# This file is auto-generated by alef — DO NOT EDIT.
# alef:hash:4e15143f4af1ae8bafbdb1506ef057da924484c66a19483966333558ad437e75
# To regenerate: alef generate
# To verify freshness: alef verify --exit-code
# Issues & docs: https://github.com/kreuzberg-dev/alef
# frozen_string_literal: true

require "json"
require "sorbet-runtime"
require "kreuzberg_rb"
module Kreuzberg
  # How chunk size is measured.
  #
  # Defaults to `Characters` (Unicode character count). When using token-based sizing,
  # chunks are sized by token count according to the specified tokenizer.
  #
  # Token-based sizing uses HuggingFace tokenizers loaded at runtime. Any tokenizer
  # available on HuggingFace Hub can be used, including OpenAI-compatible tokenizers
  # (e.g., `Xenova/gpt-4o`, `Xenova/cl100k_base`).
  module ChunkSizing
    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 "characters" then ChunkSizingCharacters.from_hash(hash)
      when "tokenizer" then ChunkSizingTokenizer.from_hash(hash)
      else raise "Unknown discriminator: #{discriminator}"
      end
    end
  end

  # Size measured in Unicode characters (default).
  ChunkSizingCharacters = Data.define do
    include ChunkSizing
    extend T::Sig

    sig { returns(T::Boolean) }
    def characters? = true

    sig { returns(T::Boolean) }
    def tokenizer? = 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

  # Size measured in tokens from a HuggingFace tokenizer.
  ChunkSizingTokenizer = Data.define(:model, :cache_dir) do
    include ChunkSizing
    extend T::Sig

    # HuggingFace model ID or path, e.g. "Xenova/gpt-4o", "bert-base-uncased".
    sig { returns(String) }
    def model = super # rubocop:disable Lint/UselessMethodDefinition

    # Optional cache directory override for tokenizer files.
    # Defaults to hf-hub's standard cache (`~/.cache/huggingface/`).
    # Can also be set via `KREUZBERG_TOKENIZER_CACHE_DIR` environment variable.
    sig { returns(T.nilable(String)) }
    def cache_dir = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def characters? = false

    sig { returns(T::Boolean) }
    def tokenizer? = 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(model: hash[:model] || hash["model"], cache_dir: hash[:cache_dir] || hash["cache_dir"])
    end
  end
end

module Kreuzberg
  # Embedding model types supported by Kreuzberg.
  module EmbeddingModelType
    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 "preset" then EmbeddingModelTypePreset.from_hash(hash)
      when "custom" then EmbeddingModelTypeCustom.from_hash(hash)
      when "llm" then EmbeddingModelTypeLlm.from_hash(hash)
      when "plugin" then EmbeddingModelTypePlugin.from_hash(hash)
      else raise "Unknown discriminator: #{discriminator}"
      end
    end
  end

  # Use a preset model configuration (recommended)
  EmbeddingModelTypePreset = Data.define(:name) do
    include EmbeddingModelType
    extend T::Sig

    # @return [String]
    sig { returns(String) }
    def name = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def preset? = true

    sig { returns(T::Boolean) }
    def custom? = false

    sig { returns(T::Boolean) }
    def llm? = false

    sig { returns(T::Boolean) }
    def plugin? = 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(name: hash[:name] || hash["name"])
    end
  end

  # Use a custom ONNX model from HuggingFace
  EmbeddingModelTypeCustom = Data.define(:model_id, :dimensions) do
    include EmbeddingModelType
    extend T::Sig

    # @return [String]
    sig { returns(String) }
    def model_id = super # rubocop:disable Lint/UselessMethodDefinition

    # @return [Integer]
    sig { returns(Integer) }
    def dimensions = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def preset? = false

    sig { returns(T::Boolean) }
    def custom? = true

    sig { returns(T::Boolean) }
    def llm? = false

    sig { returns(T::Boolean) }
    def plugin? = 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(model_id: hash[:model_id] || hash["model_id"], dimensions: hash[:dimensions] || hash["dimensions"])
    end
  end

  # Provider-hosted embedding model via liter-llm.
  #
  # Uses the model specified in the nested `LlmConfig` (e.g.,
  # `"openai/text-embedding-3-small"`).
  EmbeddingModelTypeLlm = Data.define(:llm) do
    include EmbeddingModelType
    extend T::Sig

    # @return [LlmConfig]
    sig { returns(LlmConfig) }
    def llm = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def preset? = false

    sig { returns(T::Boolean) }
    def custom? = false

    sig { returns(T::Boolean) }
    def llm? = true

    sig { returns(T::Boolean) }
    def plugin? = 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(llm: hash[:llm] || hash["llm"])
    end
  end

  # In-process embedding backend registered via the plugin system.
  #
  # The caller registers an [`EmbeddingBackend`](crate::plugins::EmbeddingBackend) once
  # (e.g. a wrapper around an already-loaded `llama-cpp-python`, `sentence-transformers`,
  # or tuned ONNX model), then references it by name in config. Kreuzberg calls back
  # into the registered backend during chunking and standalone embed requests —
  # no HuggingFace download, no ONNX Runtime requirement, no HTTP sidecar.
  #
  # When this variant is selected, only the following [`EmbeddingConfig`] fields
  # apply: `normalize` (post-call L2 normalization) and `max_embed_duration_secs`
  # (dispatcher timeout). Model-loading fields (`batch_size`, `cache_dir`,
  # `show_download_progress`, `acceleration`) are ignored — the host owns the
  # model lifecycle.
  #
  # Semantic chunking falls back to [`ChunkingConfig::max_characters`] when this variant
  # is used, since there is no preset to look a chunk-size ceiling up against — size your
  # context window via `max_characters` directly.
  #
  # See `register_embedding_backend`.
  EmbeddingModelTypePlugin = Data.define(:name) do
    include EmbeddingModelType
    extend T::Sig

    # @return [String]
    sig { returns(String) }
    def name = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def preset? = false

    sig { returns(T::Boolean) }
    def custom? = false

    sig { returns(T::Boolean) }
    def llm? = false

    sig { returns(T::Boolean) }
    def plugin? = 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(name: hash[:name] || hash["name"])
    end
  end
end

module Kreuzberg
  # Tagged enum for node content. Each variant carries only type-specific data.
  #
  # Uses `#[serde(tag = "node_type")]` to avoid "type" keyword collision in
  # Go/Java/TypeScript bindings.
  module NodeContent
    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[:node_type] || hash["node_type"]
      case discriminator
      when "title" then NodeContentTitle.from_hash(hash)
      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 "formula" then NodeContentFormula.from_hash(hash)
      when "footnote" then NodeContentFootnote.from_hash(hash)
      when "group" then NodeContentGroup.from_hash(hash)
      when "page_break" then NodeContentPageBreak.from_hash(hash)
      when "slide" then NodeContentSlide.from_hash(hash)
      when "definition_list" then NodeContentDefinitionList.from_hash(hash)
      when "definition_item" then NodeContentDefinitionItem.from_hash(hash)
      when "citation" then NodeContentCitation.from_hash(hash)
      when "admonition" then NodeContentAdmonition.from_hash(hash)
      when "raw_block" then NodeContentRawBlock.from_hash(hash)
      when "metadata_block" then NodeContentMetadataBlock.from_hash(hash)
      else raise "Unknown discriminator: #{discriminator}"
      end
    end
  end

  # Document title.
  NodeContentTitle = Data.define(:text) do
    include NodeContent
    extend T::Sig

    # @return [String]
    sig { returns(String) }
    def text = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def title? = true

    sig { returns(T::Boolean) }
    def heading? = false

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def list? = false

    sig { returns(T::Boolean) }
    def list_item? = false

    sig { returns(T::Boolean) }
    def table? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def quote? = false

    sig { returns(T::Boolean) }
    def formula? = false

    sig { returns(T::Boolean) }
    def footnote? = false

    sig { returns(T::Boolean) }
    def group? = false

    sig { returns(T::Boolean) }
    def page_break? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def definition_list? = false

    sig { returns(T::Boolean) }
    def definition_item? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def admonition? = false

    sig { returns(T::Boolean) }
    def raw_block? = false

    sig { returns(T::Boolean) }
    def metadata_block? = 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(text: hash[:text] || hash["text"])
    end
  end

  # Section heading with level (1-6).
  NodeContentHeading = Data.define(:level, :text) do
    include NodeContent
    extend T::Sig

    # @return [Integer]
    sig { returns(Integer) }
    def level = super # rubocop:disable Lint/UselessMethodDefinition

    # @return [String]
    sig { returns(String) }
    def text = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def title? = false

    sig { returns(T::Boolean) }
    def heading? = true

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def list? = false

    sig { returns(T::Boolean) }
    def list_item? = false

    sig { returns(T::Boolean) }
    def table? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def quote? = false

    sig { returns(T::Boolean) }
    def formula? = false

    sig { returns(T::Boolean) }
    def footnote? = false

    sig { returns(T::Boolean) }
    def group? = false

    sig { returns(T::Boolean) }
    def page_break? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def definition_list? = false

    sig { returns(T::Boolean) }
    def definition_item? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def admonition? = false

    sig { returns(T::Boolean) }
    def raw_block? = false

    sig { returns(T::Boolean) }
    def metadata_block? = 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(level: hash[:level] || hash["level"], text: hash[:text] || hash["text"])
    end
  end

