html-to-markdown 3.4.0 → 3.6.0.pre.rc.23

data/vendor/html-to-markdown-rs/src/visitor/types.rs

@@ -1,319 +0,0 @@
-//! Type definitions for the visitor pattern.
-//!
-//! This module contains the core data types used in the visitor pattern:
-//! - `NodeType`: Categorization of HTML elements
-//! - `NodeContext`: Metadata about the current node being visited
-//! - `VisitResult`: Control flow signals from visitor methods
-
-use std::collections::BTreeMap;
-
-/// Node type enumeration covering all HTML element types.
-///
-/// This enum categorizes all HTML elements that the converter recognizes,
-/// providing a coarse-grained classification for visitor dispatch.
-#[derive(Debug, Clone, Copy, PartialEq, Eq, Hash)]
-#[cfg_attr(feature = "serde", derive(serde::Serialize, serde::Deserialize))]
-pub enum NodeType {
-    /// Text node (most frequent - 100+ per document)
-    Text,
-
-    /// Generic element node
-    Element,
-
-    /// Heading elements (h1-h6)
-    Heading,
-    /// Paragraph element
-    Paragraph,
-    /// Generic div container
-    Div,
-    /// Blockquote element
-    Blockquote,
-    /// Preformatted text block
-    Pre,
-    /// Horizontal rule
-    Hr,
-
-    /// Ordered or unordered list (ul, ol)
-    List,
-    /// List item (li)
-    ListItem,
-    /// Definition list (dl)
-    DefinitionList,
-    /// Definition term (dt)
-    DefinitionTerm,
-    /// Definition description (dd)
-    DefinitionDescription,
-
-    /// Table element
-    Table,
-    /// Table row (tr)
-    TableRow,
-    /// Table cell (td, th)
-    TableCell,
-    /// Table header cell (th)
-    TableHeader,
-    /// Table body (tbody)
-    TableBody,
-    /// Table head (thead)
-    TableHead,
-    /// Table foot (tfoot)
-    TableFoot,
-
-    /// Anchor link (a)
-    Link,
-    /// Image (img)
-    Image,
-    /// Strong/bold (strong, b)
-    Strong,
-    /// Emphasis/italic (em, i)
-    Em,
-    /// Inline code (code)
-    Code,
-    /// Strikethrough (s, del, strike)
-    Strikethrough,
-    /// Underline (u, ins)
-    Underline,
-    /// Subscript (sub)
-    Subscript,
-    /// Superscript (sup)
-    Superscript,
-    /// Mark/highlight (mark)
-    Mark,
-    /// Small text (small)
-    Small,
-    /// Line break (br)
-    Br,
-    /// Span element
-    Span,
-
-    /// Article element
-    Article,
-    /// Section element
-    Section,
-    /// Navigation element
-    Nav,
-    /// Aside element
-    Aside,
-    /// Header element
-    Header,
-    /// Footer element
-    Footer,
-    /// Main element
-    Main,
-    /// Figure element
-    Figure,
-    /// Figure caption
-    Figcaption,
-    /// Time element
-    Time,
-    /// Details element
-    Details,
-    /// Summary element
-    Summary,
-
-    /// Form element
-    Form,
-    /// Input element
-    Input,
-    /// Select element
-    Select,
-    /// Option element
-    Option,
-    /// Button element
-    Button,
-    /// Textarea element
-    Textarea,
-    /// Label element
-    Label,
-    /// Fieldset element
-    Fieldset,
-    /// Legend element
-    Legend,
-
-    /// Audio element
-    Audio,
-    /// Video element
-    Video,
-    /// Picture element
-    Picture,
-    /// Source element
-    Source,
-    /// Iframe element
-    Iframe,
-    /// SVG element
-    Svg,
-    /// Canvas element
-    Canvas,
-
-    /// Ruby annotation
-    Ruby,
-    /// Ruby text
-    Rt,
-    /// Ruby parenthesis
-    Rp,
-    /// Abbreviation
-    Abbr,
-    /// Keyboard input
-    Kbd,
-    /// Sample output
-    Samp,
-    /// Variable
-    Var,
-    /// Citation
-    Cite,
-    /// Quote
-    Q,
-    /// Deleted text
-    Del,
-    /// Inserted text
-    Ins,
-    /// Data element
-    Data,
-    /// Meter element
-    Meter,
-    /// Progress element
-    Progress,
-    /// Output element
-    Output,
-    /// Template element
-    Template,
-    /// Slot element
-    Slot,
-
-    /// HTML root element
-    Html,
-    /// Head element
-    Head,
