misaki_ko library
Korean APIs for the pure, backend-injected Misaki g2pkc pipeline.
No MeCab implementation, Korean morphology dictionary, NLTK runtime, or CMUdict resource is bundled. Applications must inject both narrow provider contracts explicitly.
final engine = KoreanG2pkcEngine(
morphology: morphologyBackend,
cmuPronunciations: cmuPronunciationProvider,
);
final result = engine.convert('안녕하세요.');
print(result.phonemes);
Classes
- AsyncG2pEngine
- A reusable engine whose backend requires asynchronous work.
- BackendInfo
- Stable identity information reported by an injected backend.
- EnglishTokenMetadata
- Metadata produced by the English pipeline.
- G2pEngine
- A synchronous, reusable grapheme-to-phoneme engine.
- G2pResult
- Exact output from one grapheme-to-phoneme conversion.
- JapaneseTokenMetadata
- Metadata produced by the pyopenjtalk-style Japanese pipeline.
- KokoroEnglishG2pFrontend
- Kokoro 0.9.4's token-aware English G2P frontend.
- KokoroG2pChunk
- One text-and-phoneme chunk ready for Kokoro model input preparation.
- KokoroNonEnglishG2pFrontend
- Kokoro 0.9.4's pre-G2P chunking contract for non-English languages.
- KoreanCmuPronunciation
- One immutable first-choice CMUdict pronunciation.
- KoreanCmuPronunciationProvider
- Explicit CMUdict lookup boundary used by g2pkc English conversion.
- KoreanG2pkcEngine
- Pure deterministic Korean g2pkc pipeline with injected external data.
- KoreanMorphologyBackend
-
Explicit Korean morphology boundary matching
mecab.MeCab().pos. - KoreanMorphologyToken
-
One immutable
(surface, POS tag)result from Korean morphology. - MisakiBackend
- Common contract for explicitly supplied language backends.
- MisakiToken
- Immutable token shared by all language engines.
- TokenMetadata
- Language-specific, typed token details.
- UnknownMarkerG2pEngine
- A synchronous G2P engine whose unresolved-token marker is inspectable.
- VietnameseTokenMetadata
- Metadata produced by the Vietnamese pipeline.
Constants
- defaultUnknownMarker → const String
- Default replacement for text that a selected engine cannot phonemize.
- misakiUpstreamCommit → const String
- The authoritative upstream revision for fixtures and copied data.
- misakiUpstreamRepository → const String
- The upstream project whose observable behavior this package ports.
- misakiUpstreamVersion → const String
- The Python package version at misakiUpstreamCommit.
Functions
-
composeKoreanJamo(
String input) → String - Composes modern conjoining Jamo using the pinned g2pkc helper semantics.
-
convertKoreanNumerals(
String input) → String - Converts all annotated numeral spans using pinned g2pkc semantics.
-
decomposeKoreanHangul(
String input) → String - Decomposes modern precomposed Hangul syllables into conjoining Jamo.
-
koreanPhonemeInventory(
) → Set< String> - Returns the frozen modern conjoining-Jamo inventory emitted by g2pkc.
-
spellKoreanNumber(
String number, {bool sino = true}) → String - Spells one comma-grouped decimal integer using pinned g2pkc number rules.
Exceptions / Errors
- BackendFailureException
- A configured backend failed while processing input.
- A required explicitly configured backend is not available.
- InvalidConfigurationException
- The selected options do not form a supported configuration.
- MalformedDataException
- Generated or copied runtime data failed validation.
- MisakiException
- Base class for failures exposed by Misaki Dart APIs.