Skip to content

shirakaba/iPhone-libmecab

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 

Repository files navigation

Japanese and Korean Morphological Analyzer Mecab on iOS and macOS

This is a project that enables mecab-ko (a fork of Mecab that adds support for Korean) to be easily integrated into an iOS or macOS app.

Updated to work with Xcode 11, iOS 9+, macOS 10.11+, and Swift 5.

Sample app

The sample app in LibMecabSample should build both an iOS Obj-C app (with a UI) and a macOS Swift app (without any UI; check the console for results).

Japanese

Korean

Installing from Cocoapods

Specify these pods in your Podfile (you can omit the Japanese or Korean dictionary if you only plan to use Mecab with one of the two languages):

# Fork of Mecab supporting both Japanese and Korean
pod 'mecab-ko'

# Japanese dictionary
pod 'mecab-naist-jdic-utf-8'

# Korean dictionary
pod 'mecab-ko-dic-utf-8'
pod update

Installing as a Cocoapod from npm (for React Native iOS apps)

Add these npm packages (you can omit the Japanese or Korean dictionary if you only plan to use Mecab with one of the two languages):

yarn add mecab-ko mecab-naist-jdic mecab-ko-dic

# or:

npm install --save mecab-ko mecab-naist-jdic mecab-ko-dic

# WHY THE DICTIONARY NPM PACKAGE NAMES DIFFER FROM THOSE OF THE COCOAPODS:
# I had to remove the "-utf-8" suffix from them for npm publishing to succeed.

Next, specify these pods in your Podfile (you can omit the Japanese or Korean dictionary if you only plan to use Mecab with one of the two languages):

# Fork of Mecab supporting both Japanese and Korean
pod 'mecab-ko', :podspec => '../node_modules/mecab-ko/mecab-ko.podspec'

# Japanese dictionary
pod 'mecab-naist-jdic-utf-8', :podspec => '../node_modules/mecab-naist-jdic/mecab-naist-jdic-utf-8.podspec'

# Korean dictionary
pod 'mecab-ko-dic-utf-8', :podspec => '../node_modules/mecab-ko-dic/mecab-ko-dic-utf-8.podspec'

# WHY THE DICTIONARY NPM PACKAGE NAMES DIFFER FROM THOSE OF THE COCOAPODS:
# I had to remove the "-utf-8" suffix from them for npm publishing to succeed.

Don't forget to install the pods.

cd ios
pod update

Usage

Import the necessary mecab_ko headers into your class. Allocate and initialize a new Mecab object (specifying whether to use the Japanese or Korean dictionary via the DEFAULT_JAPANESE_RESOURCES_BUNDLE_NAME or DEFAULT_KOREAN_RESOURCES_BUNDLE_NAME constants) and then supply it a string to parse via the parseToNodeWithString method. It'll return an array of nodes that you can then manipulate as needed:

Swift invocation

Bridging header

If you have use_frameworks! enabled in your Podfile, then you don't need to include a bridging header.

If you don't have it enabled (because you're using only static libraries), then add these imports to your bridging header to expose the Obj-C headers to the Swift runtime:

// LibMecabSample-macos-Bridging-Header.h
// Note how the module is `mecab-ko` in this Obj-C header but `mecab_ko` in Swift!
#import <mecab-ko/MecabObjC.h>
#import <mecab-ko/MecabNode.h>

Swift file

import mecab_ko // Omit this if you have `use_frameworks!` enabled in your Podfile.

// ...

let jpBundlePath = Bundle.main.path(forResource: DEFAULT_JAPANESE_RESOURCES_BUNDLE_NAME, ofType: "bundle")
let jpBundleResourcePath = Bundle.init(path: jpBundlePath!)!.resourcePath

let mecabJapanese: Mecab = Mecab.init(dicDirPath: jpBundleResourcePath!)
let japaneseNodes: [MecabNode]? = mecabJapanese.parseToNode(with: "すもももももももものうち")
japaneseNodes?.forEach({ node in print("[\(node.surface)] \(node.feature ?? "")") })

Obj-C invocation

// If use_frameworks! is on, you may have to write mecab_ko instead of mecab-ko.
// This is because Clang modules don't support hyphens:
// http://blog.cocoapods.org/Pod-Authors-Guide-to-CocoaPods-Frameworks/
#import <mecab-ko/MecabObjC.h>
#import <mecab-ko/MecabNode.h>

// ...