-    /// Body element
-    Body,
-    /// Title element
-    Title,
-    /// Meta element
-    Meta,
-    /// Link element (not anchor)
-    LinkTag,
-    /// Style element
-    Style,
-    /// Script element
-    Script,
-    /// Base element
-    Base,
-
-    /// Custom element (web components) or unknown tag
-    Custom,
-}
-
-/// Context information passed to all visitor methods.
-///
-/// Provides comprehensive metadata about the current node being visited,
-/// including its type, attributes, position in the DOM tree, and parent context.
-#[derive(Debug, Clone)]
-#[cfg_attr(feature = "serde", derive(serde::Serialize, serde::Deserialize))]
-pub struct NodeContext {
-    /// Coarse-grained node type classification
-    pub node_type: NodeType,
-
-    /// Raw HTML tag name (e.g., "div", "h1", "custom-element")
-    pub tag_name: String,
-
-    /// All HTML attributes as key-value pairs
-    pub attributes: BTreeMap<String, String>,
-
-    /// Depth in the DOM tree (0 = root)
-    pub depth: usize,
-
-    /// Index among siblings (0-based)
-    pub index_in_parent: usize,
-
-    /// Parent element's tag name (None if root)
-    pub parent_tag: Option<String>,
-
-    /// Whether this element is treated as inline vs block
-    pub is_inline: bool,
-}
-
-/// Result of a visitor callback.
-///
-/// Allows visitors to control the conversion flow by either proceeding
-/// with default behavior, providing custom output, skipping elements,
-/// preserving HTML, or signaling errors.
-#[derive(Debug, Clone, Default)]
-#[cfg_attr(feature = "serde", derive(serde::Serialize, serde::Deserialize))]
-pub enum VisitResult {
-    #[default]
-    /// Continue with default conversion behavior
-    Continue,
-
-    /// Replace default output with custom markdown
-    ///
-    /// The visitor takes full responsibility for the markdown output
-    /// of this node and its children.
-    Custom(String),
-
-    /// Skip this element entirely (don't output anything)
-    ///
-    /// The element and all its children are ignored in the output.
-    Skip,
-
-    /// Preserve original HTML (don't convert to markdown)
-    ///
-    /// The element's raw HTML is included verbatim in the output.
-    PreserveHtml,
-
-    /// Stop conversion with an error
-    ///
-    /// The conversion process halts and returns this error message.
-    Error(String),
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-
-    #[test]
-    fn test_node_type_equality() {
-        assert_eq!(NodeType::Text, NodeType::Text);
-        assert_ne!(NodeType::Text, NodeType::Element);
-        assert_eq!(NodeType::Heading, NodeType::Heading);
-    }
-
-    #[test]
-    fn test_node_context_creation() {
-        let ctx = NodeContext {
-            node_type: NodeType::Heading,
-            tag_name: "h1".to_string(),
-            attributes: BTreeMap::new(),
-            depth: 1,
-            index_in_parent: 0,
-            parent_tag: Some("body".to_string()),
-            is_inline: false,
-        };
-
-        assert_eq!(ctx.node_type, NodeType::Heading);
-        assert_eq!(ctx.tag_name, "h1");
-        assert_eq!(ctx.depth, 1);
-        assert!(!ctx.is_inline);
-    }
-
-    #[test]
-    fn test_visit_result_variants() {
-        let continue_result = VisitResult::Continue;
-        matches!(continue_result, VisitResult::Continue);
-
-        let custom_result = VisitResult::Custom("# Custom Output".to_string());
-        if let VisitResult::Custom(output) = custom_result {
-            assert_eq!(output, "# Custom Output");
-        }
-
-        let skip_result = VisitResult::Skip;
-        matches!(skip_result, VisitResult::Skip);
-
-        let preserve_result = VisitResult::PreserveHtml;
-        matches!(preserve_result, VisitResult::PreserveHtml);
-
-        let error_result = VisitResult::Error("Test error".to_string());
-        if let VisitResult::Error(msg) = error_result {
-            assert_eq!(msg, "Test error");
-        }
-    }
-}

