Crate gukhanmun_core 

Core types and algorithms for Gukhanmun.

This crate is the home for the format-neutral intermediate representation, conversion engine, dictionary traits, lattice segmentation, and fallback hanja reading logic. Format adapters, command-line I/O, and language bindings live in separate crates.

Structs

Annotation

Metadata for a dictionary-backed hanja conversion.

ChainDictionary

A dictionary composition that preserves caller-supplied priority order.

DictionaryRecord

A complete dictionary entry exposed for batch policy analysis.

Engine

Stateful hanja conversion engine for chunked token streams.

EngineOptions

Engine-level options that affect hanja conversion before rendering.

FirstOccurrenceFilter

Streaming first-occurrence middleware.

HomophoneMarker

Streaming homophone marker middleware.

MapDictionary

A small in-memory dictionary backed by an ordered map.

Match

A dictionary match that starts at the queried cursor position.

MatchMark

Dictionary-provided rendering constraints for a match.

PlainScopeData

Scope data used by the plain-text adapter.

RecoverableInputError

A recoverable reader error plus the original source region.

RenderOptions

Rendering options that combine a RenderMode with per-mode sub-options.

Renderer

Stateful renderer for chunked OutputToken streams.

Scope

A structural scope in the format-neutral token stream.

UnihanCharDict

Per-character Unihan fallback readings exposed as a dictionary.

UserDirectives

User rules that adjust annotation presentation policy.

Enums

ContextWindow

The context boundary used by stateful annotation middlewares.

DirectiveAction

Action applied when a user directive predicate matches an annotation.

Error

Error returned by fallible core pipeline entry points.

InputToken

A token emitted by a reader before hanja conversion has run.

NumeralStrategy

Strategy for rendering hanja numerals encountered in fallback text.

OriginalGloss

Form for the gloss attached to annotations in RenderMode::Original.

OutputToken

A token emitted by the engine after hanja conversion.

Recovery

Stream-level error recovery policy.

RenderMode

The concrete rendering mode for annotated hanja words.

RenderedToken

A token emitted by a renderer after all annotations have been expanded.

RubyBase

Selects which side of a <ruby> element is the base text.

SegmentationStrategy

Strategy used to segment hanja-containing spans.

Traits

HanjaDictionary

A hanja dictionary queried by the conversion engine.

ScopeData

Adapter-owned data attached to an intermediate-representation scope.

Functions

apply_user_directives

Applies literal user directives to annotation policy flags.

apply_user_directives_iter

Lazily applies literal user directives to an output token stream.

convert_plain_text

Converts plain text through reader, engine, renderer, and writer stages.

convert_plain_text_with_options

Converts plain text with explicit hanja conversion engine options.

filter_first_occurrences

Clears repeat gloss requirements after the first occurrence of each hanja.

is_hanja

Returns whether ch is in a known CJK ideograph range.

mark_homophones

Sets homophone on dictionary annotations sharing a reading.

process_fallible_tokens

Processes fallible input tokens with default engine options.

process_fallible_tokens_with_options

Processes fallible input tokens with explicit engine options.

process_tokens

Processes input tokens with the default hanja conversion engine options.

process_tokens_iter

Processes input tokens through the default engine options and returns an iterator over the collected output.

process_tokens_iter_with_options

Processes input tokens through explicit engine options and returns an iterator over the collected output.

process_tokens_with_options

Processes input tokens with explicit hanja conversion engine options.

read_plain_text

Reads a plain-text string into the core input-token stream.

recover_input_token

Resolves one fallible reader item according to a Recovery policy.

recover_input_tokens

Resolves a fallible reader token stream into recovered input tokens.

render_tokens

Renders engine output tokens into annotation-free tokens.

render_tokens_iter

Renders engine output tokens into annotation-free tokens as an iterator.

write_plain_text

Writes rendered plain-text tokens back to a string.