html-to-markdown 3.10.6 → 3.11.2

data/lib/html_to_markdown/native.rb

@@ -1,7 +1,7 @@
 # This file is auto-generated by alef — DO NOT EDIT.
-# alef:hash:f44000653901d10b61e9069d2e901a11b70b0e643ba63d75f2222729c924ae82
+# alef:hash:7f3c74ddad62b174cb8ff3a57759bff2fc8e3af68b56501137ccef0857114d98
 # To regenerate: alef generate
-# To verify freshness: alef verify --exit-code
+# To verify freshness: alef verify
 # frozen_string_literal: true
 
 require "json"
@@ -18,7 +18,7 @@ library_candidates = dlexts.flat_map do |dlext|
     File.join(__dir__, "..", extension_name, ruby_abi, "#{extension_name}.#{dlext}"),
     File.join(__dir__, "..", extension_name, ruby_abi, "lib#{extension_name}.#{dlext}"),
     File.join(__dir__, "..", "#{extension_name}.#{dlext}"),
-    File.join(__dir__, "..", "lib#{extension_name}.#{dlext}")
+    File.join(__dir__, "..", "lib#{extension_name}.#{dlext}"),
   ]
 end
 
@@ -34,7 +34,6 @@ unless native_extension
 end
 
 require native_extension
-
 module HtmlToMarkdown
   # Re-export public types from the native extension (curated list, excludes Update/Builder types)
   ConversionOptions = HtmlToMarkdownRs.const_get(:ConversionOptions)
@@ -57,11 +56,8 @@ module HtmlToMarkdown
   TableGrid = HtmlToMarkdownRs.const_get(:TableGrid)
   TextAnnotation = HtmlToMarkdownRs.const_get(:TextAnnotation)
   # Re-export public module functions from the native extension (curated list)
-  define_singleton_method(:convert) { |*args, **kwargs, &blk|
-    HtmlToMarkdownRs.public_send(:convert, *args, **kwargs, &blk)
-  }
+  define_singleton_method(:convert) { |*args, **kwargs, &blk| HtmlToMarkdownRs.public_send(:convert, *args, **kwargs, &blk) }
 end
-
 module HtmlToMarkdown
   # The semantic content type of a document node.
   #
@@ -79,34 +75,20 @@ module HtmlToMarkdown
     def self.from_hash(hash)
       discriminator = hash[:node_type] || hash["node_type"]
       case discriminator
-      when "heading"
-        NodeContentHeading.from_hash(hash)
-      when "paragraph"
-        NodeContentParagraph.from_hash(hash)
-      when "list"
-        NodeContentList.from_hash(hash)
-      when "list_item"
-        NodeContentListItem.from_hash(hash)
-      when "table"
-        NodeContentTable.from_hash(hash)
-      when "image"
-        NodeContentImage.from_hash(hash)
-      when "code"
-        NodeContentCode.from_hash(hash)
-      when "quote"
-        NodeContentQuote.from_hash(hash)
-      when "definition_list"
-        NodeContentDefinitionList.from_hash(hash)
-      when "definition_item"
-        NodeContentDefinitionItem.from_hash(hash)
-      when "raw_block"
-        NodeContentRawBlock.from_hash(hash)
-      when "metadata_block"
-        NodeContentMetadataBlock.from_hash(hash)
-      when "group"
-        NodeContentGroup.from_hash(hash)
-      else
-        raise "Unknown discriminator: #{discriminator}"
+      when "heading" then NodeContentHeading.from_hash(hash)
+      when "paragraph" then NodeContentParagraph.from_hash(hash)
+      when "list" then NodeContentList.from_hash(hash)
+      when "list_item" then NodeContentListItem.from_hash(hash)
+      when "table" then NodeContentTable.from_hash(hash)
+      when "image" then NodeContentImage.from_hash(hash)
+      when "code" then NodeContentCode.from_hash(hash)
+      when "quote" then NodeContentQuote.from_hash(hash)
+      when "definition_list" then NodeContentDefinitionList.from_hash(hash)
+      when "definition_item" then NodeContentDefinitionItem.from_hash(hash)
+      when "raw_block" then NodeContentRawBlock.from_hash(hash)
+      when "metadata_block" then NodeContentMetadataBlock.from_hash(hash)
+      when "group" then NodeContentGroup.from_hash(hash)
+      else raise "Unknown discriminator: #{discriminator}"
       end
     end
   end
@@ -117,12 +99,10 @@ module HtmlToMarkdown
 
