kiwi

package module
v0.0.0-...-8eb509f Latest Latest
Warning

This package is not in the latest version of its module.

 Go to latest Published: Oct 15, 2025 License: LGPL-2.1 Imports: 6 Imported by: 0

README

kiwigo

CI Go Reference

https://github.com/bab2min/Kiwi for go

Dependencies

  • Install Kiwi

check out how to install Kiwi here

Example script (install using the helper script)
make install-kiwi
Example script (install from source)
git clone [email protected]:bab2min/Kiwi.git
cd Kiwi
git submodule sync
git submodule update --init --recursive
mkdir build
cd build
cmake -DCMAKE_BUILD_TYPE=Release ../
make
make install
ldconfig

Documentation

Overview

Package kiwi is a Go binding for Kiwi (https://github.com/bab2min/Kiwi) project.

Example 
package main

import (
	"fmt"

	kiwi "github.com/codingpot/kiwigo"
)

func main() {
	kb := kiwi.NewBuilder("./base", 1 /*=numThread*/, kiwi.KIWI_BUILD_INTEGRATE_ALLOMORPH /*=options*/)
	kb.AddWord("코딩냄비", "NNP", 0)

	k := kb.Build()
	defer k.Close() // don't forget to Close()!

	results, _ := k.Analyze("안녕하세요 코딩냄비입니다. 부글부글.", 1 /*=topN*/, kiwi.KIWI_MATCH_ALL)

	// Print tokens without the score to avoid floating-point output issues
	if len(results) > 0 {
		fmt.Printf("Tokens: %v\n", results[0].Tokens)
	}
}

Output:

Tokens: [{0 NNG 안녕} {2 XSA 하} {3 EF 세요} {6 NNP 코딩냄비} {10 VCP 이} {10 EF ᆸ니다} {13 SF .} {15 MAG 부글부글} {19 SF .}]

Index

Examples

Constants

This section is empty.

Variables

This section is empty.

Functions

func KiwiClearError 

func KiwiClearError()

KiwiClearError clear error.

func KiwiError 

func KiwiError() string

KiwiError returns the Error messages.

func KiwiReaderImpl 

func KiwiReaderImpl(lineNumber C.int, buffer *C.char, userData unsafe.Pointer) C.int

func KiwiVersion 

func KiwiVersion() string

KiwiVersion returns the version of the kiwi library.

Types

type AnalyzeOption 

type AnalyzeOption int

AnalyzeOption is a bitwise OR of the KiwiAnalyzeOption values. 

const (
	KIWI_MATCH_URL                  AnalyzeOption = C.KIWI_MATCH_URL
	KIWI_MATCH_EMAIL                AnalyzeOption = C.KIWI_MATCH_EMAIL
	KIWI_MATCH_HASHTAG              AnalyzeOption = C.KIWI_MATCH_HASHTAG
	KIWI_MATCH_MENTION              AnalyzeOption = C.KIWI_MATCH_MENTION
	KIWI_MATCH_ALL                  AnalyzeOption = C.KIWI_MATCH_ALL
	KIWI_MATCH_NORMALIZE_CODA       AnalyzeOption = C.KIWI_MATCH_NORMALIZE_CODA
	KIWI_MATCH_ALL_WITH_NORMALIZING AnalyzeOption = C.KIWI_MATCH_ALL_WITH_NORMALIZING
)

type BuildOption 

type BuildOption int

BuildOption is a bitwise OR of the KiwiBuildOption values. 

const (
	KIWI_BUILD_LOAD_DEFAULT_DICT   BuildOption = C.KIWI_BUILD_LOAD_DEFAULT_DICT
	KIWI_BUILD_INTEGRATE_ALLOMORPH BuildOption = C.KIWI_BUILD_INTEGRATE_ALLOMORPH
	KIWI_BUILD_DEFAULT             BuildOption = C.KIWI_BUILD_DEFAULT
)

type Kiwi 

type Kiwi struct {
	// contains filtered or unexported fields
}

Kiwi is a wrapper for the kiwi C library.

func New 

func New(modelPath string, numThread int, options BuildOption) *Kiwi

New returns a new Kiwi instance. Don't forget to call Close after this.

func (*Kiwi) Analyze 

func (k *Kiwi) Analyze(text string, topN int, options AnalyzeOption) ([]TokenResult, error)

Analyze returns the result of the analysis.

func (*Kiwi) Close 

func (k *Kiwi) Close() int

Close frees the resource allocated for Kiwi and returns the exit status. This must be called after New. Returns 0 if successful. Safe to call multiple times.

func (*Kiwi) SplitSentence 

func (k *Kiwi) SplitSentence(text string, options AnalyzeOption) ([]SplitResult, error)

SplitSentence returns the line of sentences.

type KiwiBuilder 

type KiwiBuilder struct {
	// contains filtered or unexported fields
}

KiwiBuilder is a wrapper for the kiwi C library.

func NewBuilder 

func NewBuilder(modelPath string, numThread int, options BuildOption) *KiwiBuilder

NewBuilder returns a new KiwiBuilder instance. Don't forget to call Close after this.

func (*KiwiBuilder) AddWord 

func (kb *KiwiBuilder) AddWord(word string, pos POSType, score float32) int

AddWord set custom word with word, pos, score.

func (*KiwiBuilder) Build 

func (kb *KiwiBuilder) Build() *Kiwi

Build creates kiwi instance with user word etc.