  # Body text paragraph.
  NodeContentParagraph = Data.define(:text) do
    include NodeContent
    extend T::Sig

    # @return [String]
    sig { returns(String) }
    def text = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def title? = false

    sig { returns(T::Boolean) }
    def heading? = false

    sig { returns(T::Boolean) }
    def paragraph? = true

    sig { returns(T::Boolean) }
    def list? = false

    sig { returns(T::Boolean) }
    def list_item? = false

    sig { returns(T::Boolean) }
    def table? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def quote? = false

    sig { returns(T::Boolean) }
    def formula? = false

    sig { returns(T::Boolean) }
    def footnote? = false

    sig { returns(T::Boolean) }
    def group? = false

    sig { returns(T::Boolean) }
    def page_break? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def definition_list? = false

    sig { returns(T::Boolean) }
    def definition_item? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def admonition? = false

    sig { returns(T::Boolean) }
    def raw_block? = false

    sig { returns(T::Boolean) }
    def metadata_block? = 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(text: hash[:text] || hash["text"])
    end
  end

  # List container — children are `ListItem` nodes.
  NodeContentList = Data.define(:ordered) do
    include NodeContent
    extend T::Sig

    # @return [T::Boolean]
    sig { returns(T::Boolean) }
    def ordered = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def title? = false

    sig { returns(T::Boolean) }
    def heading? = false

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def list? = true

    sig { returns(T::Boolean) }
    def list_item? = false

    sig { returns(T::Boolean) }
    def table? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def quote? = false

    sig { returns(T::Boolean) }
    def formula? = false

    sig { returns(T::Boolean) }
    def footnote? = false

    sig { returns(T::Boolean) }
    def group? = false

    sig { returns(T::Boolean) }
    def page_break? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def definition_list? = false

    sig { returns(T::Boolean) }
    def definition_item? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def admonition? = false

    sig { returns(T::Boolean) }
    def raw_block? = false

    sig { returns(T::Boolean) }
    def metadata_block? = 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(ordered: hash[:ordered] || hash["ordered"])
    end
  end

  # Individual list item.
  NodeContentListItem = Data.define(:text) do
    include NodeContent
    extend T::Sig

    # @return [String]
    sig { returns(String) }
    def text = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def title? = false

    sig { returns(T::Boolean) }
    def heading? = false

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def list? = false

    sig { returns(T::Boolean) }
    def list_item? = true

    sig { returns(T::Boolean) }
    def table? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def quote? = false

    sig { returns(T::Boolean) }
    def formula? = false

    sig { returns(T::Boolean) }
    def footnote? = false

    sig { returns(T::Boolean) }
    def group? = false

    sig { returns(T::Boolean) }
    def page_break? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def definition_list? = false

    sig { returns(T::Boolean) }
    def definition_item? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def admonition? = false

    sig { returns(T::Boolean) }
    def raw_block? = false

    sig { returns(T::Boolean) }
    def metadata_block? = 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(text: hash[:text] || hash["text"])
    end
  end

  # Table with structured cell grid.
  NodeContentTable = Data.define(:grid) do
    include NodeContent
    extend T::Sig

    # @return [TableGrid]
    sig { returns(TableGrid) }
    def grid = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def title? = false

    sig { returns(T::Boolean) }
    def heading? = false

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def list? = false

    sig { returns(T::Boolean) }
    def list_item? = false

    sig { returns(T::Boolean) }
    def table? = true

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def quote? = false

    sig { returns(T::Boolean) }
    def formula? = false

    sig { returns(T::Boolean) }
    def footnote? = false

    sig { returns(T::Boolean) }
    def group? = false

    sig { returns(T::Boolean) }
    def page_break? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def definition_list? = false

    sig { returns(T::Boolean) }
    def definition_item? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def admonition? = false

    sig { returns(T::Boolean) }
    def raw_block? = false

    sig { returns(T::Boolean) }
    def metadata_block? = 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(grid: hash[:grid] || hash["grid"])
    end
  end

  # Image reference.
  NodeContentImage = Data.define(:description, :image_index, :src) do
    include NodeContent
    extend T::Sig

    # @return [T.nilable(String)]
    sig { returns(T.nilable(String)) }
    def description = super # rubocop:disable Lint/UselessMethodDefinition

    # @return [T.nilable(Integer)]
    sig { returns(T.nilable(Integer)) }
    def image_index = super # rubocop:disable Lint/UselessMethodDefinition

    # Source URL or path of the image (from `<img src="...">` or `![](src)`).
    sig { returns(T.nilable(String)) }
    def src = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def title? = false

    sig { returns(T::Boolean) }
    def heading? = false

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def list? = false

    sig { returns(T::Boolean) }
    def list_item? = false

    sig { returns(T::Boolean) }
    def table? = false

    sig { returns(T::Boolean) }
    def image? = true

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def quote? = false

    sig { returns(T::Boolean) }
    def formula? = false

    sig { returns(T::Boolean) }
    def footnote? = false

    sig { returns(T::Boolean) }
    def group? = false

    sig { returns(T::Boolean) }
    def page_break? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def definition_list? = false

    sig { returns(T::Boolean) }
    def definition_item? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def admonition? = false

    sig { returns(T::Boolean) }
    def raw_block? = false

    sig { returns(T::Boolean) }
    def metadata_block? = 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(description: hash[:description] || hash["description"], image_index: hash[:image_index] || hash["image_index"], src: hash[:src] || hash["src"])
    end
  end

  # Code block.
  NodeContentCode = Data.define(:text, :language) do
    include NodeContent
    extend T::Sig

    # @return [String]
    sig { returns(String) }
    def text = super # rubocop:disable Lint/UselessMethodDefinition

    # @return [T.nilable(String)]
    sig { returns(T.nilable(String)) }
    def language = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def title? = false

    sig { returns(T::Boolean) }
    def heading? = false

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def list? = false

    sig { returns(T::Boolean) }
    def list_item? = false

    sig { returns(T::Boolean) }
    def table? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def code? = true

    sig { returns(T::Boolean) }
    def quote? = false

    sig { returns(T::Boolean) }
    def formula? = false

    sig { returns(T::Boolean) }
    def footnote? = false

    sig { returns(T::Boolean) }
    def group? = false

    sig { returns(T::Boolean) }
    def page_break? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def definition_list? = false

    sig { returns(T::Boolean) }
    def definition_item? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def admonition? = false

    sig { returns(T::Boolean) }
    def raw_block? = false

    sig { returns(T::Boolean) }
    def metadata_block? = 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(text: hash[:text] || hash["text"], language: hash[:language] || hash["language"])
    end
  end

  # Block quote — container, children carry the quoted content.
  NodeContentQuote = Data.define do
    include NodeContent
    extend T::Sig

    sig { returns(T::Boolean) }
    def title? = false

    sig { returns(T::Boolean) }
    def heading? = false

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def list? = false

    sig { returns(T::Boolean) }
    def list_item? = false

    sig { returns(T::Boolean) }
    def table? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def quote? = true

    sig { returns(T::Boolean) }
    def formula? = false

    sig { returns(T::Boolean) }
    def footnote? = false

    sig { returns(T::Boolean) }
    def group? = false

    sig { returns(T::Boolean) }
    def page_break? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def definition_list? = false

    sig { returns(T::Boolean) }
    def definition_item? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def admonition? = false

    sig { returns(T::Boolean) }
    def raw_block? = false

    sig { returns(T::Boolean) }
    def metadata_block? = 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

  # Mathematical formula / equation.
  NodeContentFormula = Data.define(:text) do
    include NodeContent
    extend T::Sig

    # @return [String]
    sig { returns(String) }
    def text = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def title? = false

    sig { returns(T::Boolean) }
    def heading? = false

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def list? = false

    sig { returns(T::Boolean) }
    def list_item? = false

    sig { returns(T::Boolean) }
    def table? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def quote? = false

    sig { returns(T::Boolean) }
    def formula? = true

    sig { returns(T::Boolean) }
    def footnote? = false

    sig { returns(T::Boolean) }
    def group? = false

    sig { returns(T::Boolean) }
    def page_break? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def definition_list? = false

    sig { returns(T::Boolean) }
    def definition_item? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def admonition? = false

    sig { returns(T::Boolean) }
    def raw_block? = false

    sig { returns(T::Boolean) }
    def metadata_block? = 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(text: hash[:text] || hash["text"])
    end
  end

