RubyGems - html-to-markdown - Versions diffs - 2.30.0 → 3.0.0 - Mend

html-to-markdown 2.30.0 → 3.0.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Files changed (166) hide show

data/vendor/html-to-markdown-rs/src/lib.rs CHANGED Viewed

@@ -33,7 +33,6 @@
     clippy::assigning_clones,
     clippy::uninlined_format_args
 )]
-#![allow(dead_code)]
 //! High-performance HTML to Markdown converter.
 //!
@@ -50,7 +49,6 @@
 pub mod converter;
 pub mod error;
-pub mod hocr;
 #[cfg(feature = "inline-images")]
 mod inline_images;
 #[cfg(feature = "metadata")]
@@ -58,6 +56,7 @@ pub mod metadata;
 pub mod options;
 pub mod safety;
 pub mod text;
+pub mod types;
 #[cfg(feature = "visitor")]
 pub mod visitor;
 #[cfg(feature = "visitor")]
@@ -76,6 +75,11 @@ mod validation;
 // ============================================================================
 pub use exports::*;
+pub use types::{
+    AnnotationKind, ConversionResult, DocumentNode, DocumentStructure, GridCell, NodeContent, ProcessingWarning,
+    TableGrid, TextAnnotation, WarningKind,
+};
+// Note: types::TableData will replace convert_api::TableData when convert() is refactored
 // ============================================================================
 // Main Public API Functions
@@ -90,141 +94,15 @@ pub use convert_api::{conversion_options_from_json, conversion_options_update_fr
 pub use convert_api::metadata_config_from_json;
 #[cfg(feature = "inline-images")]
-pub use convert_api::{convert_with_inline_images, inline_image_config_from_json};
-#[cfg(feature = "metadata")]
-pub use convert_api::convert_with_metadata;
+pub use convert_api::inline_image_config_from_json;
 #[cfg(feature = "visitor")]
+#[doc(hidden)]
 pub use convert_api::convert_with_visitor;
-#[cfg(feature = "visitor")]
-pub use convert_api::{ConversionWithTables, TableData, convert_with_tables};
-#[cfg(feature = "async-visitor")]
-pub use convert_api::convert_with_async_visitor;
 // Tests
 // ============================================================================
-#[cfg(all(test, feature = "metadata"))]
-mod tests {
-    use super::*;
-    #[test]
-    fn test_convert_with_metadata_full_workflow() {
-        let html = "<html lang=\"en\" dir=\"ltr\"><head><title>Test Article</title></head><body><h1 id=\"main-title\">Main Title</h1><p>This is a paragraph with a <a href=\"https://example.com\">link</a>.</p><h2>Subsection</h2><p>Another paragraph with <a href=\"#main-title\">internal link</a>.</p><img src=\"https://example.com/image.jpg\" alt=\"Test image\" title=\"Image title\"></body></html>";
-        let config = MetadataConfig {
-            extract_document: true,
-            extract_headers: true,
-            extract_links: true,
-            extract_images: true,
-            extract_structured_data: true,
-            max_structured_data_size: metadata::DEFAULT_MAX_STRUCTURED_DATA_SIZE,
-        };
-        let (markdown, metadata) = convert_with_metadata(html, None, config, None).expect("conversion should succeed");
-        assert!(!markdown.is_empty());
-        assert!(markdown.contains("Main Title"));
-        assert!(markdown.contains("Subsection"));
-        assert_eq!(metadata.document.language, Some("en".to_string()));
-        assert_eq!(metadata.headers.len(), 2);
-        assert_eq!(metadata.headers[0].level, 1);
-        assert_eq!(metadata.headers[0].text, "Main Title");
-        assert_eq!(metadata.headers[0].id, Some("main-title".to_string()));
-        assert_eq!(metadata.headers[1].level, 2);
-        assert_eq!(metadata.headers[1].text, "Subsection");
-        assert!(metadata.links.len() >= 2);
-        let external_link = metadata.links.iter().find(|l| l.link_type == LinkType::External);
-        assert!(external_link.is_some());
-        let anchor_link = metadata.links.iter().find(|l| l.link_type == LinkType::Anchor);
-        assert!(anchor_link.is_some());
-        assert_eq!(metadata.images.len(), 1);
-        assert_eq!(metadata.images[0].alt, Some("Test image".to_string()));
-        assert_eq!(metadata.images[0].title, Some("Image title".to_string()));