     # Heading level (1-6).
     sig { returns(Integer) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def level = super
+    def level = super # rubocop:disable Lint/UselessMethodDefinition
     # The heading text content.
     sig { returns(String) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def text = super
+    def text = super # rubocop:disable Lint/UselessMethodDefinition
     sig { returns(T::Boolean) }
     def heading? = true
     sig { returns(T::Boolean) }
@@ -163,8 +143,7 @@ module HtmlToMarkdown
 
     # The paragraph text content.
     sig { returns(String) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def text = super
+    def text = super # rubocop:disable Lint/UselessMethodDefinition
     sig { returns(T::Boolean) }
     def heading? = false
     sig { returns(T::Boolean) }
@@ -205,8 +184,7 @@ module HtmlToMarkdown
 
     # Whether this is an ordered list.
     sig { returns(T::Boolean) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def ordered = super
+    def ordered = super # rubocop:disable Lint/UselessMethodDefinition
     sig { returns(T::Boolean) }
     def heading? = false
     sig { returns(T::Boolean) }
@@ -247,8 +225,7 @@ module HtmlToMarkdown
 
     # The list item text content.
     sig { returns(String) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def text = super
+    def text = super # rubocop:disable Lint/UselessMethodDefinition
     sig { returns(T::Boolean) }
     def heading? = false
     sig { returns(T::Boolean) }
@@ -289,8 +266,7 @@ module HtmlToMarkdown
 
     # The table grid structure.
     sig { returns(TableGrid) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def grid = super
+    def grid = super # rubocop:disable Lint/UselessMethodDefinition
     sig { returns(T::Boolean) }
     def heading? = false
     sig { returns(T::Boolean) }
@@ -331,16 +307,13 @@ module HtmlToMarkdown
 
     # Alt text or caption.
     sig { returns(T.nilable(String)) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def description = super
+    def description = super # rubocop:disable Lint/UselessMethodDefinition
     # Image source URL.
     sig { returns(T.nilable(String)) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def src = super
+    def src = super # rubocop:disable Lint/UselessMethodDefinition
     # Index into `ConversionResult.images` when image extraction is enabled.
     sig { returns(T.nilable(Integer)) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def image_index = super
+    def image_index = super # rubocop:disable Lint/UselessMethodDefinition
     sig { returns(T::Boolean) }
     def heading? = false
     sig { returns(T::Boolean) }
@@ -371,11 +344,7 @@ module HtmlToMarkdown
     # @return [self]
     sig { params(hash: T::Hash[T.untyped, T.untyped]).returns(T.attached_class) }
     def self.from_hash(hash)
-      new(
-        description: hash[:description] || hash["description"],
-        src: hash[:src] || hash["src"],
-        image_index: hash[:image_index] || hash["image_index"]
-      )
+      new(description: hash[:description] || hash["description"], src: hash[:src] || hash["src"], image_index: hash[:image_index] || hash["image_index"])
     end
   end
   ## A code block or inline code.
@@ -385,12 +354,10 @@ module HtmlToMarkdown
 
     # The code text content.
     sig { returns(String) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def text = super
+    def text = super # rubocop:disable Lint/UselessMethodDefinition
     # Programming language (from class="language-*" or similar).
     sig { returns(T.nilable(String)) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def language = super
+    def language = super # rubocop:disable Lint/UselessMethodDefinition
     sig { returns(T::Boolean) }
     def heading? = false
     sig { returns(T::Boolean) }
@@ -507,12 +474,10 @@ module HtmlToMarkdown
 
