Skip to content

IPA辞書を拡充してローマ字駅名の未解決を61件から1件に削減 #1626

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
TinyKitten merged 2 commits into from
Aug 12, 2026
Merged

IPA辞書を拡充してローマ字駅名の未解決を61件から1件に削減 #1626

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
3324b61
IPA辞書を拡充してローマ字駅名の未解決を61件から1件に削減
TinyKitten Aug 12, 2026
630ee7e
Merge branch 'dev' into feature/ipa-dictionary-expansion
TinyKitten Aug 12, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
197 changes: 191 additions & 6 deletions stationapi/src/domain/ipa.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -397,7 +397,30 @@ fn is_name_token_char(c: char) -> bool {
c.is_ascii_alphanumeric()
|| matches!(
c,
'\'' | '.' | 'Ā' | 'Ī' | 'Ū' | 'Ē' | 'Ō' | 'ā' | 'ī' | 'ū' | 'ē' | 'ō'
'\'' | '.'
| 'Ā'
| 'Ī'
| 'Ū'
| 'Ē'
| 'Ō'
| 'ā'
| 'ī'
| 'ū'
| 'ē'
| 'ō'
// 日本式ローマ字のサーカムフレックス長音 (Kyûjômae / Sessokyô / Okuôi)。
// トークン文字に含めないと長音記号が区切り文字扱いになり、
// "Kyûjômae" が "Ky" / "j" / "mae" に割れて読みが崩れる。
| 'Â'
| 'Î'
| 'Û'
| 'Ê'
| 'Ô'
| 'â'
| 'î'
| 'û'
| 'ê'
| 'ô'
)
}

Expand All @@ -411,13 +434,15 @@ fn normalize_name_token(token: &str) -> String {
.to_lowercase()
}

/// マクロン (ā/ī/ū/ē/ō) とサーカムフレックス (â/î/û/ê/ô) の長音表記を、
/// romaji_to_katakana が解釈できる母音の並びへ展開する。
fn normalize_name_char(c: char) -> Vec<char> {
match c {
'Ā' | 'ā' => vec!['a', 'a'],
'Ī' | 'ī' => vec!['i', 'i'],
'Ū' | 'ū' => vec!['u', 'u'],
'Ē' | 'ē' => vec!['e', 'i'],
'Ō' | 'ō' => vec!['o', 'u'],
'Ā' | 'ā' | 'Â' | 'â' => vec!['a', 'a'],
'Ī' | 'ī' | 'Î' | 'î' => vec!['i', 'i'],
'Ū' | 'ū' | 'Û' | 'û' => vec!['u', 'u'],
'Ē' | 'ē' | 'Ê' | 'ê' => vec!['e', 'i'],
'Ō' | 'ō' | 'Ô' | 'ô' => vec!['o', 'u'],
_ => vec![c],
}
}
Expand Down Expand Up @@ -685,6 +710,74 @@ fn lookup_english_word_ipa(word: &str) -> Option<&'static str> {
"t" => Some("tiː"),
"trans" => Some("tɹæns"),
"zoological" => Some("zuːəlɑdʒɪkəl"),
// ここから下は `make ipa-audit` で未解決として報告された語。
// 英語・カタカナ外来語由来のもの。
"bayside" => Some("beɪsaɪd"),
"disney" => Some("dɪzni"),
"disneyland" => Some("dɪznilænd"),
// "DisneySea" は camelCase 分割で "Disney" + "Sea" になる。"sea" を登録しないと
// romaji_to_katakana が「セア」と読んでしまい、監査には出ないまま誤読が残る。
"sea" => Some("siː"),
"girls'" => Some("gɝlz"),
"gran" => Some("gɹɑːn"),
"playpia" => Some("pleɪpiə"),
"portopia" => Some("pɔɹtoʊpiə"),
"pref.art" => Some("pɹɛf ɑɹt"),
"pref.sports" => Some("pɹɛf spɔɹts"),
"race" => Some("ɹeɪs"),
"resort" => Some("ɹɪzɔɹt"),
"river" => Some("ɹɪvɚ"),
"royce'" => Some("ɹɔɪs"),
"ruins" => Some("ɹuːɪnz"),
"screen" => Some("skɹiːn"),
"shop" => Some("ʃɑp"),
"socio" => Some("soʊʃioʊ"),
// 「ウェスパ椿山」の "WeSPa" は camelCase 分割で "We" + "SPa" に割れるため、
// 英単語の we (wiː) ではなくカタカナ読みに寄せた wɛ を割り当てる。
"we" => Some("wɛ"),
"spa" => Some("spɑː"),
"st" => Some("seɪnt"),
"tennis" => Some("tɛnɪs"),
"toyopet" => Some("toʊjoʊpɛt"),
"univ" => Some("juːnəvɚsəti"),
"vogel" => Some("foʊgəl"),
"yomiuriland" => Some("jomiɯɾilænd"),
"yrp" => Some("waɪ ɑɹ piː"),
// 訓令式・非ヘボン式のローマ字表記など、romaji_to_katakana が解釈できない日本語読み。
// IPA は katakana_to_ipa() の出力に合わせ、長音のみ既存エントリと同じ ː 表記に畳んでいる。
"3choume" => Some("sant͡ɕoːme"),
"4choume" => Some("jont͡ɕoːme"),
"daiunzi" => Some("daiɯɲdʑi"),
"etchuujima" => Some("ett͡ɕɯːdʑima"),
"golfjomae" => Some("gɑlf dʑoːmae"),
"hasirano" => Some("haɕiɾano"),
"hatcho" => Some("hatt͡ɕoː"),
"hatchome" => Some("hatt͡ɕoːme"),
"hatchonawate" => Some("hatt͡ɕoːnawate"),
"hatchoubori" => Some("hatt͡ɕoːboɾi"),
"hatcomuta" => Some("hatt͡ɕoːmɯta"),
"inadazutumi" => Some("inadadzɯt͡sɯmi"),
"inuboh" => Some("inɯboː"),
"itchumae" => Some("itt͡ɕɯːmae"),
"itihana" => Some("it͡ɕihana"),
"juitchome" => Some("dʑɯːitt͡ɕoːme"),
"kohjiro" => Some("koːdʑiɾo"),
"kozyou" => Some("kodʑoː"),
"mitumatu" => Some("mit͡sɯmat͡sɯ"),
"mitusawa" => Some("mit͡sɯsawa"),
"nijuhatchome" => Some("ɲidʑɯːhatt͡ɕoːme"),
"oh" => Some("oː"),
"ohda" => Some("oːda"),
"ohmisaki" => Some("oːmisakʲi"),
"ohshou" => Some("oːɕoː"),
"saidaizi" => Some("saidaidʑi"),
"saigoh" => Some("saigoː"),
"shicikenjaya" => Some("ɕit͡ɕikeɲdʑaja"),
"sirokanedai" => Some("ɕiɾokanedai"),
"sizuoka" => Some("ɕizɯoka"),
"suzi" => Some("sɯdʑi"),
"tamati" => Some("tamat͡ɕi"),
"tsudumigaura" => Some("t͡sɯdzɯmigaɯɾa"),
_ => None,
}
}
Expand Down Expand Up @@ -1552,6 +1645,98 @@ mod tests {
);
}

