mecab neologd 名詞,固有名詞,一般で第「数字」,￥「数字」,¥「数字」もスキップ

Cargo.toml

@@ -5,7 +5,7 @@ members = [
 resolver = "2"
 
 [workspace.package]
-version = "0.4.5"
+version = "0.4.6"
 authors = ["Masato TOYOSHIMA", "phoepsilonix <[email protected]>"]
 edition = "2021"
 rust-version = "1.82"

crates/dict-to-mozc/src/lib.rs

@@ -570,7 +570,8 @@ fn id_expr(clsexpr: &str, _id_def: &mut IdDef, class_map: &mut MyIndexMap<String
 
     //static KANA_CHECK: Lazy<Regex> = Lazy::new(|| Regex::new(r"^[\p{Hiragana}\p{Katakana}ーゝゞヽヾ゛゜・]+$").unwrap());
     static KANA_CHECK: Lazy<Regex> = Lazy::new(|| Regex::new(r"^[(ぁ-ゖ)ゐゑゐ゙ゑ゙(ァ-ヺ)ー・゛゜]+$").unwrap());
-    static START_SUUJI_CHECK: Lazy<Regex> = Lazy::new(|| Regex::new(r"^[0-9]+").unwrap());
+    //static START_SUUJI_CHECK: Lazy<Regex> = Lazy::new(|| Regex::new(r"^(?:\d|￥\d|¥\d|第\d)+").unwrap());
+    static START_SUUJI_CHECK: Lazy<Regex> = Lazy::new(|| Regex::new(r"^(\d|￥\d|¥\d|第\d)+").unwrap());
     //static EISUU_CHECK: Lazy<Regex> = Lazy::new(|| Regex::new(r"^[a-zA-Z0-9' ]+$").unwrap());
     static KIGOU_CHECK: Lazy<Regex> = Lazy::new(|| Regex::new(r"^[a-zA-Z' ]+$").unwrap());
     // 地名チェックに用いる日本語判定