NSString *jpDicBundlePath = [[NSBundle mainBundle] pathForResource:DEFAULT_JAPANESE_RESOURCES_BUNDLE_NAME ofType:@"bundle"];
NSString *jpDicBundleResourcePath = [[NSBundle alloc] initWithPath:jpDicBundlePath].resourcePath;

self.mecab = [[Mecab alloc] initWithDicDirPath:jpDicBundleResourcePath];
NSArray<MecabNode *> results = [mecab parseToNodeWithString:@"すもももももももものうち"];

This would give you an array of nodes with the following raw metadata:

すもも: 名詞,一般,*,*,*,*,すもも,スモモ,スモモ  
も: 助詞,係助詞,*,*,*,*,も,モ,モ  
もも: 名詞,一般,*,*,*,*,もも,モモ,モモ  
も: 助詞,係助詞,*,*,*,*,も,モ,モ  
もも: 名詞,一般,*,*,*,*,もも,モモ,モモ  
の: 助詞,連体化,*,*,*,*,の,ノ,ノ  
うち: 名詞,非自立,副詞可能,*,*,*,うち,ウチ,ウチ

You can access each piece of metadata by getting the full comma-separated feature string node.feature or indexing into the pre-processed features array, node.features.

@interface MecabNode : NSObject {
    /* What text the node consists of. */
    NSString *surface;

    /* The raw comma-separated string of metadata about the node. */
    NSString *feature;

    /* All the comma-separated features converted into an array, for 
     * direct access. */
    NSArray<NSString *> *features;

    /* The leading whitespace in front of this node (strictly 0 for the
     * first node, as Mecab always trims whitespace at the start of input
     * text). */
    int leadingWhitespaceLength;

    /* Trailing whitespace is only calculated if passing YES into
     * calculateTrailingWhitespace when calling this method on the mecab
     * instance:
     * 
     * - (NSArray<MecabNode *> *)parseToNodeWithString:(NSString *)string calculateTrailingWhitespace:(BOOL)calculateTrailingWhitespace;
     *
     */
    NSString *trailingWhitespace;
}

Examples

Japanese

Feature schema

It's based on IPADIC, and the tags are fully documented here https://osdn.net/projects/ipadic/docs (English manual):

index role (Japanese) role (English)
0 品詞 part-of-speech
1 品詞細分類1 part-of-speech subtype 1
2 品詞細分類2 part-of-speech subtype 2
3 品詞細分類3 part-of-speech subtype 3
4 活用形 inflected form
5 活用型 inflection type
6 原形 original form
7 読み reading
8 発音 pronunciation

⚠️ When accessing these features, e.g. feature[3], most of them safely return "*" when there is no data. However, certain ones may return null for certain words (usually these are corner-cases like pronunciation for foreign tokens). As I'm not too sure which ones return null, so please treat everything as potentially null to avoid crashes!

Example feature strings

欲しがっていた

欲し: 形容詞,自立,*,*,形容詞・イ段,ガル接続,欲しい,ホシ,ホシ  
がっ: 動詞,接尾,*,*,五段・ラ行,連用タ接続,がる,ガッ,ガッ  
て: 助詞,接続助詞,*,*,*,*,て,テ,テ  
い: 動詞,非自立,*,*,一段,連用形,いる,イ,イ  
た: 助動詞,*,*,*,特殊・タ,基本形,た,タ,タ  

通ったんだろうな

通っ: 動詞,自立,*,*,五段・ラ行,連用タ接続,通る,トオッ,トーッ  
た: 助動詞,*,*,*,特殊・タ,基本形,た,タ,タ  
ん: 名詞,非自立,一般,*,*,*,ん,ン,ン  
だろ: 助動詞,*,*,*,特殊・ダ,未然形,だ,ダロ,ダロ  
う: 助動詞,*,*,*,不変化型,基本形,う,ウ,ウ  
な: 助詞,終助詞,*,*,*,*,な,ナ,ナ  

光らせておくように

光らせ: 動詞,自立,*,*,一段,連用形,光らせる,ヒカラセ,ヒカラセ  
て: 助詞,接続助詞,*,*,*,*,て,テ,テ  
おく: 動詞,非自立,*,*,五段・カ行イ音便,基本形,おく,オク,オク  
よう: 名詞,非自立,助動詞語幹,*,*,*,よう,ヨウ,ヨー  
に: 助詞,副詞化,*,*,*,*,に,ニ,ニ  

Korean

Feature schema

Information about the dictionary format and part-of-speech tags used by mecab-ko-dic is documented in this Google Spreadsheet, linked to from mecab-ko-dic's repo readme.