-        assert_eq!(metadata.images[0].image_type, ImageType::External);
-    }
-    #[test]
-    fn test_convert_with_metadata_document_fields() {
-        let html = "<html lang=\"en\"><head><title>Test Article</title><meta name=\"description\" content=\"Desc\"><meta name=\"author\" content=\"Author\"><meta property=\"og:title\" content=\"OG Title\"><meta property=\"og:description\" content=\"OG Desc\"></head><body><h1>Heading</h1></body></html>";
-        let (_markdown, metadata) =
-            convert_with_metadata(html, None, MetadataConfig::default(), None).expect("conversion should succeed");
-        assert_eq!(
-            metadata.document.title,
-            Some("Test Article".to_string()),
-            "document: {:?}",
-            metadata.document
-        );
-        assert_eq!(metadata.document.description, Some("Desc".to_string()));
-        assert_eq!(metadata.document.author, Some("Author".to_string()));
-        assert_eq!(metadata.document.language, Some("en".to_string()));
-        assert_eq!(metadata.document.open_graph.get("title"), Some(&"OG Title".to_string()));
-        assert_eq!(
-            metadata.document.open_graph.get("description"),
-            Some(&"OG Desc".to_string())
-        );
-    }
-    #[test]
-    fn test_convert_with_metadata_empty_config() {
-        let html = "<html lang=\"en\"><head><title>Test</title></head><body><h1>Title</h1><a href=\"#\">Link</a></body></html>";
-        let config = MetadataConfig {
-            extract_document: false,
-            extract_headers: false,
-            extract_links: false,
-            extract_images: false,
-            extract_structured_data: false,
-            max_structured_data_size: 0,
-        };
-        let (_markdown, metadata) = convert_with_metadata(html, None, config, None).expect("conversion should succeed");
-        assert!(metadata.headers.is_empty());
-        assert!(metadata.links.is_empty());
-        assert!(metadata.images.is_empty());
-        assert_eq!(metadata.document.language, None);
-    }
-    #[test]
-    fn test_convert_with_metadata_data_uri_image() {
-        let html = "<html><body><img src=\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mNk+M9QDwADhgGAWjR9awAAAABJRU5ErkJggg==\" alt=\"Pixel\"></body></html>";
-        let config = MetadataConfig::default();
-        let (_markdown, metadata) = convert_with_metadata(html, None, config, None).expect("conversion should succeed");
-        assert_eq!(metadata.images.len(), 1);
-        assert_eq!(metadata.images[0].image_type, ImageType::DataUri);
-        assert_eq!(metadata.images[0].alt, Some("Pixel".to_string()));
-    }
-    #[test]
-    fn test_convert_with_metadata_relative_paths() {
-        let html = r#"<html><body><a href="/page">Internal</a><a href="../other">Relative</a></body></html>"#;
-        let config = MetadataConfig::default();
-        let (_markdown, metadata) = convert_with_metadata(html, None, config, None).expect("conversion should succeed");
-        let internal_link_count = metadata
-            .links
-            .iter()
-            .filter(|l| l.link_type == LinkType::Internal)
-            .count();
-        assert_eq!(internal_link_count, 2);
-    }
-}
 #[cfg(test)]
 mod basic_tests {
     use super::*;
@@ -253,7 +131,8 @@ mod basic_tests {
     #[test]
     fn test_plain_text_allowed() {
         let result = convert("Just text", None).unwrap();
-        assert!(result.contains("Just text"));
+        let content = result.content.unwrap_or_default();
+        assert!(content.contains("Just text"));
     }
     #[test]
@@ -264,7 +143,8 @@ mod basic_tests {
             ..ConversionOptions::default()
         };
         let result = convert("Text *asterisks* _underscores_", Some(options)).unwrap();
-        assert!(result.contains(r"\*asterisks\*"));
-        assert!(result.contains(r"\_underscores\_"));
+        let content = result.content.unwrap_or_default();
+        assert!(content.contains(r"\*asterisks\*"));
+        assert!(content.contains(r"\_underscores\_"));
     }
 }