     # The term being defined.
     sig { returns(String) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def term = super
+    def term = super # rubocop:disable Lint/UselessMethodDefinition
     # The definition text.
     sig { returns(String) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def definition = super
+    def definition = super # rubocop:disable Lint/UselessMethodDefinition
     sig { returns(T::Boolean) }
     def heading? = false
     sig { returns(T::Boolean) }
@@ -553,12 +518,10 @@ module HtmlToMarkdown
 
     # The format of the raw content (e.g. "html", "css", "javascript").
     sig { returns(String) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def format = super
+    def format = super # rubocop:disable Lint/UselessMethodDefinition
     # The raw content.
     sig { returns(String) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def content = super
+    def content = super # rubocop:disable Lint/UselessMethodDefinition
     sig { returns(T::Boolean) }
     def heading? = false
     sig { returns(T::Boolean) }
@@ -599,8 +562,7 @@ module HtmlToMarkdown
 
     # Key-value metadata pairs.
     sig { returns(T::Array[MetadataEntry]) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def entries = super
+    def entries = super # rubocop:disable Lint/UselessMethodDefinition
     sig { returns(T::Boolean) }
     def heading? = false
     sig { returns(T::Boolean) }
@@ -641,16 +603,13 @@ module HtmlToMarkdown
 
     # Optional section label.
     sig { returns(T.nilable(String)) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def label = super
+    def label = super # rubocop:disable Lint/UselessMethodDefinition
     # The heading level that created this group.
     sig { returns(T.nilable(Integer)) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def heading_level = super
+    def heading_level = super # rubocop:disable Lint/UselessMethodDefinition
     # The heading text that created this group.
     sig { returns(T.nilable(String)) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def heading_text = super
+    def heading_text = super # rubocop:disable Lint/UselessMethodDefinition
     sig { returns(T::Boolean) }
     def heading? = false
     sig { returns(T::Boolean) }
@@ -681,11 +640,7 @@ module HtmlToMarkdown
     # @return [self]
     sig { params(hash: T::Hash[T.untyped, T.untyped]).returns(T.attached_class) }
     def self.from_hash(hash)
-      new(
-        label: hash[:label] || hash["label"],
-        heading_level: hash[:heading_level] || hash["heading_level"],
-        heading_text: hash[:heading_text] || hash["heading_text"]
-      )
+      new(label: hash[:label] || hash["label"], heading_level: hash[:heading_level] || hash["heading_level"], heading_text: hash[:heading_text] || hash["heading_text"])
     end
   end
 end
@@ -707,26 +662,16 @@ module HtmlToMarkdown
     def self.from_hash(hash)
       discriminator = hash[:annotation_type] || hash["annotation_type"]
       case discriminator
-      when "bold"
-        AnnotationKindBold.from_hash(hash)
-      when "italic"
-        AnnotationKindItalic.from_hash(hash)
-      when "underline"
-        AnnotationKindUnderline.from_hash(hash)
-      when "strikethrough"
-        AnnotationKindStrikethrough.from_hash(hash)
-      when "code"
-        AnnotationKindCode.from_hash(hash)
-      when "subscript"
-        AnnotationKindSubscript.from_hash(hash)
-      when "superscript"
-        AnnotationKindSuperscript.from_hash(hash)
-      when "highlight"
-        AnnotationKindHighlight.from_hash(hash)
-      when "link"
-        AnnotationKindLink.from_hash(hash)
-      else
-        raise "Unknown discriminator: #{discriminator}"
