v0.0.1.5: mecab-ko 0.999 support, Windows build fix, dict_index()

- Fix Windows build: add $(SHLIB):$(MECAB_OBJS) dependency so MeCab
  .o files are compiled before linking; patch dllimport, _stdcall,
  size_t overload, progress_bar gutting, and HAVE_WINDOWS_H for
  MinGW/Rtools45 compatibility.

- Korean builds now use mecab-ko-msvc 0.999 (Pusnow/mecab-ko-msvc)
  instead of mecab-ko 0.9.2. Japanese builds continue using
  taku910/mecab 0.996. Selected via MECAB_LANG env var (default: ko).

- Add dict_index() R function wrapping mecab_dict_index, allowing
  user dictionary compilation directly from R without the external
  mecab-dict-index command-line tool.

- Update CI to install mecab-ko 0.999 pre-built binaries on Ubuntu
  and macOS for Korean builds.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: junhewk

.github/workflows/R-CMD-check.yml

@@ -29,32 +29,32 @@ jobs:
           sudo apt-get update -y -qq
           sudo apt-get install -y mecab libmecab-dev mecab-ipadic-utf8
           sudo ldconfig
-      - name: Install MeCab (Korean, Ubuntu)
+      - name: Install MeCab-Ko (Korean, Ubuntu)
         if: ${{ matrix.platform == 'ubuntu-latest' && matrix.mecab_lang == 'ko' }}
         run: |
-          sudo apt-get update -y -qq
-          sudo apt-get install -y autoconf automake
-          MECAB_KO_BUILD=$(mktemp -d)
-          cd "$MECAB_KO_BUILD"
-          curl -fsSL "https://github.com/junhewk/RcppMeCab/releases/download/0.0.1.0/mecab-0.996-ko-0.9.2.tar.gz" -o mecab-ko.tar.gz
-          tar xzf mecab-ko.tar.gz
-          cd mecab-0.996-ko-0.9.2
-          ./configure
-          make
-          sudo make install
+          curl -fsSL "https://github.com/Pusnow/mecab-ko-msvc/releases/download/release-0.999/mecab-ko-linux-x86_64.tar.gz" -o mecab-ko.tar.gz
+          sudo tar xzf mecab-ko.tar.gz -C /usr/local --strip-components=1
           sudo ldconfig
-          cd "$MECAB_KO_BUILD"
-          curl -fsSL "https://github.com/junhewk/RcppMeCab/releases/download/0.0.1.0/mecab-ko-dic-2.1.1-20180720.tar.gz" -o mecab-ko-dic.tar.gz
-          tar xzf mecab-ko-dic.tar.gz
-          cd mecab-ko-dic-2.1.1-20180720
-          ./autogen.sh
-          ./configure
-          make
-          sudo make install
-          rm -rf "$MECAB_KO_BUILD"
+          curl -fsSL "https://github.com/Pusnow/mecab-ko-msvc/releases/download/release-0.999/mecab-ko-dic.tar.gz" -o mecab-ko-dic.tar.gz
+          sudo mkdir -p /usr/local/lib/mecab/dic
+          sudo tar xzf mecab-ko-dic.tar.gz -C /usr/local/lib/mecab/dic
       - name: Install MeCab (Japanese, macOS)
         if: ${{ matrix.platform == 'macos-latest' && matrix.mecab_lang == 'ja' }}
         run: brew install mecab mecab-ipadic
+      - name: Install MeCab-Ko (Korean, macOS)
+        if: ${{ matrix.platform == 'macos-latest' && matrix.mecab_lang == 'ko' }}
+        run: |
+          ARCH=$(uname -m)
+          if [ "$ARCH" = "arm64" ]; then
+            VARIANT="macos-arm64"
+          else
+            VARIANT="macos-x86_64"
+          fi
+          curl -fsSL "https://github.com/Pusnow/mecab-ko-msvc/releases/download/release-0.999/mecab-ko-${VARIANT}.tar.gz" -o mecab-ko.tar.gz
+          sudo tar xzf mecab-ko.tar.gz -C /usr/local --strip-components=1
+          curl -fsSL "https://github.com/Pusnow/mecab-ko-msvc/releases/download/release-0.999/mecab-ko-dic.tar.gz" -o mecab-ko-dic.tar.gz
+          sudo mkdir -p /usr/local/lib/mecab/dic
+          sudo tar xzf mecab-ko-dic.tar.gz -C /usr/local/lib/mecab/dic
       - name: Install dependencies
         run: |
           install.packages(c("remotes", "rcmdcheck"))

