RcppMeCab

This repo is a fork from junhewk/RcppMeCab.

As of v0.1.0, this fork has been changed to wrap the gibasa package, instead of wrapping the ‘MeCab’ API via ‘Rcpp’ by itself. If you need the old version, use v0.0.x branch.

---

This package, RcppMeCab, is an ‘Rcpp’ wrapper for the part-of-speech morphological analyzer MeCab. It supports native utf-8 encoding in C++ code and CJK (Chinese, Japanese, and Korean) MeCab library. This package fully utilizes the power Rcpp brings R computation to analyze texts faster.

Installation

# Enable repository from paithiov909
options(repos = c(
  paithiov909 = "https://paithiov909.r-universe.dev",
  CRAN = "https://cloud.r-project.org"))

# Download and install RcppMeCab in R
install.packages("RcppMeCab")

# Or build from source package
Sys.setenv(MECAB_DEFAULT_RC = "/fullpath/to/your/mecabrc") # if necessary
remotes::install_github("paithiov909/RcppMeCab")

To use this package requires the MeCab library and its dictionary installed and available.

In case using Linux or OSX, you can install them with their package managers, or build and install from the source by yourself.

In case using Windows, use installer built for 32bit or built for 64bit.

Usage

This package has pos and posParallel function.

sentence <- "雨にも負けず、風にも負けず"

## テキストだけ与える場合、デフォルトの戻り値はnamed list of character vectors.
## リストの各要素は、表層形（surface form）と素性情報の1番目（IPA辞書では「品詞」）を'/'で区切ってつなげた文字列ベクトルになる。
pos(sentence) # returns list
#> Warning: This fork does not support `pos()`. Using `posParallel()` instead.
#> $`1`
#>  [1] "雨/名詞"   "に/助詞"   "も/助詞"   "負け/動詞" "ず/助動詞" "、/記号"  
#>  [7] "風/名詞"   "に/助詞"   "も/助詞"   "負け/動詞" "ず/助動詞"

## 'join = FALSE'を指定すると、戻り値はnamed list of named character vectorsになる。
## Neologd辞書などでは収録されている語彙そのものに'/'が含まれていることがあるため、使用ケースによって使い分けるとよい。
pos(sentence, join = FALSE) # for yielding morphemes only (tags will be given on the vector names)
#> Warning: This fork does not support `pos()`. Using `posParallel()` instead.
#> $`1`
#>   名詞   助詞   助詞   動詞 助動詞   記号   名詞   助詞   助詞   動詞 助動詞 
#>   "雨"   "に"   "も" "負け"   "ず"   "、"   "風"   "に"   "も" "負け"   "ず"

## 'format = data.frame'にすると、戻り値は以下のようなデータフレームになる。
## pos列・subtype列は素性情報の1~2番目（IPA辞書では「品詞」と「品詞細分類1」）、
## analytic列は素性情報の8番目（IPA辞書の「読み」）だが、
## 未知語で推定されない素性だった場合などには'NA_character_'が含まれることがある。
pos(sentence, format = "data.frame") # the result will returned as a data frame format
#> Warning: This fork does not support `pos()`. Using `posParallel()` instead.
#> # A tibble: 11 × 7
#>    doc_id sentence_id token_id token pos    subtype analytic 
#>    <fct>        <int>    <int> <chr> <chr>  <chr>   <chr>    
#>  1 1                1        1 雨    名詞   一般    アメ,アメ
#>  2 1                1        2 に    助詞   格助詞  ニ,ニ    
#>  3 1                1        3 も    助詞   係助詞  モ,モ    
#>  4 1                1        4 負け  動詞   自立    マケ,マケ
#>  5 1                1        5 ず    助動詞 <NA>    ズ,ズ    
#>  6 1                1        6 、    記号   読点    、,、    
#>  7 1                1        7 風    名詞   一般    カゼ,カゼ
#>  8 1                1        8 に    助詞   格助詞  ニ,ニ    
#>  9 1                1        9 も    助詞   係助詞  モ,モ    
#> 10 1                1       10 負け  動詞   自立    マケ,マケ
#> 11 1                1       11 ず    助動詞 <NA>    ズ,ズ

posParallel(sentence) # parallelized version uses more memory, but much faster than the loop in single threading
#> $`1`
#>  [1] "雨/名詞"   "に/助詞"   "も/助詞"   "負け/動詞" "ず/助動詞" "、/記号"  
#>  [7] "風/名詞"   "に/助詞"   "も/助詞"   "負け/動詞" "ず/助動詞"

both two funtions can get same arguments:

• sentence: a text for analyzing
• join: If it gets TRUE, output form is (morpheme/tag). If it gets FALSE, output form is (morpheme) + tag in attribute.
• format: The default is a list. If you set this as "data.frame", the function will return the result in a data frame format.
• sys_dic: a directory in which dicrc file is located, default value is NULL or you can set your default value using options(mecabSysDic = "/path/to/your/system_dictionary")
• user_dic: a user dictionary file compiled by mecab_dict_index, default value is also “”

Changes from the origin

• When format="data.frame", the function returns ‘doc_id’ and ‘sentence_id’ as factors.
• When options(mecabSplit = TRUE), the function splits and flattens character vector into sentences with stringi::stri_split_boundaries(type = "sentence") before analyze them.
• Fixed the behavior when input sentence is an unnamed vector.
• Refactored C++ source for performance.
• Included MeCab source code in package src directory for installation.

License

GPL (>=3)