  # Footnote reference content.
  NodeContentFootnote = Data.define(:text) do
    include NodeContent
    extend T::Sig

    # @return [String]
    sig { returns(String) }
    def text = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def title? = false

    sig { returns(T::Boolean) }
    def heading? = false

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def list? = false

    sig { returns(T::Boolean) }
    def list_item? = false

    sig { returns(T::Boolean) }
    def table? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def quote? = false

    sig { returns(T::Boolean) }
    def formula? = false

    sig { returns(T::Boolean) }
    def footnote? = true

    sig { returns(T::Boolean) }
    def group? = false

    sig { returns(T::Boolean) }
    def page_break? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def definition_list? = false

    sig { returns(T::Boolean) }
    def definition_item? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def admonition? = false

    sig { returns(T::Boolean) }
    def raw_block? = false

    sig { returns(T::Boolean) }
    def metadata_block? = 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(text: hash[:text] || hash["text"])
    end
  end

  # Logical grouping container (section, key-value area).
  #
  # `heading_level` + `heading_text` capture the section heading directly
  # rather than relying on a first-child positional convention.
  NodeContentGroup = Data.define(:label, :heading_level, :heading_text) do
    include NodeContent
    extend T::Sig

    # @return [T.nilable(String)]
    sig { returns(T.nilable(String)) }
    def label = super # rubocop:disable Lint/UselessMethodDefinition

    # @return [T.nilable(Integer)]
    sig { returns(T.nilable(Integer)) }
    def heading_level = super # rubocop:disable Lint/UselessMethodDefinition

    # @return [T.nilable(String)]
    sig { returns(T.nilable(String)) }
    def heading_text = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def title? = false

    sig { returns(T::Boolean) }
    def heading? = false

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def list? = false

    sig { returns(T::Boolean) }
    def list_item? = false

    sig { returns(T::Boolean) }
    def table? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def quote? = false

    sig { returns(T::Boolean) }
    def formula? = false

    sig { returns(T::Boolean) }
    def footnote? = false

    sig { returns(T::Boolean) }
    def group? = true

    sig { returns(T::Boolean) }
    def page_break? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def definition_list? = false

    sig { returns(T::Boolean) }
    def definition_item? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def admonition? = false

    sig { returns(T::Boolean) }
    def raw_block? = false

    sig { returns(T::Boolean) }
    def metadata_block? = 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(label: hash[:label] || hash["label"], heading_level: hash[:heading_level] || hash["heading_level"], heading_text: hash[:heading_text] || hash["heading_text"])
    end
  end

  # Page break marker.
  NodeContentPageBreak = Data.define do
    include NodeContent
    extend T::Sig

    sig { returns(T::Boolean) }
    def title? = false

    sig { returns(T::Boolean) }
    def heading? = false

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def list? = false

    sig { returns(T::Boolean) }
    def list_item? = false

    sig { returns(T::Boolean) }
    def table? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def quote? = false

    sig { returns(T::Boolean) }
    def formula? = false

    sig { returns(T::Boolean) }
    def footnote? = false

    sig { returns(T::Boolean) }
    def group? = false

    sig { returns(T::Boolean) }
    def page_break? = true

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def definition_list? = false

    sig { returns(T::Boolean) }
    def definition_item? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def admonition? = false

    sig { returns(T::Boolean) }
    def raw_block? = false

    sig { returns(T::Boolean) }
    def metadata_block? = 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

  # Presentation slide container — children are the slide's content nodes.
  NodeContentSlide = Data.define(:number, :title) do
    include NodeContent
    extend T::Sig

    # 1-indexed slide number.
    sig { returns(Integer) }
    def number = super # rubocop:disable Lint/UselessMethodDefinition

    # @return [T.nilable(String)]
    sig { returns(T.nilable(String)) }
    def title = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def title? = false

    sig { returns(T::Boolean) }
    def heading? = false

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def list? = false

    sig { returns(T::Boolean) }
    def list_item? = false

    sig { returns(T::Boolean) }
    def table? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def quote? = false

    sig { returns(T::Boolean) }
    def formula? = false

    sig { returns(T::Boolean) }
    def footnote? = false

    sig { returns(T::Boolean) }
    def group? = false

    sig { returns(T::Boolean) }
    def page_break? = false

    sig { returns(T::Boolean) }
    def slide? = true

    sig { returns(T::Boolean) }
    def definition_list? = false

    sig { returns(T::Boolean) }
    def definition_item? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def admonition? = false

    sig { returns(T::Boolean) }
    def raw_block? = false

    sig { returns(T::Boolean) }
    def metadata_block? = 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(number: hash[:number] || hash["number"], title: hash[:title] || hash["title"])
    end
  end

  # Definition list container — children are `DefinitionItem` nodes.
  NodeContentDefinitionList = Data.define do
    include NodeContent
    extend T::Sig

    sig { returns(T::Boolean) }
    def title? = false

    sig { returns(T::Boolean) }
    def heading? = false

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def list? = false

    sig { returns(T::Boolean) }
    def list_item? = false

    sig { returns(T::Boolean) }
    def table? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def quote? = false

    sig { returns(T::Boolean) }
    def formula? = false

    sig { returns(T::Boolean) }
    def footnote? = false

    sig { returns(T::Boolean) }
    def group? = false

    sig { returns(T::Boolean) }
    def page_break? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def definition_list? = true

    sig { returns(T::Boolean) }
    def definition_item? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def admonition? = false

    sig { returns(T::Boolean) }
    def raw_block? = false

    sig { returns(T::Boolean) }
    def metadata_block? = 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

  # Individual definition list entry with term and definition.
  NodeContentDefinitionItem = Data.define(:term, :definition) do
    include NodeContent
    extend T::Sig

    # @return [String]
    sig { returns(String) }
    def term = super # rubocop:disable Lint/UselessMethodDefinition

    # @return [String]
    sig { returns(String) }
    def definition = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def title? = false

    sig { returns(T::Boolean) }
    def heading? = false

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def list? = false

    sig { returns(T::Boolean) }
    def list_item? = false

    sig { returns(T::Boolean) }
    def table? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def quote? = false

    sig { returns(T::Boolean) }
    def formula? = false

    sig { returns(T::Boolean) }
    def footnote? = false

    sig { returns(T::Boolean) }
    def group? = false

    sig { returns(T::Boolean) }
    def page_break? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def definition_list? = false

    sig { returns(T::Boolean) }
    def definition_item? = true

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def admonition? = false

    sig { returns(T::Boolean) }
    def raw_block? = false

    sig { returns(T::Boolean) }
    def metadata_block? = 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(term: hash[:term] || hash["term"], definition: hash[:definition] || hash["definition"])
    end
  end

  # Citation or bibliographic reference.
  NodeContentCitation = Data.define(:key, :text) do
    include NodeContent
    extend T::Sig

    # @return [String]
    sig { returns(String) }
    def key = super # rubocop:disable Lint/UselessMethodDefinition

    # @return [String]
    sig { returns(String) }
    def text = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def title? = false

    sig { returns(T::Boolean) }
    def heading? = false

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def list? = false

    sig { returns(T::Boolean) }
    def list_item? = false

    sig { returns(T::Boolean) }
    def table? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def quote? = false

    sig { returns(T::Boolean) }
    def formula? = false

    sig { returns(T::Boolean) }
    def footnote? = false

    sig { returns(T::Boolean) }
    def group? = false

    sig { returns(T::Boolean) }
    def page_break? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def definition_list? = false

    sig { returns(T::Boolean) }
    def definition_item? = false

    sig { returns(T::Boolean) }
    def citation? = true

    sig { returns(T::Boolean) }
    def admonition? = false

    sig { returns(T::Boolean) }
    def raw_block? = false

    sig { returns(T::Boolean) }
    def metadata_block? = 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(key: hash[:key] || hash["key"], text: hash[:text] || hash["text"])
    end
  end

  # Admonition / callout container (note, warning, tip, etc.).
  #
  # Children carry the admonition body content.
  NodeContentAdmonition = Data.define(:kind, :title) do
    include NodeContent
    extend T::Sig

    # Kind of admonition (e.g. "note", "warning", "tip", "danger").
    sig { returns(String) }
    def kind = super # rubocop:disable Lint/UselessMethodDefinition

    # @return [T.nilable(String)]
    sig { returns(T.nilable(String)) }
    def title = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def title? = false

    sig { returns(T::Boolean) }
    def heading? = false

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def list? = false

    sig { returns(T::Boolean) }
    def list_item? = false

    sig { returns(T::Boolean) }
    def table? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def quote? = false

    sig { returns(T::Boolean) }
    def formula? = false

    sig { returns(T::Boolean) }
    def footnote? = false

    sig { returns(T::Boolean) }
    def group? = false

    sig { returns(T::Boolean) }
    def page_break? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def definition_list? = false

    sig { returns(T::Boolean) }
    def definition_item? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def admonition? = true

    sig { returns(T::Boolean) }
    def raw_block? = false

    sig { returns(T::Boolean) }
    def metadata_block? = 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(kind: hash[:kind] || hash["kind"], title: hash[:title] || hash["title"])
    end
  end