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: RcppMeCab
 Title: 'rcpp' Wrapper for 'mecab' Library
-Version: 0.0.1.4
+Version: 0.0.1.5
 Authors@R: c(person("Junhewk", "Kim", role = c("aut", "cre"),
                     email = "junhewk.kim@gmail.com"),
              person("Taku", "Kudo", role = c("aut"),
@@ -17,7 +17,7 @@ Depends: R (>= 3.4.0)
 License: GPL
 Encoding: UTF-8
 BugReports: https://github.com/junhewk/RcppMeCab/issues
-RoxygenNote: 7.1.1
+RoxygenNote: 7.3.3
 Language: en-US
 LinkingTo:
     Rcpp,
@@ -30,4 +30,5 @@ Suggests:
     testthat,
     spelling
 SystemRequirements:
-    MeCab 0.996 (or mecab-ko 0.9.2) or higher (libmecab-dev (deb), mecab-devel (rpm))
+    MeCab 0.996 or higher for Japanese (libmecab-dev (deb), mecab-devel (rpm)),
+    mecab-ko 0.999 (https://github.com/Pusnow/mecab-ko-msvc) for Korean

NAMESPACE

@@ -1,5 +1,6 @@
 # Generated by roxygen2: do not edit by hand
 
+export(dict_index)
 export(pos)
 export(posParallel)
 import(Rcpp)

R/RcppExports.R

@@ -1,6 +1,10 @@
 # Generated by using Rcpp::compileAttributes() -> do not edit by hand
 # Generator token: 10BE3573-1514-4C36-9D1C-5A225CD40393
 
+dictIndexRcpp <- function(args) {
+    .Call(`_RcppMeCab_dictIndexRcpp`, args)
+}
+
 posParallelJoinRcpp <- function(text, sys_dic, user_dic) {
     .Call(`_RcppMeCab_posParallelJoinRcpp`, text, sys_dic, user_dic)
 }

R/dict.R

@@ -0,0 +1,64 @@
+#' Compile a MeCab user dictionary
+#'
+#' \code{dict_index} compiles a user dictionary CSV file into a binary
+#' dictionary that can be used with \code{pos} and \code{posParallel}.
+#'
+#' This function wraps MeCab's \code{mecab-dict-index} internally, so you
+#' do not need the command-line tool installed separately.
+#'
+#' @param dic_csv Character scalar. Path to the user dictionary CSV file(s).
+#'   Multiple CSV files can be provided as a character vector.
+#' @param out_dic Character scalar. Path for the output compiled dictionary file.
+#' @param dic_dir Character scalar. Path to the system dictionary directory.
+#'   This is required so that MeCab can reference the system dictionary
+#'   configuration during compilation.
+#' @param dic_charset Character scalar. Charset of the input CSV file.
+#'   Default is \code{"utf-8"}.
+#' @param out_charset Character scalar. Charset of the output dictionary.
+#'   Default is \code{"utf-8"}.
+#' @return Invisible \code{TRUE} on success.
+#'
+#' @examples
+#' \dontrun{
+#' dict_index(
+#'   dic_csv = "user_words.csv",
+#'   out_dic = "user.dic",
+#'   dic_dir = "/usr/local/lib/mecab/dic/ipadic"
+#' )
+#'
+#' # Then use the compiled dictionary:
+#' pos("some text", user_dic = "user.dic")
+#' }
+#'
+#' @export
+dict_index <- function(dic_csv, out_dic, dic_dir,
+                       dic_charset = "utf-8", out_charset = "utf-8") {
+  if (!is.character(dic_csv) || length(dic_csv) < 1)
+    stop("dic_csv must be a character vector of CSV file path(s)")
+  if (!is.character(out_dic) || length(out_dic) != 1)
+    stop("out_dic must be a single output file path")
+  if (!is.character(dic_dir) || length(dic_dir) != 1)
+    stop("dic_dir must be a single directory path")
+
+  for (f in dic_csv) {
+    if (!file.exists(f))
+      stop("CSV file not found: ", f)
+  }
+  if (!dir.exists(dic_dir))
+    stop("System dictionary directory not found: ", dic_dir)
+
+  out_dir <- dirname(out_dic)
+  if (!dir.exists(out_dir))
+    stop("Output directory does not exist: ", out_dir)
+
+  args <- c("mecab-dict-index",
+            "-d", normalizePath(dic_dir, mustWork = TRUE),
+            "-u", normalizePath(out_dic, mustWork = FALSE),
+            "-f", dic_charset,
+            "-t", out_charset,
+            normalizePath(dic_csv, mustWork = TRUE))
+
+  dictIndexRcpp(args)
+
+  invisible(TRUE)
+}

README.md

@@ -10,6 +10,17 @@ This package, RcppMeCab, is a `Rcpp` wrapper for the part-of-speech morphologica
 
 __Please see [this](README_kr.md) for easy installation and usage examples in Korean.__
 
+## MeCab backends
+
+RcppMeCab builds MeCab from source at install time. The MeCab variant is selected by the `MECAB_LANG` environment variable:
+
+| `MECAB_LANG` | Backend | Version | Source |
+|---|---|---|---|
+| `ko` (default) | [mecab-ko-msvc](https://github.com/Pusnow/mecab-ko-msvc) | 0.999 | Pusnow/mecab-ko-msvc |
+| `ja` | [MeCab](http://taku910.github.io/mecab/) | 0.996 | taku910/mecab |
+
+On Linux and macOS, if MeCab is already installed system-wide (detected via `mecab-config`), RcppMeCab uses the system installation regardless of `MECAB_LANG`.
+
 ## Installation
 
 ### Linux, macOS, and Windows
@@ -26,9 +37,9 @@ devtools::install_github("junhewk/RcppMeCab")
 
 If you already have MeCab installed (e.g. via `brew install mecab` on macOS, or `apt install libmecab-dev` on Linux), RcppMeCab will use your system installation.
 