data/vendor/html-to-markdown-rs/src/visitor_helpers/helpers/callbacks/mod.rs

@@ -1 +0,0 @@
-//! Callback management for visitor pattern.

data/vendor/html-to-markdown-rs/src/visitor_helpers/helpers/content.rs

@@ -1,126 +0,0 @@
-//! Content extraction and result handling.
-//!
-//! This module provides the `VisitorDispatch` enum and helper methods for
-//! processing the results of visitor callbacks.
-
-/// Result of dispatching a visitor callback.
-///
-/// This enum represents the outcome of a visitor callback dispatch,
-/// providing a more ergonomic interface for control flow than the
-/// raw `VisitResult` type.
-#[allow(dead_code)]
-#[derive(Debug)]
-pub enum VisitorDispatch {
-    /// Continue with default conversion behavior
-    Continue,
-
-    /// Replace default output with custom markdown
-    Custom(String),
-
-    /// Skip this element entirely (don't output anything)
-    Skip,
-
-    /// Preserve original HTML (don't convert to markdown)
-    PreserveHtml,
-}
-
-impl VisitorDispatch {
-    /// Check if this dispatch result indicates continuation.
-    #[allow(dead_code)]
-    #[inline]
-    #[must_use]
-    pub const fn is_continue(&self) -> bool {
-        matches!(self, Self::Continue)
-    }
-
-    /// Check if this dispatch result contains custom output.
-    #[allow(dead_code)]
-    #[inline]
-    #[must_use]
-    pub const fn is_custom(&self) -> bool {
-        matches!(self, Self::Custom(_))
-    }
-
-    /// Check if this dispatch result indicates skipping.
-    #[allow(dead_code)]
-    #[inline]
-    #[must_use]
-    pub const fn is_skip(&self) -> bool {
-        matches!(self, Self::Skip)
-    }
-
-    /// Check if this dispatch result indicates HTML preservation.
-    #[allow(dead_code)]
-    #[inline]
-    #[must_use]
-    pub const fn is_preserve_html(&self) -> bool {
-        matches!(self, Self::PreserveHtml)
-    }
-
-    /// Extract custom output if present.
-    #[allow(dead_code)]
-    #[inline]
-    #[must_use]
-    pub fn into_custom(self) -> Option<String> {
-        match self {
-            Self::Custom(output) => Some(output),
-            _ => None,
-        }
-    }
-
-    /// Extract custom output reference if present.
-    #[allow(dead_code)]
-    #[inline]
-    #[must_use]
-    pub fn as_custom(&self) -> Option<&str> {
-        match self {
-            Self::Custom(output) => Some(output),
-            _ => None,
-        }
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-
-    #[test]
-    fn test_visitor_dispatch_predicates() {
-        let continue_dispatch = VisitorDispatch::Continue;
-        assert!(continue_dispatch.is_continue());
-        assert!(!continue_dispatch.is_custom());
-        assert!(!continue_dispatch.is_skip());
-
-        let custom_dispatch = VisitorDispatch::Custom("output".to_string());
-        assert!(!custom_dispatch.is_continue());
-        assert!(custom_dispatch.is_custom());
-        assert!(!custom_dispatch.is_skip());
-
-        let skip_dispatch = VisitorDispatch::Skip;
-        assert!(!skip_dispatch.is_continue());
-        assert!(!skip_dispatch.is_custom());
-        assert!(skip_dispatch.is_skip());
-
-        let preserve_dispatch = VisitorDispatch::PreserveHtml;
-        assert!(!preserve_dispatch.is_continue());
-        assert!(preserve_dispatch.is_preserve_html());
-    }
-
-    #[test]
-    fn test_visitor_dispatch_into_custom() {
-        let custom = VisitorDispatch::Custom("test".to_string());
-        assert_eq!(custom.into_custom(), Some("test".to_string()));
-
-        let continue_dispatch = VisitorDispatch::Continue;
-        assert_eq!(continue_dispatch.into_custom(), None);
-    }
-
-    #[test]
-    fn test_visitor_dispatch_as_custom() {
-        let custom = VisitorDispatch::Custom("test".to_string());
-        assert_eq!(custom.as_custom(), Some("test"));
-
-        let skip = VisitorDispatch::Skip;
-        assert_eq!(skip.as_custom(), None);
-    }
-}