  # Raw block preserved verbatim from the source format.
  #
  # Used for content that cannot be mapped to a semantic node type
  # (e.g. JSX in MDX, raw LaTeX in markdown, embedded HTML).
  NodeContentRawBlock = Data.define(:format, :content) do
    include NodeContent
    extend T::Sig

    # Source format identifier (e.g. "html", "latex", "jsx").
    sig { returns(String) }
    def format = super # rubocop:disable Lint/UselessMethodDefinition

    # @return [String]
    sig { returns(String) }
    def content = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def title? = false

    sig { returns(T::Boolean) }
    def heading? = false

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def list? = false

    sig { returns(T::Boolean) }
    def list_item? = false

    sig { returns(T::Boolean) }
    def table? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def quote? = false

    sig { returns(T::Boolean) }
    def formula? = false

    sig { returns(T::Boolean) }
    def footnote? = false

    sig { returns(T::Boolean) }
    def group? = false

    sig { returns(T::Boolean) }
    def page_break? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def definition_list? = false

    sig { returns(T::Boolean) }
    def definition_item? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def admonition? = false

    sig { returns(T::Boolean) }
    def raw_block? = true

    sig { returns(T::Boolean) }
    def metadata_block? = 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(format: hash[:format] || hash["format"], content: hash[:content] || hash["content"])
    end
  end

  # Structured metadata block (email headers, YAML frontmatter, etc.).
  NodeContentMetadataBlock = Data.define(:entries) do
    include NodeContent
    extend T::Sig

    # @return [T::Array[T::Array[String]]]
    sig { returns(T::Array[T::Array[String]]) }
    def entries = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def title? = false

    sig { returns(T::Boolean) }
    def heading? = false

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def list? = false

    sig { returns(T::Boolean) }
    def list_item? = false

    sig { returns(T::Boolean) }
    def table? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def quote? = false

    sig { returns(T::Boolean) }
    def formula? = false

    sig { returns(T::Boolean) }
    def footnote? = false

    sig { returns(T::Boolean) }
    def group? = false

    sig { returns(T::Boolean) }
    def page_break? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def definition_list? = false

    sig { returns(T::Boolean) }
    def definition_item? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def admonition? = false

    sig { returns(T::Boolean) }
    def raw_block? = false

    sig { returns(T::Boolean) }
    def metadata_block? = 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(entries: hash[:entries] || hash["entries"])
    end
  end
end

module Kreuzberg
  # Types of inline text annotations.
  module AnnotationKind
    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[:annotation_type] || hash["annotation_type"]
      case 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 "link" then AnnotationKindLink.from_hash(hash)
      when "highlight" then AnnotationKindHighlight.from_hash(hash)
      when "color" then AnnotationKindColor.from_hash(hash)
      when "font_size" then AnnotationKindFontSize.from_hash(hash)
      when "custom" then AnnotationKindCustom.from_hash(hash)
      else raise "Unknown discriminator: #{discriminator}"
      end
    end
  end

  # Variant AnnotationKindBold of the AnnotationKind sum type.
  AnnotationKindBold = Data.define do
    include AnnotationKind
    extend T::Sig

    sig { returns(T::Boolean) }
    def bold? = true

    sig { returns(T::Boolean) }
    def italic? = false

    sig { returns(T::Boolean) }
    def underline? = false

    sig { returns(T::Boolean) }
    def strikethrough? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def subscript? = false

    sig { returns(T::Boolean) }
    def superscript? = false

    sig { returns(T::Boolean) }
    def link? = false

    sig { returns(T::Boolean) }
    def highlight? = false

    sig { returns(T::Boolean) }
    def color? = false

    sig { returns(T::Boolean) }
    def font_size? = false

    sig { returns(T::Boolean) }
    def custom? = 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

  # Variant AnnotationKindItalic of the AnnotationKind sum type.
  AnnotationKindItalic = Data.define do
    include AnnotationKind
    extend T::Sig

    sig { returns(T::Boolean) }
    def bold? = false

    sig { returns(T::Boolean) }
    def italic? = true

    sig { returns(T::Boolean) }
    def underline? = false

    sig { returns(T::Boolean) }
    def strikethrough? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def subscript? = false

    sig { returns(T::Boolean) }
    def superscript? = false

    sig { returns(T::Boolean) }
    def link? = false

    sig { returns(T::Boolean) }
    def highlight? = false

    sig { returns(T::Boolean) }
    def color? = false

    sig { returns(T::Boolean) }
    def font_size? = false

    sig { returns(T::Boolean) }
    def custom? = 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

  # Variant AnnotationKindUnderline of the AnnotationKind sum type.
  AnnotationKindUnderline = Data.define do
    include AnnotationKind
    extend T::Sig

    sig { returns(T::Boolean) }
    def bold? = false

    sig { returns(T::Boolean) }
    def italic? = false

    sig { returns(T::Boolean) }
    def underline? = true

    sig { returns(T::Boolean) }
    def strikethrough? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def subscript? = false

    sig { returns(T::Boolean) }
    def superscript? = false

    sig { returns(T::Boolean) }
    def link? = false

    sig { returns(T::Boolean) }
    def highlight? = false

    sig { returns(T::Boolean) }
    def color? = false

    sig { returns(T::Boolean) }
    def font_size? = false

    sig { returns(T::Boolean) }
    def custom? = 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

  # Variant AnnotationKindStrikethrough of the AnnotationKind sum type.
  AnnotationKindStrikethrough = Data.define do
    include AnnotationKind
    extend T::Sig

    sig { returns(T::Boolean) }
    def bold? = false

    sig { returns(T::Boolean) }
    def italic? = false

    sig { returns(T::Boolean) }
    def underline? = false

    sig { returns(T::Boolean) }
    def strikethrough? = true

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def subscript? = false

    sig { returns(T::Boolean) }
    def superscript? = false

    sig { returns(T::Boolean) }
    def link? = false

    sig { returns(T::Boolean) }
    def highlight? = false

    sig { returns(T::Boolean) }
    def color? = false

    sig { returns(T::Boolean) }
    def font_size? = false

    sig { returns(T::Boolean) }
    def custom? = 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

  # Variant AnnotationKindCode of the AnnotationKind sum type.
  AnnotationKindCode = Data.define do
    include AnnotationKind
    extend T::Sig

    sig { returns(T::Boolean) }
    def bold? = false

    sig { returns(T::Boolean) }
    def italic? = false

    sig { returns(T::Boolean) }
    def underline? = false

    sig { returns(T::Boolean) }
    def strikethrough? = false

    sig { returns(T::Boolean) }
    def code? = true

    sig { returns(T::Boolean) }
    def subscript? = false

    sig { returns(T::Boolean) }
    def superscript? = false

    sig { returns(T::Boolean) }
    def link? = false

    sig { returns(T::Boolean) }
    def highlight? = false

    sig { returns(T::Boolean) }
    def color? = false

    sig { returns(T::Boolean) }
    def font_size? = false

    sig { returns(T::Boolean) }
    def custom? = 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

  # Variant AnnotationKindSubscript of the AnnotationKind sum type.
  AnnotationKindSubscript = Data.define do
    include AnnotationKind
    extend T::Sig

    sig { returns(T::Boolean) }
    def bold? = false

    sig { returns(T::Boolean) }
    def italic? = false

    sig { returns(T::Boolean) }
    def underline? = false

    sig { returns(T::Boolean) }
    def strikethrough? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def subscript? = true

    sig { returns(T::Boolean) }
    def superscript? = false

    sig { returns(T::Boolean) }
    def link? = false

    sig { returns(T::Boolean) }
    def highlight? = false

    sig { returns(T::Boolean) }
    def color? = false

    sig { returns(T::Boolean) }
    def font_size? = false

    sig { returns(T::Boolean) }
    def custom? = 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

  # Variant AnnotationKindSuperscript of the AnnotationKind sum type.
  AnnotationKindSuperscript = Data.define do
    include AnnotationKind
    extend T::Sig

    sig { returns(T::Boolean) }
    def bold? = false

    sig { returns(T::Boolean) }
    def italic? = false

    sig { returns(T::Boolean) }
    def underline? = false

    sig { returns(T::Boolean) }
    def strikethrough? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def subscript? = false

    sig { returns(T::Boolean) }
    def superscript? = true

    sig { returns(T::Boolean) }
    def link? = false

    sig { returns(T::Boolean) }
    def highlight? = false

    sig { returns(T::Boolean) }
    def color? = false

    sig { returns(T::Boolean) }
    def font_size? = false

    sig { returns(T::Boolean) }
    def custom? = 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

  # Variant AnnotationKindLink of the AnnotationKind sum type.
  AnnotationKindLink = Data.define(:url, :title) do
    include AnnotationKind
    extend T::Sig