-### Language selection (Windows)
+### Language selection
 
-On Windows, set `MECAB_LANG` before installation to choose the MeCab language variant. The default is `ko` (Korean).
+Set `MECAB_LANG` before installation to choose the MeCab variant:
 
 ```r
 # Korean (default)
@@ -44,7 +55,7 @@ install.packages("RcppMeCab", type = "source")
 You need a MeCab dictionary for your target language:
 
 + **Japanese**: Install [MeCab](http://taku910.github.io/mecab/) and IPAdic, or on macOS: `brew install mecab mecab-ipadic`
-+ **Korean**: Install [mecab-ko](https://bitbucket.org/eunjeon/mecab-ko) and [mecab-ko-dic](https://bitbucket.org/eunjeon/mecab-ko-dic). Mirrors of these files are also available on the [RcppMeCab releases page](https://github.com/junhewk/RcppMeCab/releases/tag/0.0.1.0) in case Bitbucket is unavailable. On Windows: install [mecab-ko-msvc](https://github.com/Pusnow/mecab-ko-msvc) and [mecab-ko-dic-msvc](https://github.com/Pusnow/mecab-ko-dic-msvc) in `C:\mecab`
++ **Korean**: Install [mecab-ko-dic](https://github.com/Pusnow/mecab-ko-msvc/releases) (available as `mecab-ko-dic.zip`/`mecab-ko-dic.tar.gz` from mecab-ko-msvc releases)
 + **Chinese**: Install MeCab with [MeCab Chinese Dic](http://www.52nlp.cn/%E7%94%A8mecab%E6%89%93%E9%80%A0%E4%B8%80%E5%A5%97%E5%AE%9E%E7%94%A8%E7%9A%84%E4%B8%AD%E6%96%87%E5%88%86%E8%AF%8D%E7%B3%BB%E7%BB%9F%E4%B8%89%EF%BC%9Amecab-chinese)
 
 ## Usage
@@ -65,32 +76,27 @@ posParallel(sentence)                # parallelized, faster for large inputs
 + `join`: if `TRUE` (default), output is `morpheme/tag`; if `FALSE`, output is `morpheme` with tag as attribute
 + `format`: `"list"` (default) or `"data.frame"`
 + `sys_dic`: directory containing `dicrc`, `model.bin`, etc. Set a default with `options(mecabSysDic = "/path/to/dic")`
-+ `user_dic`: path to a user dictionary compiled by `mecab-dict-index`
++ `user_dic`: path to a user dictionary compiled by `dict_index()`
 
 Note: provide full paths for `sys_dic` and `user_dic` (no tilde `~/` expansion).
 
 ## Compiling a user dictionary
 
-MeCab's `DictionaryCompiler` API calls `die()`, which would crash the R session, so it is not exposed through RcppMeCab. Use the `mecab-dict-index` command-line tool instead.
-
-You need a `model_file` for automatic cost estimation:
-
-+ Japanese: [model_file in ipadic](https://drive.google.com/uc?export=download&id=0B4y35FiV1wh7bnc5aFZSTE9qNnM)
-+ Korean: `model.bin` in [mecab-ko-dic](https://bitbucket.org/eunjeon/mecab-ko-dic)
+RcppMeCab provides the `dict_index()` function to compile user dictionaries directly from R, without needing the `mecab-dict-index` command-line tool.
 
 Prepare your entries as a CSV file ([Japanese format](http://taku910.github.io/mecab/dic.html), [Korean format](https://bitbucket.org/eunjeon/mecab-ko-dic/src/master/final/user-dic/README.md)), then compile:
 
-```sh
-/usr/local/libexec/mecab/mecab-dict-index \
-  -m /path/to/model.bin \
-  -d /path/to/mecab-dic \
-  -u userdic.dic \
-  -f utf8 -t utf8 \
-  entries.csv
+```r
+dict_index(
+  dic_csv = "entries.csv",
+  out_dic = "userdic.dic",
+  dic_dir = "/path/to/mecab-dic"
+)
+
+# Then use the compiled dictionary:
+pos("some text", user_dic = "userdic.dic")
 ```
 
-On Windows, use `mecab-dict-index.exe` bundled with [mecab-ko-msvc](https://github.com/Pusnow/mecab-ko-msvc) or the MeCab binary installer.
-
 ## Author
 
 Junhewk Kim (junhewk.kim@gmail.com)