data/vendor/html-to-markdown-rs/src/visitor_helpers/helpers/mod.rs

@@ -1,27 +0,0 @@
-//! Helper functions for visitor pattern integration.
-//!
-//! This module provides efficient utilities for building visitor contexts,
-//! dispatching visitor callbacks, and handling visitor results during the
-//! HTML→Markdown conversion process.
-//!
-//! # Design Goals
-//!
-//! - **Zero allocation when possible**: Reuse existing data structures
-//! - **Minimal overhead**: Inline hot paths, avoid unnecessary clones
-//! - **Type safety**: Leverage Rust's type system for correct visitor handling
-//! - **Ergonomics**: Reduce boilerplate for common visitor patterns
-//!
-//! # Usage
-//!
-//! These helpers are designed to be used within the converter module during
-//! the DOM traversal. They bridge the gap between the internal conversion
-//! state and the public visitor API.
-
-pub mod callbacks;
-pub mod content;
-pub mod state;
-pub mod traversal;
-
-pub use content::VisitorDispatch;
-pub use state::build_node_context;
-pub use traversal::dispatch_visitor;

data/vendor/html-to-markdown-rs/src/visitor_helpers/helpers/state.rs

@@ -1,110 +0,0 @@
-//! Visitor state management and context building.
-//!
-//! This module handles construction of `NodeContext` objects that represent
-//! the state of DOM nodes during traversal.
-
-use std::collections::BTreeMap;
-
-use crate::visitor::NodeContext;
-use crate::visitor::NodeType;
-
-/// Build a `NodeContext` from current parsing state.
-///
-/// Creates a complete `NodeContext` suitable for passing to visitor callbacks.
-/// This function collects metadata about the current node from various sources:
-/// - Tag name and attributes from the HTML element
-/// - Depth and parent information from the DOM tree
-/// - Index among siblings for positional awareness
-/// - Inline/block classification
-///
-/// # Parameters
-///
-/// - `node_type`: Coarse-grained classification (Link, Image, Heading, etc.)
-/// - `tag_name`: Raw HTML tag name (e.g., "div", "h1", "custom-element")
-/// - `attributes`: All HTML attributes as key-value pairs
-/// - `depth`: Nesting depth in the DOM tree (0 = root)
-/// - `index_in_parent`: Zero-based index among siblings
-/// - `parent_tag`: Parent element's tag name (None if root)
-/// - `is_inline`: Whether this element is treated as inline vs block
-///
-/// # Returns
-///
-/// A fully populated `NodeContext` ready for visitor dispatch.
-///
-/// # Performance
-///
-/// This function performs minimal allocations:
-/// - Clones `tag_name` (typically 2-10 bytes)
-/// - Clones `parent_tag` if present (typically 2-10 bytes)
-/// - Clones the attributes `BTreeMap` (heap allocation if non-empty)
-///
-/// For text nodes and simple elements without attributes, allocations are minimal.
-///
-/// # Examples
-///
-/// ```text
-/// let ctx = build_node_context(
-///     NodeType::Heading,
-///     "h1",
-///     &attrs,
-///     1,
-///     0,
-///     Some("body"),
-///     false,
-/// );
-/// ```
-#[allow(dead_code)]
-#[inline]
-pub fn build_node_context(
-    node_type: NodeType,
-    tag_name: &str,
-    attributes: &BTreeMap<String, String>,
-    depth: usize,
-    index_in_parent: usize,
-    parent_tag: Option<&str>,
-    is_inline: bool,
-) -> NodeContext {
-    NodeContext {
-        node_type,
-        tag_name: tag_name.to_string(),
-        attributes: attributes.clone(),
-        depth,
-        index_in_parent,
-        parent_tag: parent_tag.map(String::from),
-        is_inline,
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use super::*;
-
-    #[test]
-    fn test_build_node_context() {
-        let mut attrs = BTreeMap::new();
-        attrs.insert("id".to_string(), "main".to_string());
-        attrs.insert("class".to_string(), "container".to_string());
-
-        let ctx = build_node_context(NodeType::Div, "div", &attrs, 2, 3, Some("body"), false);
-
-        assert_eq!(ctx.node_type, NodeType::Div);
-        assert_eq!(ctx.tag_name, "div");
-        assert_eq!(ctx.depth, 2);
-        assert_eq!(ctx.index_in_parent, 3);
-        assert_eq!(ctx.parent_tag, Some("body".to_string()));
-        assert!(!ctx.is_inline);
-        assert_eq!(ctx.attributes.len(), 2);
-        assert_eq!(ctx.attributes.get("id"), Some(&"main".to_string()));
-    }
-
-    #[test]
-    fn test_build_node_context_no_parent() {
-        let attrs = BTreeMap::new();
-
-        let ctx = build_node_context(NodeType::Html, "html", &attrs, 0, 0, None, false);
-
-        assert_eq!(ctx.node_type, NodeType::Html);
-        assert_eq!(ctx.parent_tag, None);
-        assert!(ctx.attributes.is_empty());
-    }
-}