    # @return [String]
    sig { returns(String) }
    def url = super # rubocop:disable Lint/UselessMethodDefinition

    # @return [T.nilable(String)]
    sig { returns(T.nilable(String)) }
    def title = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def bold? = false

    sig { returns(T::Boolean) }
    def italic? = false

    sig { returns(T::Boolean) }
    def underline? = false

    sig { returns(T::Boolean) }
    def strikethrough? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def subscript? = false

    sig { returns(T::Boolean) }
    def superscript? = false

    sig { returns(T::Boolean) }
    def link? = true

    sig { returns(T::Boolean) }
    def highlight? = false

    sig { returns(T::Boolean) }
    def color? = false

    sig { returns(T::Boolean) }
    def font_size? = false

    sig { returns(T::Boolean) }
    def custom? = 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(url: hash[:url] || hash["url"], title: hash[:title] || hash["title"])
    end
  end

  # Highlighted text (PDF highlights, HTML `<mark>`).
  AnnotationKindHighlight = Data.define do
    include AnnotationKind
    extend T::Sig

    sig { returns(T::Boolean) }
    def bold? = false

    sig { returns(T::Boolean) }
    def italic? = false

    sig { returns(T::Boolean) }
    def underline? = false

    sig { returns(T::Boolean) }
    def strikethrough? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def subscript? = false

    sig { returns(T::Boolean) }
    def superscript? = false

    sig { returns(T::Boolean) }
    def link? = false

    sig { returns(T::Boolean) }
    def highlight? = true

    sig { returns(T::Boolean) }
    def color? = false

    sig { returns(T::Boolean) }
    def font_size? = false

    sig { returns(T::Boolean) }
    def custom? = 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

  # Text color (CSS-compatible value, e.g. "#ff0000", "red").
  AnnotationKindColor = Data.define(:value) do
    include AnnotationKind
    extend T::Sig

    # @return [String]
    sig { returns(String) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def bold? = false

    sig { returns(T::Boolean) }
    def italic? = false

    sig { returns(T::Boolean) }
    def underline? = false

    sig { returns(T::Boolean) }
    def strikethrough? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def subscript? = false

    sig { returns(T::Boolean) }
    def superscript? = false

    sig { returns(T::Boolean) }
    def link? = false

    sig { returns(T::Boolean) }
    def highlight? = false

    sig { returns(T::Boolean) }
    def color? = true

    sig { returns(T::Boolean) }
    def font_size? = false

    sig { returns(T::Boolean) }
    def custom? = 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[:value] || hash["value"])
    end
  end

  # Font size with units (e.g. "12pt", "1.2em", "16px").
  AnnotationKindFontSize = Data.define(:value) do
    include AnnotationKind
    extend T::Sig

    # @return [String]
    sig { returns(String) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def bold? = false

    sig { returns(T::Boolean) }
    def italic? = false

    sig { returns(T::Boolean) }
    def underline? = false

    sig { returns(T::Boolean) }
    def strikethrough? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def subscript? = false

    sig { returns(T::Boolean) }
    def superscript? = false

    sig { returns(T::Boolean) }
    def link? = false

    sig { returns(T::Boolean) }
    def highlight? = false

    sig { returns(T::Boolean) }
    def color? = false

    sig { returns(T::Boolean) }
    def font_size? = true

    sig { returns(T::Boolean) }
    def custom? = 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[:value] || hash["value"])
    end
  end

  # Extensible annotation for format-specific styling.
  AnnotationKindCustom = Data.define(:name, :value) do
    include AnnotationKind
    extend T::Sig

    # @return [String]
    sig { returns(String) }
    def name = super # rubocop:disable Lint/UselessMethodDefinition

    # @return [T.nilable(String)]
    sig { returns(T.nilable(String)) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def bold? = false

    sig { returns(T::Boolean) }
    def italic? = false

    sig { returns(T::Boolean) }
    def underline? = false

    sig { returns(T::Boolean) }
    def strikethrough? = false

    sig { returns(T::Boolean) }
    def code? = false

    sig { returns(T::Boolean) }
    def subscript? = false

    sig { returns(T::Boolean) }
    def superscript? = false

    sig { returns(T::Boolean) }
    def link? = false

    sig { returns(T::Boolean) }
    def highlight? = false

    sig { returns(T::Boolean) }
    def color? = false

    sig { returns(T::Boolean) }
    def font_size? = false

    sig { returns(T::Boolean) }
    def custom? = 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(name: hash[:name] || hash["name"], value: hash[:value] || hash["value"])
    end
  end
end

module Kreuzberg
  # Format-specific metadata (discriminated union).
  #
  # Only one format type can exist per extraction result. This provides
  # type-safe, clean metadata without nested optionals.
  module FormatMetadata
    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[:format_type] || hash["format_type"]
      case discriminator
      when "pdf" then FormatMetadataPdf.from_hash(hash)
      when "docx" then FormatMetadataDocx.from_hash(hash)
      when "excel" then FormatMetadataExcel.from_hash(hash)
      when "email" then FormatMetadataEmail.from_hash(hash)
      when "pptx" then FormatMetadataPptx.from_hash(hash)
      when "archive" then FormatMetadataArchive.from_hash(hash)
      when "image" then FormatMetadataImage.from_hash(hash)
      when "xml" then FormatMetadataXml.from_hash(hash)
      when "text" then FormatMetadataText.from_hash(hash)
      when "html" then FormatMetadataHtml.from_hash(hash)
      when "ocr" then FormatMetadataOcr.from_hash(hash)
      when "csv" then FormatMetadataCsv.from_hash(hash)
      when "bibtex" then FormatMetadataBibtex.from_hash(hash)
      when "citation" then FormatMetadataCitation.from_hash(hash)
      when "fiction_book" then FormatMetadataFictionBook.from_hash(hash)
      when "dbf" then FormatMetadataDbf.from_hash(hash)
      when "jats" then FormatMetadataJats.from_hash(hash)
      when "epub" then FormatMetadataEpub.from_hash(hash)
      when "pst" then FormatMetadataPst.from_hash(hash)
      when "code" then FormatMetadataCode.from_hash(hash)
      else raise "Unknown discriminator: #{discriminator}"
      end
    end
  end

  # Variant FormatMetadataPdf of the FormatMetadata sum type.
  FormatMetadataPdf = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [PdfMetadata]
    sig { returns(PdfMetadata) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = true

    sig { returns(T::Boolean) }
    def docx? = false

    sig { returns(T::Boolean) }
    def excel? = false

    sig { returns(T::Boolean) }
    def email? = false

    sig { returns(T::Boolean) }
    def pptx? = false

    sig { returns(T::Boolean) }
    def archive? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def xml? = false

    sig { returns(T::Boolean) }
    def text? = false

    sig { returns(T::Boolean) }
    def html? = false

    sig { returns(T::Boolean) }
    def ocr? = false

    sig { returns(T::Boolean) }
    def csv? = false

    sig { returns(T::Boolean) }
    def bibtex? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def fiction_book? = false

    sig { returns(T::Boolean) }
    def dbf? = false

    sig { returns(T::Boolean) }
    def jats? = false

    sig { returns(T::Boolean) }
    def epub? = false

    sig { returns(T::Boolean) }
    def pst? = false

    sig { returns(T::Boolean) }
    def code? = 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

  # Variant FormatMetadataDocx of the FormatMetadata sum type.
  FormatMetadataDocx = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [DocxMetadata]
    sig { returns(DocxMetadata) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = false

    sig { returns(T::Boolean) }
    def docx? = true

    sig { returns(T::Boolean) }
    def excel? = false

    sig { returns(T::Boolean) }
    def email? = false

    sig { returns(T::Boolean) }
    def pptx? = false

    sig { returns(T::Boolean) }
    def archive? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def xml? = false

    sig { returns(T::Boolean) }
    def text? = false

    sig { returns(T::Boolean) }
    def html? = false

    sig { returns(T::Boolean) }
    def ocr? = false

    sig { returns(T::Boolean) }
    def csv? = false

    sig { returns(T::Boolean) }
    def bibtex? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def fiction_book? = false

    sig { returns(T::Boolean) }
    def dbf? = false

    sig { returns(T::Boolean) }
    def jats? = false

    sig { returns(T::Boolean) }
    def epub? = false

    sig { returns(T::Boolean) }
    def pst? = false

    sig { returns(T::Boolean) }
    def code? = 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

  # Variant FormatMetadataExcel of the FormatMetadata sum type.
  FormatMetadataExcel = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [ExcelMetadata]
    sig { returns(ExcelMetadata) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = false

    sig { returns(T::Boolean) }
    def docx? = false

    sig { returns(T::Boolean) }
    def excel? = true

    sig { returns(T::Boolean) }
    def email? = false

    sig { returns(T::Boolean) }
    def pptx? = false

    sig { returns(T::Boolean) }
    def archive? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def xml? = false

    sig { returns(T::Boolean) }
    def text? = false

    sig { returns(T::Boolean) }
    def html? = false

    sig { returns(T::Boolean) }
    def ocr? = false

    sig { returns(T::Boolean) }
    def csv? = false

    sig { returns(T::Boolean) }
    def bibtex? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def fiction_book? = false

    sig { returns(T::Boolean) }
    def dbf? = false

    sig { returns(T::Boolean) }
    def jats? = false

    sig { returns(T::Boolean) }
    def epub? = false

    sig { returns(T::Boolean) }
    def pst? = false

    sig { returns(T::Boolean) }
    def code? = 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