Note how ko-dic has one less feature column than NAIST JDIC, and has an altogether different set of information (e.g. doesn't provide the "original form" of the morpheme).

The tags are a slight modification of those specified by 세종 (Sejong), whatever that is. The mappings from Sejong to mecab-ko-dic's tag names are given in tab "태그 v2.0" on the above-linked spreadsheet.

The dictionary format is specified fully (in Korean) in tab "사전 형식 v2.0" of the spreadsheet. Any blank values default to *.

index role (Korean) role (English) notes
0 품사 태그 part-of-speech tag See "태그 v2.0" tab on spreadsheet
1 의미 부류 meaning (too few examples for me to be sure)
2 종성 유무 presence or absence T for true; F for false; else *
3 읽기 reading usually matches surface, but may differ for foreign words e.g. Chinese character words
4 타입 type One of: Inflect (활용); Compound (복합명사); or Preanalysis (기분석)
5 첫번째 품사 first part-of-speech e.g. given a part-of-speech tag of "VV+EM+VX+EP", would return VV
6 마지막 품사 last part-of-speech e.g. given a part-of-speech tag of "VV+EM+VX+EP", would return EP
7 표현 expression "활용, 복합명사, 기분석이 어떻게 구성되는지 알려주는 필드" – Fields that tell how usage, compound nouns, and key analysis are organized

⚠️ As I'm mentioned for NAIST JDic above, I am unsure whether any of these fields may return null in certain circumstances. It is safest to handle all fields as if they may be optional in practice.

Example feature strings

mecab-ko-dic은 MeCab을 사용하여, 한국어 형태소 분석을 하기 위한 프로젝트입니다.

mecab    SL,*,*,*,*,*,*,*
-    SY,*,*,*,*,*,*,*
ko    SL,*,*,*,*,*,*,*
-    SY,*,*,*,*,*,*,*
dic    SL,*,*,*,*,*,*,*
은    JX,*,T,은,*,*,*,*
MeCab    SL,*,*,*,*,*,*,*
을    JKO,*,T,을,*,*,*,*
사용    NNG,행위,T,사용,*,*,*,*
하    XSV,*,F,하,*,*,*,*
여    EC,*,F,여,*,*,*,*
,    SC,*,*,*,*,*,*,*
한국어    NNG,*,F,한국어,Compound,*,*,한국/NNG/*+어/NNG/*
형태소    NNG,*,F,형태소,Compound,*,*,형태/NNG/*+소/NNG/*
분석    NNG,행위,T,분석,*,*,*,*
을    JKO,*,T,을,*,*,*,*
하    VV,*,F,하,*,*,*,*
기    ETN,*,F,기,*,*,*,*
위한    VV+ETM,*,T,위한,Inflect,VV,ETM,위하/VV/*+ᆫ/ETM/*
프로젝트    NNG,*,F,프로젝트,*,*,*,*
입니다    VCP+EF,*,F,입니다,Inflect,VCP,EF,이/VCP/*+ᄇ니다/EF/*
.    SF,*,*,*,*,*,*,*
EOS

License

iPhone-libmecab

Feel free to use this sample project under the terms of the licenses inherited from mecab-ko.

mecab-ko

mecab-ko is free software; I can only specify one license in the metadata for the Cocoapods mecab-ko.podspec and npm package.json, so I have specified BSD, but it can be used under the GPL, LGPL, and/or BSD licenses; please feel free to do so despite the limitations of what I can write into the metadata.

For details, please check the COPYING, GPL, LGPL, and BSD files in mecab-ko/Assets.

mecab-ko-dic-utf-8

mecab-ko-dic-utf-8 is available only under the Apache 2.0 licence: mecab-ko-dic-utf-8/bundleContents/COPYING.

mecab-naist-jdic-utf-8

mecab-naist-jdic-utf-8 is available only under the BSD licence: mecab-naist-jdic-utf-8/bundleContents/COPYING.

Contributors

This repository has had a long history (9 years?), being contributed to by (to my knowledge):

  • Toshinori Watanabe (link to fork now broken, but was re-hosted at some point by Richard North here)
  • Matthew Long (fork here)
  • Jamie Birch (the current fork)

Special thank you to Joseph Toronto on StackOverflow for assisting both Matthew Long and myself by providing setup steps.

Acknowledgments

This repository is based on:

Mecab
iPhone-libmecab

See also

About

A sample iOS application utilizing the Mecab morphological analyzer as a static library.

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Objective-C 52.7%
  • Swift 23.2%
  • C++ 19.3%
  • Ruby 4.8%