data/vendor/html-to-markdown-rs/src/metadata/collector.rs CHANGED Viewed

@@ -2,7 +2,7 @@
 use super::config::MetadataConfig;
 use super::extraction::{extract_document_metadata, extract_structured_data};
-use super::types::{ExtendedMetadata, ImageMetadata, ImageType, LinkMetadata};
+use super::types::{HtmlMetadata, ImageMetadata, ImageType, LinkMetadata};
 use std::collections::BTreeMap;
 /// Internal metadata collector for single-pass extraction.
@@ -256,13 +256,13 @@ impl MetadataCollector {
     /// Finish collection and return all extracted metadata.
     ///
     /// Performs final processing, validation, and consolidation of all
-    /// collected data into the [`ExtendedMetadata`] output structure.
+    /// collected data into the [`HtmlMetadata`] output structure.
     #[allow(dead_code)]
-    pub(crate) fn finish(self) -> ExtendedMetadata {
+    pub(crate) fn finish(self) -> HtmlMetadata {
         let structured_data = extract_structured_data(self.json_ld);
         let document = extract_document_metadata(self.head_metadata, self.lang, self.dir);
-        ExtendedMetadata {
+        HtmlMetadata {
             document,
             headers: self.headers,
             links: self.links,

data/vendor/html-to-markdown-rs/src/metadata/mod.rs CHANGED Viewed

@@ -40,14 +40,14 @@
 //! - [`ImageMetadata`]: Image element with src, alt, title, dimensions, type, and attributes
 //! - [`StructuredData`]: Structured data block with type and raw JSON
 //! - [`MetadataConfig`]: Configuration controlling extraction granularity and size limits
-//! - [`ExtendedMetadata`]: Top-level result containing all extracted metadata
+//! - [`HtmlMetadata`]: Top-level result containing all extracted metadata
 //!
 //! # Examples
 //!
-//! ## Basic Usage with `convert_with_metadata`
+//! ## Basic Usage with `convert()`
 //!
 //! ```ignore
-//! use html_to_markdown_rs::{convert_with_metadata, MetadataConfig};
+//! use html_to_markdown_rs::convert;
 //!
 //! let html = r#"
 //!   <html lang="en">
@@ -63,8 +63,8 @@
 //!   </html>
 //! "#;
 //!
-//! let config = MetadataConfig::default();
-//! let (markdown, metadata) = convert_with_metadata(html, None, config)?;
+//! let result = convert(html, None)?;
+//! let metadata = result.metadata.unwrap();
 //!
 //! // Access document metadata
 //! assert_eq!(metadata.document.title, Some("My Article".to_string()));
@@ -88,28 +88,26 @@
 //! ## Selective Extraction
 //!
 //! ```ignore
-//! use html_to_markdown_rs::{convert_with_metadata, MetadataConfig};
+//! use html_to_markdown_rs::{convert, ConversionOptions};
 //!
-//! let config = MetadataConfig {
-//!     extract_headers: true,
-//!     extract_links: true,
-//!     extract_images: false,  // Skip images
-//!     extract_structured_data: false,  // Skip structured data
-//!     max_structured_data_size: 0,
+//! let options = ConversionOptions {
+//!     extract_metadata: false,  // Disable metadata extraction
+//!     ..Default::default()
 //! };
 //!
-//! let (markdown, metadata) = convert_with_metadata(html, None, config)?;
-//! assert_eq!(metadata.images.len(), 0);  // Images not extracted
+//! let result = convert(html, Some(options))?;
+//! assert!(result.metadata.is_none());  // Metadata not extracted
 //! # Ok::<(), html_to_markdown_rs::ConversionError>(())
 //! ```
 //!
 //! ## Analyzing Link Types
 //!
 //! ```ignore
-//! use html_to_markdown_rs::{convert_with_metadata, MetadataConfig};
+//! use html_to_markdown_rs::convert;
 //! use html_to_markdown_rs::metadata::LinkType;
 //!
-//! let (_markdown, metadata) = convert_with_metadata(html, None, MetadataConfig::default())?;
+//! let result = convert(html, None)?;
+//! let metadata = result.metadata.unwrap();
 //!
 //! for link in &metadata.links {
 //!     match link.link_type {
@@ -129,11 +127,13 @@
 //! This enables easy export to JSON, YAML, or other formats:
 //!
 //! ```ignore
-//! use html_to_markdown_rs::{convert_with_metadata, MetadataConfig};
+//! use html_to_markdown_rs::convert;
 //!
-//! let (_markdown, metadata) = convert_with_metadata(html, None, MetadataConfig::default())?;
-//! let json = serde_json::to_string_pretty(&metadata)?;
-//! println!("{}", json);
+//! let result = convert(html, None)?;
+//! if let Some(metadata) = &result.metadata {
+//!     let json = serde_json::to_string_pretty(metadata)?;
+//!     println!("{}", json);
+//! }
 //! # Ok::<(), Box<dyn std::error::Error>>(())
 //! ```
@@ -146,8 +146,8 @@ pub mod types;
 pub use collector::MetadataCollector;
 pub use config::{DEFAULT_MAX_STRUCTURED_DATA_SIZE, MetadataConfig, MetadataConfigUpdate};
 pub use types::{
-    DocumentMetadata, ExtendedMetadata, HeaderMetadata, ImageMetadata, ImageType, LinkMetadata, LinkType,
-    StructuredData, StructuredDataType, TextDirection,
+    DocumentMetadata, HeaderMetadata, HtmlMetadata, ImageMetadata, ImageType, LinkMetadata, LinkType, StructuredData,
+    StructuredDataType, TextDirection,
 };
 // Internal handle type for shared mutable access during tree traversal

data/vendor/html-to-markdown-rs/src/metadata/types.rs CHANGED Viewed

@@ -446,8 +446,8 @@ pub struct StructuredData {
 /// # Examples
 ///
 /// ```
-/// # use html_to_markdown_rs::metadata::ExtendedMetadata;
-/// let metadata = ExtendedMetadata {
+/// # use html_to_markdown_rs::metadata::HtmlMetadata;
+/// let metadata = HtmlMetadata {
 ///     document: Default::default(),
 ///     headers: Vec::new(),
 ///     links: Vec::new(),
@@ -459,7 +459,7 @@ pub struct StructuredData {
 /// ```
 #[derive(Debug, Clone, Default)]
 #[cfg_attr(feature = "metadata", derive(serde::Serialize, serde::Deserialize))]
-pub struct ExtendedMetadata {
+pub struct HtmlMetadata {
     /// Document-level metadata (title, description, canonical, etc.)
     pub document: DocumentMetadata,