  # Variant FormatMetadataEmail of the FormatMetadata sum type.
  FormatMetadataEmail = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [EmailMetadata]
    sig { returns(EmailMetadata) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = false

    sig { returns(T::Boolean) }
    def docx? = false

    sig { returns(T::Boolean) }
    def excel? = false

    sig { returns(T::Boolean) }
    def email? = true

    sig { returns(T::Boolean) }
    def pptx? = false

    sig { returns(T::Boolean) }
    def archive? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def xml? = false

    sig { returns(T::Boolean) }
    def text? = false

    sig { returns(T::Boolean) }
    def html? = false

    sig { returns(T::Boolean) }
    def ocr? = false

    sig { returns(T::Boolean) }
    def csv? = false

    sig { returns(T::Boolean) }
    def bibtex? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def fiction_book? = false

    sig { returns(T::Boolean) }
    def dbf? = false

    sig { returns(T::Boolean) }
    def jats? = false

    sig { returns(T::Boolean) }
    def epub? = false

    sig { returns(T::Boolean) }
    def pst? = false

    sig { returns(T::Boolean) }
    def code? = 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

  # Variant FormatMetadataPptx of the FormatMetadata sum type.
  FormatMetadataPptx = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [PptxMetadata]
    sig { returns(PptxMetadata) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = false

    sig { returns(T::Boolean) }
    def docx? = false

    sig { returns(T::Boolean) }
    def excel? = false

    sig { returns(T::Boolean) }
    def email? = false

    sig { returns(T::Boolean) }
    def pptx? = true

    sig { returns(T::Boolean) }
    def archive? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def xml? = false

    sig { returns(T::Boolean) }
    def text? = false

    sig { returns(T::Boolean) }
    def html? = false

    sig { returns(T::Boolean) }
    def ocr? = false

    sig { returns(T::Boolean) }
    def csv? = false

    sig { returns(T::Boolean) }
    def bibtex? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def fiction_book? = false

    sig { returns(T::Boolean) }
    def dbf? = false

    sig { returns(T::Boolean) }
    def jats? = false

    sig { returns(T::Boolean) }
    def epub? = false

    sig { returns(T::Boolean) }
    def pst? = false

    sig { returns(T::Boolean) }
    def code? = 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

  # Variant FormatMetadataArchive of the FormatMetadata sum type.
  FormatMetadataArchive = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [ArchiveMetadata]
    sig { returns(ArchiveMetadata) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = false

    sig { returns(T::Boolean) }
    def docx? = false

    sig { returns(T::Boolean) }
    def excel? = false

    sig { returns(T::Boolean) }
    def email? = false

    sig { returns(T::Boolean) }
    def pptx? = false

    sig { returns(T::Boolean) }
    def archive? = true

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def xml? = false

    sig { returns(T::Boolean) }
    def text? = false

    sig { returns(T::Boolean) }
    def html? = false

    sig { returns(T::Boolean) }
    def ocr? = false

    sig { returns(T::Boolean) }
    def csv? = false

    sig { returns(T::Boolean) }
    def bibtex? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def fiction_book? = false

    sig { returns(T::Boolean) }
    def dbf? = false

    sig { returns(T::Boolean) }
    def jats? = false

    sig { returns(T::Boolean) }
    def epub? = false

    sig { returns(T::Boolean) }
    def pst? = false

    sig { returns(T::Boolean) }
    def code? = 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

  # Variant FormatMetadataImage of the FormatMetadata sum type.
  FormatMetadataImage = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [ImageMetadata]
    sig { returns(ImageMetadata) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = false

    sig { returns(T::Boolean) }
    def docx? = false

    sig { returns(T::Boolean) }
    def excel? = false

    sig { returns(T::Boolean) }
    def email? = false

    sig { returns(T::Boolean) }
    def pptx? = false

    sig { returns(T::Boolean) }
    def archive? = false

    sig { returns(T::Boolean) }
    def image? = true

    sig { returns(T::Boolean) }
    def xml? = false

    sig { returns(T::Boolean) }
    def text? = false

    sig { returns(T::Boolean) }
    def html? = false

    sig { returns(T::Boolean) }
    def ocr? = false

    sig { returns(T::Boolean) }
    def csv? = false

    sig { returns(T::Boolean) }
    def bibtex? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def fiction_book? = false

    sig { returns(T::Boolean) }
    def dbf? = false

    sig { returns(T::Boolean) }
    def jats? = false

    sig { returns(T::Boolean) }
    def epub? = false

    sig { returns(T::Boolean) }
    def pst? = false

    sig { returns(T::Boolean) }
    def code? = 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

  # Variant FormatMetadataXml of the FormatMetadata sum type.
  FormatMetadataXml = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [XmlMetadata]
    sig { returns(XmlMetadata) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = false

    sig { returns(T::Boolean) }
    def docx? = false

    sig { returns(T::Boolean) }
    def excel? = false

    sig { returns(T::Boolean) }
    def email? = false

    sig { returns(T::Boolean) }
    def pptx? = false

    sig { returns(T::Boolean) }
    def archive? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def xml? = true

    sig { returns(T::Boolean) }
    def text? = false

    sig { returns(T::Boolean) }
    def html? = false

    sig { returns(T::Boolean) }
    def ocr? = false

    sig { returns(T::Boolean) }
    def csv? = false

    sig { returns(T::Boolean) }
    def bibtex? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def fiction_book? = false

    sig { returns(T::Boolean) }
    def dbf? = false

    sig { returns(T::Boolean) }
    def jats? = false

    sig { returns(T::Boolean) }
    def epub? = false

    sig { returns(T::Boolean) }
    def pst? = false

    sig { returns(T::Boolean) }
    def code? = 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

  # Variant FormatMetadataText of the FormatMetadata sum type.
  FormatMetadataText = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [TextMetadata]
    sig { returns(TextMetadata) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = false

    sig { returns(T::Boolean) }
    def docx? = false

    sig { returns(T::Boolean) }
    def excel? = false

    sig { returns(T::Boolean) }
    def email? = false

    sig { returns(T::Boolean) }
    def pptx? = false

    sig { returns(T::Boolean) }
    def archive? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def xml? = false

    sig { returns(T::Boolean) }
    def text? = true

    sig { returns(T::Boolean) }
    def html? = false

    sig { returns(T::Boolean) }
    def ocr? = false

    sig { returns(T::Boolean) }
    def csv? = false

    sig { returns(T::Boolean) }
    def bibtex? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def fiction_book? = false

    sig { returns(T::Boolean) }
    def dbf? = false

    sig { returns(T::Boolean) }
    def jats? = false

    sig { returns(T::Boolean) }
    def epub? = false

    sig { returns(T::Boolean) }
    def pst? = false

    sig { returns(T::Boolean) }
    def code? = 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

  # Variant FormatMetadataHtml of the FormatMetadata sum type.
  FormatMetadataHtml = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [HtmlMetadata]
    sig { returns(HtmlMetadata) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = false

    sig { returns(T::Boolean) }
    def docx? = false

    sig { returns(T::Boolean) }
    def excel? = false

    sig { returns(T::Boolean) }
    def email? = false

    sig { returns(T::Boolean) }
    def pptx? = false

    sig { returns(T::Boolean) }
    def archive? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def xml? = false

    sig { returns(T::Boolean) }
    def text? = false

    sig { returns(T::Boolean) }
    def html? = true

    sig { returns(T::Boolean) }
    def ocr? = false

    sig { returns(T::Boolean) }
    def csv? = false

    sig { returns(T::Boolean) }
    def bibtex? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def fiction_book? = false

    sig { returns(T::Boolean) }
    def dbf? = false

    sig { returns(T::Boolean) }
    def jats? = false

    sig { returns(T::Boolean) }
    def epub? = false

    sig { returns(T::Boolean) }
    def pst? = false

    sig { returns(T::Boolean) }
    def code? = 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

  # Variant FormatMetadataOcr of the FormatMetadata sum type.
  FormatMetadataOcr = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [OcrMetadata]
    sig { returns(OcrMetadata) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = false

    sig { returns(T::Boolean) }
    def docx? = false

    sig { returns(T::Boolean) }
    def excel? = false

    sig { returns(T::Boolean) }
    def email? = false

    sig { returns(T::Boolean) }
    def pptx? = false

    sig { returns(T::Boolean) }
    def archive? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def xml? = false

    sig { returns(T::Boolean) }
    def text? = false

    sig { returns(T::Boolean) }
    def html? = false

    sig { returns(T::Boolean) }
    def ocr? = true

    sig { returns(T::Boolean) }
    def csv? = false

    sig { returns(T::Boolean) }
    def bibtex? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def fiction_book? = false

    sig { returns(T::Boolean) }
    def dbf? = false

    sig { returns(T::Boolean) }
    def jats? = false

    sig { returns(T::Boolean) }
    def epub? = false

    sig { returns(T::Boolean) }
    def pst? = false

    sig { returns(T::Boolean) }
    def code? = 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

