBQYFBAYGBQYHBwYIChAKCgkJChQODwwQFxQYGBcUFhYaHSUfGhsjHBYWICwgIyYnKSopGR8tMC0oMCUoKSj/2wBDAQcHBwoIChMKChMoGhYaKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCj/wAARCAABAAEDASIAAhEBAxEB/8QAFQABAQAAAAAAAAAAAAAAAAAAAAv/xAAUEAEAAAAAAAAAAAAAAAAAAAAA/8QAFQEBAQAAAAAAAAAAAAAAAAAAAAX/xAAUEQEAAAAAAAAAAAAAAAAAAAAA/9oADAMBAAIRAxEAPwCdABmX/9k= 2x"
                        alt="Data URL image">
            </picture>
        </article>
        "#;

        let item = parse_html_to_item(html)?;

        let photo_property = item
            .properties
            .get("photo")
            .ok_or_else(|| crate::Error::HtmlCodegen("No photo property found".to_string()))?;

        assert_eq!(photo_property.len(), 1);

        match &photo_property[0] {
            PropertyValue::Image(image) => {
                assert!(image.value.as_str().starts_with("data:image/jpeg;base64,"));
                assert_eq!(image.alt, Some("Data URL image".to_string()));
            }
            _ => panic!(
                "Expected Image property value, got: {:?}",
                photo_property[0]
            ),
        }

        Ok(())
    }

    #[test]
    fn test_realistic_blog_post() -> Result<(), crate::Error> {
        let html = r#"
        <!DOCTYPE html>
        <html>
        <head>
            <title>My Travel Blog</title>
        </head>
        <body>
            <article class="h-entry" id="travel-post-1">
                <h1 class="p-name">Amazing Trip to Japan</h1>
                <div class="p-author h-card">
                    <img class="u-photo" src="/images/author-avatar.jpg" alt="John Traveler">
                    <span class="p-name">John Traveler</span>
                </div>

                <picture class="u-photo">
                    <source srcset="/images/japan-hero-1200w.jpg 1200w, /images/japan-hero-800w.jpg 800w, /images/japan-hero-400w.jpg 400w"
                            media="(min-width: 1200px)" type="image/jpeg">
                    <source srcset="/images/japan-hero-tablet-800w.jpg 800w, /images/japan-hero-tablet-400w.jpg 400w"
                            media="(min-width: 768px)" type="image/webp">
                    <source srcset="/images/japan-hero-mobile-400w.jpg 400w, /images/japan-hero-mobile-200w.jpg 200w"
                            media="(min-width: 320px)" type="image/png">
                    <img src="/images/japan-hero-fallback.jpg" alt="Beautiful cherry blossoms in Tokyo">
                </picture>

                <div class="e-content">
                    <p>I recently visited Japan and it was absolutely amazing! The cherry blossoms were in full bloom.</p>

                    <picture class="u-photo">
                        <source srcset="/images/cherry-blossom-600w.jpg 600w"
                                media="(min-width: 600px)" type="image/jpeg">
                        <img src="/images/cherry-blossom-300w.jpg" alt="Close-up of cherry blossoms">
                    </picture>

                    <p>The temples were breathtaking and the food was incredible.</p>
                </div>
            </article>
        </body>
        </html>
        "#;

        let base_url = Url::parse("https://mytravelblog.example.com/").unwrap();
        let mut parser = crate::parse::Parser::from_html(html.to_string())?;
        let document = parser.into_document(Some(base_url))?;

        // Debug: Check what items we actually get
        println!("Number of items found: {}", document.items.len());
        for (i, item) in document.items.iter().enumerate() {
            println!("Item {}: types={:?}", i, item.r#type);
            println!(
                "  Properties: {:?}",
                item.properties.keys().collect::<Vec<_>>()
            );
        }

        // For now, let's be flexible about the number of items
        // The h-card might be nested and not created as a separate item
        assert!(
            document.items.len() >= 1,
            "Should have at least the h-entry item"
        );

        // Check h-entry (first item)
        let h_entry = &document.items[0];
        assert_eq!(h_entry.r#type, vec![Class::from_str("h-entry").unwrap()]);
        assert_eq!(h_entry.id, Some("travel-post-1".to_string()));

        // Verify properties
        let expected_properties = ["name", "photo", "content"];
        for prop in &expected_properties {
            assert!(
                h_entry.properties.contains_key(&prop.to_string()),
                "Missing property: {}",
                prop
            );
        }

        // Check photo property (should have 2 images from the two picture elements)
        let photo_property = h_entry
            .properties
            .get("photo")
            .ok_or_else(|| crate::Error::HtmlCodegen("No photo property found".to_string()))?;
        assert_eq!(photo_property.len(), 2); // Two picture elements only