+      when "bold" then AnnotationKindBold.from_hash(hash)
+      when "italic" then AnnotationKindItalic.from_hash(hash)
+      when "underline" then AnnotationKindUnderline.from_hash(hash)
+      when "strikethrough" then AnnotationKindStrikethrough.from_hash(hash)
+      when "code" then AnnotationKindCode.from_hash(hash)
+      when "subscript" then AnnotationKindSubscript.from_hash(hash)
+      when "superscript" then AnnotationKindSuperscript.from_hash(hash)
+      when "highlight" then AnnotationKindHighlight.from_hash(hash)
+      when "link" then AnnotationKindLink.from_hash(hash)
+      else raise "Unknown discriminator: #{discriminator}"
       end
     end
   end
@@ -982,15 +927,13 @@ module HtmlToMarkdown
     # written in the source. Callers that need an absolute URL must resolve it against the
     # document base URL themselves.
     sig { returns(String) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def url = super
+    def url = super # rubocop:disable Lint/UselessMethodDefinition
     # The `title` attribute of the `<a>` element, if present.
     #
     # `None` when the `<a>` tag has no `title="..."` attribute. When present, the value
     # is copied verbatim — HTML entities within the title are not decoded.
     sig { returns(T.nilable(String)) }
-    # rubocop:disable Lint/UselessMethodDefinition
-    def title = super
+    def title = super # rubocop:disable Lint/UselessMethodDefinition
     sig { returns(T::Boolean) }
     def bold? = false
     sig { returns(T::Boolean) }
@@ -1017,3 +960,158 @@ module HtmlToMarkdown
     end
   end
 end
+
+module HtmlToMarkdown
+  # 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.
+  module VisitResult
+    extend T::Helpers
+    extend T::Sig
+
+    interface!
+
+    # Dispatch from a Hash to the appropriate variant constructor.
+    # @param hash [Hash] with discriminator field and variant-specific fields
+    # @return [variant_class] an instance of the appropriate variant
+    sig { params(hash: T::Hash[T.untyped, T.untyped]).returns(T.untyped) }
+    def self.from_hash(hash)
+      discriminator = hash[:type] || hash["type"]
+      case discriminator
+      when "continue" then VisitResultContinue.from_hash(hash)
+      when "custom" then VisitResultCustom.from_hash(hash)
+      when "skip" then VisitResultSkip.from_hash(hash)
+      when "preserve_html" then VisitResultPreserveHtml.from_hash(hash)
+      when "error" then VisitResultError.from_hash(hash)
+      else raise "Unknown discriminator: #{discriminator}"
+      end
+    end
+  end
+  ## Continue with default conversion behavior
+  VisitResultContinue = Data.define do
+    include VisitResult
+    extend T::Sig
+
+    sig { returns(T::Boolean) }
+    def continue? = true
+    sig { returns(T::Boolean) }
+    def custom? = false
+    sig { returns(T::Boolean) }
+    def skip? = false
+    sig { returns(T::Boolean) }
+    def preserve_html? = false
+    sig { returns(T::Boolean) }
+    def error? = false
+    # @param hash [Hash] deserialized from the native extension
+    # @return [self]
+    sig { params(hash: T::Hash[T.untyped, T.untyped]).returns(T.attached_class) }
+    def self.from_hash(hash)
+      new
+    end
+  end
+  ## Replace default output with custom markdown
+  ##
+  ## The visitor takes full responsibility for the markdown output
+  ## of this node and its children.
+  VisitResultCustom = Data.define(:value) do
+    include VisitResult
+    extend T::Sig
+
+    # @return [String]
+    sig { returns(String) }
+    def value = super # rubocop:disable Lint/UselessMethodDefinition
+    sig { returns(T::Boolean) }
+    def continue? = false
+    sig { returns(T::Boolean) }
+    def custom? = true
+    sig { returns(T::Boolean) }