  # Variant FormatMetadataCsv of the FormatMetadata sum type.
  FormatMetadataCsv = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [CsvMetadata]
    sig { returns(CsvMetadata) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = false

    sig { returns(T::Boolean) }
    def docx? = false

    sig { returns(T::Boolean) }
    def excel? = false

    sig { returns(T::Boolean) }
    def email? = false

    sig { returns(T::Boolean) }
    def pptx? = false

    sig { returns(T::Boolean) }
    def archive? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def xml? = false

    sig { returns(T::Boolean) }
    def text? = false

    sig { returns(T::Boolean) }
    def html? = false

    sig { returns(T::Boolean) }
    def ocr? = false

    sig { returns(T::Boolean) }
    def csv? = true

    sig { returns(T::Boolean) }
    def bibtex? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def fiction_book? = false

    sig { returns(T::Boolean) }
    def dbf? = false

    sig { returns(T::Boolean) }
    def jats? = false

    sig { returns(T::Boolean) }
    def epub? = false

    sig { returns(T::Boolean) }
    def pst? = false

    sig { returns(T::Boolean) }
    def code? = 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

  # Variant FormatMetadataBibtex of the FormatMetadata sum type.
  FormatMetadataBibtex = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [BibtexMetadata]
    sig { returns(BibtexMetadata) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = false

    sig { returns(T::Boolean) }
    def docx? = false

    sig { returns(T::Boolean) }
    def excel? = false

    sig { returns(T::Boolean) }
    def email? = false

    sig { returns(T::Boolean) }
    def pptx? = false

    sig { returns(T::Boolean) }
    def archive? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def xml? = false

    sig { returns(T::Boolean) }
    def text? = false

    sig { returns(T::Boolean) }
    def html? = false

    sig { returns(T::Boolean) }
    def ocr? = false

    sig { returns(T::Boolean) }
    def csv? = false

    sig { returns(T::Boolean) }
    def bibtex? = true

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def fiction_book? = false

    sig { returns(T::Boolean) }
    def dbf? = false

    sig { returns(T::Boolean) }
    def jats? = false

    sig { returns(T::Boolean) }
    def epub? = false

    sig { returns(T::Boolean) }
    def pst? = false

    sig { returns(T::Boolean) }
    def code? = 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

  # Variant FormatMetadataCitation of the FormatMetadata sum type.
  FormatMetadataCitation = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [CitationMetadata]
    sig { returns(CitationMetadata) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = false

    sig { returns(T::Boolean) }
    def docx? = false

    sig { returns(T::Boolean) }
    def excel? = false

    sig { returns(T::Boolean) }
    def email? = false

    sig { returns(T::Boolean) }
    def pptx? = false

    sig { returns(T::Boolean) }
    def archive? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def xml? = false

    sig { returns(T::Boolean) }
    def text? = false

    sig { returns(T::Boolean) }
    def html? = false

    sig { returns(T::Boolean) }
    def ocr? = false

    sig { returns(T::Boolean) }
    def csv? = false

    sig { returns(T::Boolean) }
    def bibtex? = false

    sig { returns(T::Boolean) }
    def citation? = true

    sig { returns(T::Boolean) }
    def fiction_book? = false

    sig { returns(T::Boolean) }
    def dbf? = false

    sig { returns(T::Boolean) }
    def jats? = false

    sig { returns(T::Boolean) }
    def epub? = false

    sig { returns(T::Boolean) }
    def pst? = false

    sig { returns(T::Boolean) }
    def code? = 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

  # Variant FormatMetadataFictionBook of the FormatMetadata sum type.
  FormatMetadataFictionBook = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [FictionBookMetadata]
    sig { returns(FictionBookMetadata) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = false

    sig { returns(T::Boolean) }
    def docx? = false

    sig { returns(T::Boolean) }
    def excel? = false

    sig { returns(T::Boolean) }
    def email? = false

    sig { returns(T::Boolean) }
    def pptx? = false

    sig { returns(T::Boolean) }
    def archive? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def xml? = false

    sig { returns(T::Boolean) }
    def text? = false

    sig { returns(T::Boolean) }
    def html? = false

    sig { returns(T::Boolean) }
    def ocr? = false

    sig { returns(T::Boolean) }
    def csv? = false

    sig { returns(T::Boolean) }
    def bibtex? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def fiction_book? = true

    sig { returns(T::Boolean) }
    def dbf? = false

    sig { returns(T::Boolean) }
    def jats? = false

    sig { returns(T::Boolean) }
    def epub? = false

    sig { returns(T::Boolean) }
    def pst? = false

    sig { returns(T::Boolean) }
    def code? = 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

  # Variant FormatMetadataDbf of the FormatMetadata sum type.
  FormatMetadataDbf = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [DbfMetadata]
    sig { returns(DbfMetadata) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = false

    sig { returns(T::Boolean) }
    def docx? = false

    sig { returns(T::Boolean) }
    def excel? = false

    sig { returns(T::Boolean) }
    def email? = false

    sig { returns(T::Boolean) }
    def pptx? = false

    sig { returns(T::Boolean) }
    def archive? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def xml? = false

    sig { returns(T::Boolean) }
    def text? = false

    sig { returns(T::Boolean) }
    def html? = false

    sig { returns(T::Boolean) }
    def ocr? = false

    sig { returns(T::Boolean) }
    def csv? = false

    sig { returns(T::Boolean) }
    def bibtex? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def fiction_book? = false

    sig { returns(T::Boolean) }
    def dbf? = true

    sig { returns(T::Boolean) }
    def jats? = false

    sig { returns(T::Boolean) }
    def epub? = false

    sig { returns(T::Boolean) }
    def pst? = false

    sig { returns(T::Boolean) }
    def code? = 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

  # Variant FormatMetadataJats of the FormatMetadata sum type.
  FormatMetadataJats = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [JatsMetadata]
    sig { returns(JatsMetadata) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = false

    sig { returns(T::Boolean) }
    def docx? = false

    sig { returns(T::Boolean) }
    def excel? = false

    sig { returns(T::Boolean) }
    def email? = false

    sig { returns(T::Boolean) }
    def pptx? = false

    sig { returns(T::Boolean) }
    def archive? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def xml? = false

    sig { returns(T::Boolean) }
    def text? = false

    sig { returns(T::Boolean) }
    def html? = false

    sig { returns(T::Boolean) }
    def ocr? = false

    sig { returns(T::Boolean) }
    def csv? = false

    sig { returns(T::Boolean) }
    def bibtex? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def fiction_book? = false

    sig { returns(T::Boolean) }
    def dbf? = false

    sig { returns(T::Boolean) }
    def jats? = true

    sig { returns(T::Boolean) }
    def epub? = false

    sig { returns(T::Boolean) }
    def pst? = false

    sig { returns(T::Boolean) }
    def code? = 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

  # Variant FormatMetadataEpub of the FormatMetadata sum type.
  FormatMetadataEpub = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [EpubMetadata]
    sig { returns(EpubMetadata) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = false

    sig { returns(T::Boolean) }
    def docx? = false

    sig { returns(T::Boolean) }
    def excel? = false

    sig { returns(T::Boolean) }
    def email? = false

    sig { returns(T::Boolean) }
    def pptx? = false

    sig { returns(T::Boolean) }
    def archive? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def xml? = false

    sig { returns(T::Boolean) }
    def text? = false

    sig { returns(T::Boolean) }
    def html? = false

    sig { returns(T::Boolean) }
    def ocr? = false

    sig { returns(T::Boolean) }
    def csv? = false

    sig { returns(T::Boolean) }
    def bibtex? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def fiction_book? = false

    sig { returns(T::Boolean) }
    def dbf? = false

    sig { returns(T::Boolean) }
    def jats? = false

    sig { returns(T::Boolean) }
    def epub? = true

    sig { returns(T::Boolean) }
    def pst? = false

    sig { returns(T::Boolean) }
    def code? = 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

  # Variant FormatMetadataPst of the FormatMetadata sum type.
  FormatMetadataPst = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [PstMetadata]
    sig { returns(PstMetadata) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = false

    sig { returns(T::Boolean) }
    def docx? = false

    sig { returns(T::Boolean) }
    def excel? = false

    sig { returns(T::Boolean) }
    def email? = false

    sig { returns(T::Boolean) }
    def pptx? = false

    sig { returns(T::Boolean) }
    def archive? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def xml? = false

    sig { returns(T::Boolean) }
    def text? = false

    sig { returns(T::Boolean) }
    def html? = false

    sig { returns(T::Boolean) }
    def ocr? = false

    sig { returns(T::Boolean) }
    def csv? = false

    sig { returns(T::Boolean) }
    def bibtex? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def fiction_book? = false

    sig { returns(T::Boolean) }
    def dbf? = false

    sig { returns(T::Boolean) }
    def jats? = false

    sig { returns(T::Boolean) }
    def epub? = false

    sig { returns(T::Boolean) }
    def pst? = true

    sig { returns(T::Boolean) }
    def code? = 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