        // Verify first image (hero image from picture element)
        match &photo_property[0] {
            PropertyValue::Image(image) => {
                assert_eq!(
                    image.value.as_str(),
                    "https://mytravelblog.example.com/images/japan-hero-fallback.jpg"
                );
                assert_eq!(
                    image.alt,
                    Some("Beautiful cherry blossoms in Tokyo".to_string())
                );
            }
            _ => panic!("Expected Image property value for hero image"),
        }

        // Verify second image (cherry blossom close-up from picture element)
        match &photo_property[1] {
            PropertyValue::Image(image) => {
                assert_eq!(
                    image.value.as_str(),
                    "https://mytravelblog.example.com/images/cherry-blossom-300w.jpg"
                );
                assert_eq!(image.alt, Some("Close-up of cherry blossoms".to_string()));
            }
            _ => panic!("Expected Image property value for cherry blossom image"),
        }

        // If we have a second item, check h-card
        if document.items.len() > 1 {
            let h_card = &document.items[1];
            assert_eq!(h_card.r#type, vec![Class::from_str("h-card").unwrap()]);

            let h_card_photo = h_card.properties.get("photo").ok_or_else(|| {
                crate::Error::HtmlCodegen("No photo property found in h-card".to_string())
            })?;
            assert_eq!(h_card_photo.len(), 1);

            match &h_card_photo[0] {
                PropertyValue::Image(image) => {
                    assert_eq!(
                        image.value.as_str(),
                        "https://mytravelblog.example.com/images/author-avatar.jpg"
                    );
                    assert_eq!(image.alt, Some("John Traveler".to_string()));
                }
                _ => panic!("Expected Image property value for author avatar"),
            }
        }

        Ok(())
    }

    #[test]
    fn test_picture_edge_cases() -> Result<(), crate::Error> {
        let html = r#"
        <article class="h-entry">
            <h1 class="p-name">Post with Edge Cases</h1>

            <!-- Picture with malformed srcset -->
            <picture class="u-photo">
                <source srcset="valid.jpg 400w, broken, valid2.jpg 800w"
                        media="(min-width: 800px)" type="image/jpeg">
                <img src="/images/fallback.jpg" alt="Edge case image">
            </picture>

            <!-- Picture with empty picture element -->
            <picture class="u-photo">
            </picture>

            <!-- Picture with only sources, no img -->
            <picture class="u-photo">
                <source srcset="/images/only-sources.jpg 400w" type="image/jpeg">
            </picture>
        </article>
        "#;

        let base_url = Url::parse("https://example.com/").unwrap();
        let mut parser = crate::parse::Parser::from_html(html.to_string())?;
        let document = parser.into_document(Some(base_url))?;

        assert_eq!(document.items.len(), 1);
        let item = &document.items[0];

        // Should still find at least the first picture with valid fallback
        let photo_property = item.properties.get("photo");
        assert!(photo_property.is_some());

        Ok(())
    }

    #[test]
    fn test_picture_performance() -> Result<(), crate::Error> {
        use std::time::Instant;

        // Generate HTML with many picture elements
        let mut html =
            String::from(r#"<article class="h-entry"><h1 class="p-name">Performance Test</h1>"#);

        // Add 50 picture elements
        for i in 0..50 {
            html.push_str(&format!(
                r#"
            <picture class="u-photo">
                <source srcset="/images/{}-800w.jpg 800w, /images/{}-400w.jpg 400w"
                        media="(min-width: 800px)" type="image/jpeg">
                <img src="/images/{}-fallback.jpg" alt="Image {}">
            </picture>
            "#,
                i, i, i, i
            ));
        }

        html.push_str("</article>");

        let start = Instant::now();

        let base_url = Url::parse("https://example.com/").unwrap();
        let mut parser = crate::parse::Parser::from_html(html)?;
        let document = parser.into_document(Some(base_url))?;

        let duration = start.elapsed();

        // Should parse within reasonable time (adjust threshold as needed)
        assert!(
            duration.as_millis() < 5000,
            "Picture parsing took too long: {:?}",
            duration
        );

        // Verify all images were parsed
        assert_eq!(document.items.len(), 1);
        let item = &document.items[0];
        let photo_property = item.properties.get("photo").unwrap();
        assert_eq!(photo_property.len(), 50);

        println!("Parsed 50 picture elements in {:?}", duration);