+    def skip? = false
+    sig { returns(T::Boolean) }
+    def preserve_html? = false
+    sig { returns(T::Boolean) }
+    def error? = false
+    # @param hash [Hash] deserialized from the native extension
+    # @return [self]
+    sig { params(hash: T::Hash[T.untyped, T.untyped]).returns(T.attached_class) }
+    def self.from_hash(hash)
+      new(value: hash[:_0] || hash["_0"])
+    end
+  end
+  ## Skip this element entirely (don't output anything)
+  ##
+  ## The element and all its children are ignored in the output.
+  VisitResultSkip = Data.define do
+    include VisitResult
+    extend T::Sig
+
+    sig { returns(T::Boolean) }
+    def continue? = false
+    sig { returns(T::Boolean) }
+    def custom? = false
+    sig { returns(T::Boolean) }
+    def skip? = true
+    sig { returns(T::Boolean) }
+    def preserve_html? = false
+    sig { returns(T::Boolean) }
+    def error? = false
+    # @param hash [Hash] deserialized from the native extension
+    # @return [self]
+    sig { params(hash: T::Hash[T.untyped, T.untyped]).returns(T.attached_class) }
+    def self.from_hash(hash)
+      new
+    end
+  end
+  ## Preserve original HTML (don't convert to markdown)
+  ##
+  ## The element's raw HTML is included verbatim in the output.
+  VisitResultPreserveHtml = Data.define do
+    include VisitResult
+    extend T::Sig
+
+    sig { returns(T::Boolean) }
+    def continue? = false
+    sig { returns(T::Boolean) }
+    def custom? = false
+    sig { returns(T::Boolean) }
+    def skip? = false
+    sig { returns(T::Boolean) }
+    def preserve_html? = true
+    sig { returns(T::Boolean) }
+    def error? = false
+    # @param hash [Hash] deserialized from the native extension
+    # @return [self]
+    sig { params(hash: T::Hash[T.untyped, T.untyped]).returns(T.attached_class) }
+    def self.from_hash(hash)
+      new
+    end
+  end
+  ## Stop conversion with an error
+  ##
+  ## The conversion process halts and returns this error message.
+  VisitResultError = Data.define(:value) do
+    include VisitResult
+    extend T::Sig
+
+    # @return [String]
+    sig { returns(String) }
+    def value = super # rubocop:disable Lint/UselessMethodDefinition
+    sig { returns(T::Boolean) }
+    def continue? = false
+    sig { returns(T::Boolean) }
+    def custom? = false
+    sig { returns(T::Boolean) }
+    def skip? = false
+    sig { returns(T::Boolean) }
+    def preserve_html? = false
+    sig { returns(T::Boolean) }
+    def error? = true
+    # @param hash [Hash] deserialized from the native extension
+    # @return [self]
+    sig { params(hash: T::Hash[T.untyped, T.untyped]).returns(T.attached_class) }
+    def self.from_hash(hash)
+      new(value: hash[:_0] || hash["_0"])
+    end
+  end
+end

data/lib/html_to_markdown/version.rb

@@ -1,10 +1,10 @@
 # This file is auto-generated by alef — DO NOT EDIT.
-# alef:hash:f44000653901d10b61e9069d2e901a11b70b0e643ba63d75f2222729c924ae82
+# alef:hash:b91f80d853da057a497660797e08ed766f3540e488f2c58b16c5f100dd6d8f04
 # To regenerate: alef generate
-# To verify freshness: alef verify --exit-code
+# To verify freshness: alef verify
 # frozen_string_literal: true
 
 module HtmlToMarkdown
   ## The version string for this package.
-  VERSION = "3.10.6"
+  VERSION = "3.11.2"
 end

data/lib/html_to_markdown.rb

@@ -1,7 +1,7 @@
 # This file is auto-generated by alef — DO NOT EDIT.
-# alef:hash:f44000653901d10b61e9069d2e901a11b70b0e643ba63d75f2222729c924ae82
+# alef:hash:cc885f84ff886e0b8811b1fd3b76314a149b879662c55f3a50a694f4126c029f
 # To regenerate: alef generate
-# To verify freshness: alef verify --exit-code
+# To verify freshness: alef verify
 # frozen_string_literal: true
 
 require_relative "html_to_markdown/version"