func (*KiwiBuilder) Close 

func (kb *KiwiBuilder) Close() int

Close frees the resource allocated for KiwiBuilder and returns the exit status. This must be called after New but not need to called after Build. Returns 0 if successful. Safe to call multiple times.

func (*KiwiBuilder) ExtractWords 

func (kb *KiwiBuilder) ExtractWords(readSeeker io.ReadSeeker, minCnt int, maxWordLen int, minScore float32, posThreshold float32) ([]WordInfo, error)

ExtractWords returns the result of extract word.

func (*KiwiBuilder) LoadDict 

func (kb *KiwiBuilder) LoadDict(dictPath string) int

LoadDict loads user dict with dict file path.

type POSType 

type POSType string
const (
	POS_UNKNOWN POSType = "UN"

	POS_NNG POSType = "NNG"
	POS_NNP POSType = "NNP"
	POS_NNB POSType = "NNB"
	POS_NR  POSType = "NR"
	POS_NP  POSType = "NP"

	POS_VV  POSType = "VV"
	POS_VA  POSType = "VA"
	POS_VX  POSType = "VX"
	POS_VCP POSType = "VCP"
	POS_VCN POSType = "VCN"

	POS_MM POSType = "MM"

	POS_MAG POSType = "MAG"
	POS_MAJ POSType = "MAJ"

	POS_IC POSType = "IC"

	POS_JKS POSType = "JKS"
	POS_JKC POSType = "JKC"
	POS_JKG POSType = "JKG"
	POS_JKO POSType = "JKO"
	POS_JKB POSType = "JKB"
	POS_JKV POSType = "JKV"
	POS_JKQ POSType = "JKQ"
	POS_JX  POSType = "JX"
	POS_JC  POSType = "JC"

	POS_EP  POSType = "EP"
	POS_EF  POSType = "EF"
	POS_EC  POSType = "EC"
	POS_ETN POSType = "ETN"
	POS_ETM POSType = "ETM"

	POS_XPN POSType = "XPN"

	POS_XSN POSType = "XSN"
	POS_XSV POSType = "XSV"
	POS_XSA POSType = "XSA"
	POS_XSM POSType = "XSM"

	POS_XR POSType = "XR"

	POS_SF  POSType = "SF"
	POS_SP  POSType = "SP"
	POS_SS  POSType = "SS"
	POS_SSO POSType = "SSO"
	POS_SSC POSType = "SSC"
	POS_SE  POSType = "SE"
	POS_SO  POSType = "SO"
	POS_SW  POSType = "SW"
	POS_SL  POSType = "SL"
	POS_SH  POSType = "SH"
	POS_SN  POSType = "SN"
	POS_SB  POSType = "SB"

	POS_W_URL     POSType = "W_URL"
	POS_W_EMAIL   POSType = "W_EMAIL"
	POS_W_HASHTAG POSType = "W_HASHTAG"
	POS_W_MENTION POSType = "W_MENTION"
	POS_W_SERIAL  POSType = "W_SERIAL"
	POS_W_EMOJI   POSType = "W_EMOJI"

	POS_Z_CODA POSType = "Z_CODA"
	POS_Z_SIOT POSType = "Z_SIOT"

	POS_USER_0 POSType = "USER0"
	POS_USER_1 POSType = "USER1"
	POS_USER_2 POSType = "USER2"
	POS_USER_3 POSType = "USER3"
	POS_USER_4 POSType = "USER4"

	POS_VV_I  POSType = "VV-I"
	POS_VA_I  POSType = "VA-I"
	POS_VX_I  POSType = "VX-I"
	POS_XSA_I POSType = "XSA-I"

	POS_VV_R  POSType = "VV-R"
	POS_VA_R  POSType = "VA-R"
	POS_VX_R  POSType = "VX-R"
	POS_XSA_R POSType = "XSA-R"

	POS_V POSType = "V"
)

NOTE: update isValid function when adding a new POSType. POS type definitions from Kiwi C++ library: - Type https://github.com/bab2min/Kiwi/blob/18b4062f98cf3beaedac149fc511a2708891f71d/include/kiwi/Types.h#L188-L219 - String https://github.com/bab2min/Kiwi/blob/18b4062f98cf3beaedac149fc511a2708891f71d/src/Utils.cpp#L300

func ParsePOSType 

func ParsePOSType(t string) (POSType, error)

ParsePOSType return POS Tag for resault

type SplitResult 

type SplitResult struct {
	Text  string
	Begin int
	End   int
}

SplitResult returns the Sentences.

type TokenInfo 

type TokenInfo struct {
	// Position is the index of this token appears in the original text.
	Position int

	// Tag represents a type of this token (e.g. VV, NNG, ...).
	Tag POSType

	// Form is the actual string of this token.
	Form string
}

TokenInfo returns the token info for the given token(Str).

type TokenResult 

type TokenResult struct {
	Tokens []TokenInfo
	Score  float32
}

TokenResult is a result for Analyze.

type WordInfo 

type WordInfo struct {
	Form     string
	Freq     int
	POSScore float32
	Score    float32
}

WordInfo returns the token info for the given token(Str).

Source Files

View all Source files

Directories

Path Synopsis

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL
go.dev uses cookies from Google to deliver and enhance the quality of its services and to analyze traffic. Learn more.