  # Variant FormatMetadataCode of the FormatMetadata sum type.
  FormatMetadataCode = Data.define(:value) do
    include FormatMetadata
    extend T::Sig

    # @return [String]
    sig { returns(String) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def pdf? = false

    sig { returns(T::Boolean) }
    def docx? = false

    sig { returns(T::Boolean) }
    def excel? = false

    sig { returns(T::Boolean) }
    def email? = false

    sig { returns(T::Boolean) }
    def pptx? = false

    sig { returns(T::Boolean) }
    def archive? = false

    sig { returns(T::Boolean) }
    def image? = false

    sig { returns(T::Boolean) }
    def xml? = false

    sig { returns(T::Boolean) }
    def text? = false

    sig { returns(T::Boolean) }
    def html? = false

    sig { returns(T::Boolean) }
    def ocr? = false

    sig { returns(T::Boolean) }
    def csv? = false

    sig { returns(T::Boolean) }
    def bibtex? = false

    sig { returns(T::Boolean) }
    def citation? = false

    sig { returns(T::Boolean) }
    def fiction_book? = false

    sig { returns(T::Boolean) }
    def dbf? = false

    sig { returns(T::Boolean) }
    def jats? = false

    sig { returns(T::Boolean) }
    def epub? = false

    sig { returns(T::Boolean) }
    def pst? = false

    sig { returns(T::Boolean) }
    def code? = 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

module Kreuzberg
  # Bounding geometry for an OCR element.
  #
  # Supports both axis-aligned rectangles (from Tesseract) and 4-point quadrilaterals
  # (from PaddleOCR and rotated text detection).
  module OcrBoundingGeometry
    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 "rectangle" then OcrBoundingGeometryRectangle.from_hash(hash)
      when "quadrilateral" then OcrBoundingGeometryQuadrilateral.from_hash(hash)
      else raise "Unknown discriminator: #{discriminator}"
      end
    end
  end

  # Axis-aligned bounding box (typical for Tesseract output).
  OcrBoundingGeometryRectangle = Data.define(:left, :top, :width, :height) do
    include OcrBoundingGeometry
    extend T::Sig

    # Left x-coordinate in pixels
    sig { returns(Integer) }
    def left = super # rubocop:disable Lint/UselessMethodDefinition

    # Top y-coordinate in pixels
    sig { returns(Integer) }
    def top = super # rubocop:disable Lint/UselessMethodDefinition

    # Width in pixels
    sig { returns(Integer) }
    def width = super # rubocop:disable Lint/UselessMethodDefinition

    # Height in pixels
    sig { returns(Integer) }
    def height = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def rectangle? = true

    sig { returns(T::Boolean) }
    def quadrilateral? = 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(left: hash[:left] || hash["left"], top: hash[:top] || hash["top"], width: hash[:width] || hash["width"], height: hash[:height] || hash["height"])
    end
  end

  # 4-point quadrilateral for rotated/skewed text (PaddleOCR).
  #
  # Points are in clockwise order starting from top-left:
  # `[top_left, top_right, bottom_right, bottom_left]`
  OcrBoundingGeometryQuadrilateral = Data.define(:points) do
    include OcrBoundingGeometry
    extend T::Sig

    # Four corner points as `[(x, y), ...]` in clockwise order
    sig { returns(String) }
    def points = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def rectangle? = false

    sig { returns(T::Boolean) }
    def quadrilateral? = 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(points: hash[:points] || hash["points"])
    end
  end
end

module Kreuzberg
  # A single line in a unified-diff hunk.
  #
  # Defined here (rather than only in `crate::diff`) so `RevisionDelta` can
  # reference it unconditionally, without requiring the `diff` Cargo feature.
  # `crate::diff` re-exports this type verbatim.
  module DiffLine
    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[:kind] || hash["kind"]
      case discriminator
      when "context" then DiffLineContext.from_hash(hash)
      when "added" then DiffLineAdded.from_hash(hash)
      when "removed" then DiffLineRemoved.from_hash(hash)
      else raise "Unknown discriminator: #{discriminator}"
      end
    end
  end

  # Unchanged context line.
  DiffLineContext = Data.define(:value) do
    include DiffLine
    extend T::Sig

    # @return [String]
    sig { returns(String) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def context? = true

    sig { returns(T::Boolean) }
    def added? = false

    sig { returns(T::Boolean) }
    def removed? = 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

  # Line added in the "after" version.
  DiffLineAdded = Data.define(:value) do
    include DiffLine
    extend T::Sig

    # @return [String]
    sig { returns(String) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def context? = false

    sig { returns(T::Boolean) }
    def added? = true

    sig { returns(T::Boolean) }
    def removed? = 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

  # Line removed from the "before" version.
  DiffLineRemoved = Data.define(:value) do
    include DiffLine
    extend T::Sig

    # @return [String]
    sig { returns(String) }
    def value = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def context? = false

    sig { returns(T::Boolean) }
    def added? = false

    sig { returns(T::Boolean) }
    def removed? = 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

module Kreuzberg
  # Best-effort document location for a revision.
  module RevisionAnchor
    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 "paragraph" then RevisionAnchorParagraph.from_hash(hash)
      when "table_cell" then RevisionAnchorTableCell.from_hash(hash)
      when "page" then RevisionAnchorPage.from_hash(hash)
      when "slide" then RevisionAnchorSlide.from_hash(hash)
      when "sheet" then RevisionAnchorSheet.from_hash(hash)
      else raise "Unknown discriminator: #{discriminator}"
      end
    end
  end

  # Body paragraph, identified by its zero-based index in the document flow.
  RevisionAnchorParagraph = Data.define(:index) do
    include RevisionAnchor
    extend T::Sig

    # Zero-based index of the paragraph in document order.
    sig { returns(Integer) }
    def index = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def paragraph? = true

    sig { returns(T::Boolean) }
    def table_cell? = false

    sig { returns(T::Boolean) }
    def page? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def sheet? = 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(index: hash[:index] || hash["index"])
    end
  end

  # Cell inside a table.
  RevisionAnchorTableCell = Data.define(:row, :col, :table_index) do
    include RevisionAnchor
    extend T::Sig

    # Zero-based row index within the table.
    sig { returns(Integer) }
    def row = super # rubocop:disable Lint/UselessMethodDefinition

    # Zero-based column index within the table.
    sig { returns(Integer) }
    def col = super # rubocop:disable Lint/UselessMethodDefinition

    # Zero-based index of the table in document order.
    sig { returns(Integer) }
    def table_index = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def table_cell? = true

    sig { returns(T::Boolean) }
    def page? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def sheet? = 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(row: hash[:row] || hash["row"], col: hash[:col] || hash["col"], table_index: hash[:table_index] || hash["table_index"])
    end
  end

  # Page, identified by its zero-based index.
  RevisionAnchorPage = Data.define(:index) do
    include RevisionAnchor
    extend T::Sig

    # Zero-based page index.
    sig { returns(Integer) }
    def index = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def table_cell? = false

    sig { returns(T::Boolean) }
    def page? = true

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def sheet? = 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(index: hash[:index] || hash["index"])
    end
  end

  # Presentation slide, identified by its zero-based index.
  RevisionAnchorSlide = Data.define(:index) do
    include RevisionAnchor
    extend T::Sig

    # Zero-based slide index.
    sig { returns(Integer) }
    def index = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def table_cell? = false

    sig { returns(T::Boolean) }
    def page? = false

    sig { returns(T::Boolean) }
    def slide? = true

    sig { returns(T::Boolean) }
    def sheet? = 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(index: hash[:index] || hash["index"])
    end
  end

  # Spreadsheet cell or range, identified by sheet index and optional name.
  RevisionAnchorSheet = Data.define(:index, :name) do
    include RevisionAnchor
    extend T::Sig

    # Zero-based sheet index.
    sig { returns(Integer) }
    def index = super # rubocop:disable Lint/UselessMethodDefinition

    # Sheet display name when available.
    sig { returns(T.nilable(String)) }
    def name = super # rubocop:disable Lint/UselessMethodDefinition

    sig { returns(T::Boolean) }
    def paragraph? = false

    sig { returns(T::Boolean) }
    def table_cell? = false

    sig { returns(T::Boolean) }
    def page? = false

    sig { returns(T::Boolean) }
    def slide? = false

    sig { returns(T::Boolean) }
    def sheet? = 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(index: hash[:index] || hash["index"], name: hash[:name] || hash["name"])
    end
  end
end