        Ok(())
    }

    #[test]
    fn test_picture_empty_srcset() -> Result<(), crate::Error> {
        let html = r#"
        <article class="h-entry">
            <h1 class="p-name">Post with Empty Source</h1>
            <picture class="u-photo">
                <source srcset="" media="(min-width: 800px)" type="image/jpeg">
                <img src="/images/fallback.jpg" alt="Fallback image">
            </picture>
        </article>
        "#;

        let item = parse_html_to_item(html)?;

        let photo_property = item.properties.get("photo");

        // Should still have the fallback image even with empty srcset
        assert!(photo_property.is_some());
        assert_eq!(photo_property.unwrap().len(), 1);

        Ok(())
    }

    #[test]
    fn test_picture_no_alt_text() -> Result<(), crate::Error> {
        let html = r#"
        <article class="h-entry">
            <h1 class="p-name">Post with Image No Alt</h1>
            <picture class="u-photo">
                <source srcset="/images/hero-800w.jpg 800w, /images/hero-400w.jpg 400w"
                        media="(min-width: 800px)" type="image/jpeg">
                <img src="/images/fallback.jpg">
            </picture>
        </article>
        "#;

        let item = parse_html_to_item(html)?;

        let photo_property = item
            .properties
            .get("photo")
            .ok_or_else(|| crate::Error::HtmlCodegen("No photo property found".to_string()))?;

        assert_eq!(photo_property.len(), 1);

        match &photo_property[0] {
            PropertyValue::Image(image) => {
                assert_eq!(
                    image.value.as_str(),
                    "https://example.com/images/fallback.jpg"
                );
                assert_eq!(image.alt, None); // No alt text should be None
            }
            PropertyValue::Url(url_value) => {
                // If no alt text, it should return as URL instead of Image
                assert_eq!(
                    url_value.as_str(),
                    "https://example.com/images/fallback.jpg"
                );
            }
            _ => panic!(
                "Expected Image or Url property value, got: {:?}",
                photo_property[0]
            ),
        }

        Ok(())
    }

    #[test]
    fn test_picture_in_hcard() -> Result<(), crate::Error> {
        let html = r#"
        <div class="h-card">
            <img class="u-photo" src="/avatar.jpg" alt="John Doe">
        </div>
        <article class="h-entry">
            <h1 class="p-name">Post by John</h1>
            <picture class="u-photo">
                <source srcset="/images/author-400w.jpg 400w, /images/author-200w.jpg 200w"
                        media="(min-width: 400px)" type="image/jpeg">
                <img src="/images/author-fallback.jpg" alt="John's photo">
            </picture>
        </article>
        "#;

        let base_url = Url::parse("https://example.com/").unwrap();
        let mut parser = crate::parse::Parser::from_html(html.to_string())?;
        let document = parser.into_document(Some(base_url))?;

        // Should have 2 items: h-card and h-entry
        assert_eq!(document.items.len(), 2);

        // Check h-card (first item)
        let h_card = &document.items[0];
        assert_eq!(h_card.r#type, vec![Class::from_str("h-card").unwrap()]);

        let h_card_photo = h_card.properties.get("photo").ok_or_else(|| {
            crate::Error::HtmlCodegen("No photo property found in h-card".to_string())
        })?;
        assert_eq!(h_card_photo.len(), 1);

        // Check h-entry (second item)
        let h_entry = &document.items[1];
        assert_eq!(h_entry.r#type, vec![Class::from_str("h-entry").unwrap()]);

        let h_entry_photo = h_entry.properties.get("photo").ok_or_else(|| {
            crate::Error::HtmlCodegen("No photo property found in h-entry".to_string())
        })?;
        assert_eq!(h_entry_photo.len(), 1);

        Ok(())
    }

    #[test]
    fn test_picture_relative_urls() -> Result<(), crate::Error> {
        let html = r#"
        <article class="h-entry">
            <h1 class="p-name">Post with Relative URLs</h1>
            <picture class="u-photo">
                <source srcset="images/hero-800w.jpg 800w, images/hero-400w.jpg 400w"
                        media="(min-width: 800px)" type="image/jpeg">
                <img src="images/fallback.jpg" alt="Relative path image">
            </picture>
        </article>
        "#;

        let item = parse_html_to_item(html)?;

        let photo_property = item
            .properties
            .get("photo")
            .ok_or_else(|| crate::Error::HtmlCodegen("No photo property found".to_string()))?;

        assert_eq!(photo_property.len(), 1);

        match &photo_property[0] {
            microformats_types::PropertyValue::Image(image) => {
                // Relative URLs should be resolved against base URL
                assert_eq!(
                    image.value.as_str(),
                    "https://example.com/images/fallback.jpg"
                );
                assert_eq!(image.alt, Some("Relative path image".to_string()));
            }
            _ => panic!(
                "Expected Image property value, got: {:?}",
                photo_property[0]
            ),
        }

        Ok(())
    }
}