#[test]
fn test_station_name_ipa_supports_kunrei_style_romanization() {
// 訓令式 (si/ti/tu/zi) は romaji_to_katakana が解釈できないため辞書で補う。
assert_eq!(
station_name_to_ipa("シロカネダイ", Some("Sirokanedai")),
Some("ɕilokanedai".to_string())
);
assert_eq!(
station_name_to_ipa("タマチ", Some("Tamati")),
Some("tamat͡ɕi".to_string())
);
assert_eq!(
station_name_to_ipa("シンサイダイジチョウスジ", Some("Shin-saidaizi-cho-suzi")),
Some("ɕiɴ saidaidʑi t͡ɕo sɯdʑi".to_string())
);
}

#[test]
fn test_station_name_ipa_expands_circumflex_long_vowels() {
// サーカムフレックスをトークン文字として扱わないと "Ky" / "j" / "mae" に割れる。
assert_eq!(
station_name_to_ipa("キュウジョウマエ", Some("Kyûjômae")),
Some("kʲɯ.ɯdʑo.ɯma.e".to_string())
);
assert_eq!(
station_name_to_ipa("セッソキョウオンセン", Some("Sessokyô Onsen")),
Some("sessokʲo.ɯ onseɴ".to_string())
);
}

#[test]
fn test_station_name_ipa_supports_oh_style_long_vowels() {
// 「オウ」を oh と綴る表記 (犬吠 / 西郷 / 海の王迎)。
assert_eq!(
station_name_to_ipa("イヌボウ", Some("Inuboh")),
Some("inɯboː".to_string())
);
assert_eq!(
station_name_to_ipa("ウミノオウムカエ", Some("Umino-oh-mukae")),
Some("ɯmino oː mɯka.e".to_string())
);
}

#[test]
fn test_station_name_ipa_supports_chome_variants() {
assert_eq!(
station_name_to_ipa("ニシハッチョウメ", Some("Nishi hatchome")),
Some("ɲiɕi hatt͡ɕoːme".to_string())
);
assert_eq!(
station_name_to_ipa("テヅカヤマサンチョウメ", Some("Tezukayama-3chōme")),
Some("tezɯkajama sant͡ɕoːme".to_string())
);
}

#[test]
fn test_station_name_ipa_supports_facility_brand_words() {
assert_eq!(
station_name_to_ipa(
"トウキョウディズニーランド・ステーション",
Some("Tokyo Disneyland Station")
),
Some("tokʲo dɪznilænd steɪʃən".to_string())
);
// "DisneySea" は camelCase 分割で "Disney" + "Sea" になる。
assert_eq!(
station_name_to_ipa(
"トウキョウディズニーシー・ステーション",
Some("Tokyo DisneySea Station")
),
Some("tokʲo dɪzni siː steɪʃən".to_string())
);
}

#[test]
fn test_station_name_ipa_supports_abbreviated_english_words() {
// "St." / "Univ." は normalize_name_token が末尾のピリオドを落とす。
assert_eq!(
station_name_to_ipa("セイマリアビョウインマエ", Some("St. Mary's Hospital")),
Some("seɪnt mɛɹiz hɑspɪtəl".to_string())
);
assert_eq!(
station_name_to_ipa("アイダイイガクブミナミグチ", Some("Ehime Univ. Hospital")),
Some("eçime juːnəvɚsəti hɑspɪtəl".to_string())
);
// "Pref.Art" はピリオドが語中にあるため 1 トークンのまま引かれる。
assert_eq!(
station_name_to_ipa("ケンリツビジュツカンマエ", Some("Pref.Art Museum")),
Some("pɹɛf ɑɹt mjuːziəm".to_string())
);
}

#[test]
fn test_station_name_ipa_supports_meitetsu_prefixed_station_names_from_csv() {
let cases = [
Expand Down
Loading