chaffjs 0.12.0 → 0.13.0

package/dist/detectors/acronym-context.d.ts

@@ -1,13 +1,25 @@
 /**
  * 大文字の語が略語ではなく、決まった書き方の一部として読める所（3:30 PM、1pm ET、USD 1,000、Kansas City, MO 64108）。
  * どれも閉じた集合で、しかも隣の数字があるときだけ認める。PM（project manager）、CT、CA（certificate authority）は
- * 数字の隣でなければ略語のまま数える。
+ * 数字の隣でなければ略語のまま数える。語は言語パッケージの語彙表から読み、書き方の組み立てだけをここに置く。
  */
 type Span = {
     readonly start: number;
     readonly end: number;
 };
-/** 文の中で、略語として数えない大文字を含む範囲。略語は範囲にまるごと入るときだけ外れるので、語の境目は見なくてよい（NOTE の中の NOT）。 */
-export declare const notAcronymSpans: (text: string) => Span[];
+/** 言語パッケージの語彙表から読む、数字の隣でだけ略語から外す語と、単独で強調として書かれる語。字面どおりに照らす。 */
+export type NotationWords = {
+    readonly meridiem: readonly string[];
+    readonly timeZones: readonly string[];
+    readonly currencies: readonly string[];
+    readonly usStates: readonly string[];
+    readonly emphasis: readonly string[];
+};
+export type NotAcronymSpans = (text: string) => Span[];
+/**
+ * 文の中で、略語として数えない大文字を含む範囲。語彙表から一度だけ組み立てて、文ごとに呼ぶ。
+ * 略語は範囲にまるごと入るときだけ外れるので、語の境目は見なくてよい（NOTE の中の NOT）。
+ */
+export declare const notAcronymSpansOf: (words: NotationWords) => NotAcronymSpans;
 export {};
 //# sourceMappingURL=acronym-context.d.ts.map

package/dist/detectors/acronym-context.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"acronym-context.d.ts","sourceRoot":"","sources":["../../src/detectors/acronym-context.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAEH,KAAK,IAAI,GAAG;IAAE,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAA;CAAE,CAAC;AA4C7D,8EAA8E;AAC9E,eAAO,MAAM,eAAe,GAAI,MAAM,MAAM,KAAG,IAAI,EACsF,CAAC"}
+{"version":3,"file":"acronym-context.d.ts","sourceRoot":"","sources":["../../src/detectors/acronym-context.ts"],"names":[],"mappings":"AAEA;;;;GAIG;AAEH,KAAK,IAAI,GAAG;IAAE,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAA;CAAE,CAAC;AAE7D,+DAA+D;AAC/D,MAAM,MAAM,aAAa,GAAG;IAC1B,QAAQ,CAAC,QAAQ,EAAE,SAAS,MAAM,EAAE,CAAC;IACrC,QAAQ,CAAC,SAAS,EAAE,SAAS,MAAM,EAAE,CAAC;IACtC,QAAQ,CAAC,UAAU,EAAE,SAAS,MAAM,EAAE,CAAC;IACvC,QAAQ,CAAC,QAAQ,EAAE,SAAS,MAAM,EAAE,CAAC;IACrC,QAAQ,CAAC,QAAQ,EAAE,SAAS,MAAM,EAAE,CAAC;CACtC,CAAC;AAuBF,MAAM,MAAM,eAAe,GAAG,CAAC,IAAI,EAAE,MAAM,KAAK,IAAI,EAAE,CAAC;AAEvD;;;GAGG;AACH,eAAO,MAAM,iBAAiB,GAAI,OAAO,aAAa,KAAG,eAGxD,CAAC"}

package/dist/detectors/acronym-context.js

@@ -1,26 +1,6 @@
-/**
- * 大文字の語が略語ではなく、決まった書き方の一部として読める所（3:30 PM、1pm ET、USD 1,000、Kansas City, MO 64108）。
- * どれも閉じた集合で、しかも隣の数字があるときだけ認める。PM（project manager）、CT、CA（certificate authority）は
- * 数字の隣でなければ略語のまま数える。
- */
-const MERIDIEM = ["AM", "PM"];
-/** 業務文書で説明なしに書かれる時間帯。ここに無いもの（ACT など）は略語として数える。 */
-const TIME_ZONE = ["UTC", "GMT", "JST", "ET", "EST", "EDT", "CT", "CST", "CDT", "MT", "MST", "MDT", "PT", "PST", "PDT", "CET", "CEST"];
-/** ISO 4217 のうち、業務文書の読み手が説明なしで通じる主要な通貨。CAD は数の隣でも computer-aided design と読める（3 CAD files）ので入れない。 */
-const CURRENCY = ["USD", "EUR", "JPY", "GBP", "CNY", "CHF", "AUD", "KRW", "HKD", "SGD", "INR"];
-/** 米国の州・特別区・準州の郵便略号。住所の書き方（, CA 94720）の中でだけ見る。 */
-const US_STATE = [
-    ...["AL", "AK", "AZ", "AR", "CA", "CO", "CT", "DE", "FL", "GA", "HI", "ID", "IL", "IN", "IA", "KS", "KY"],
-    ...["LA", "ME", "MD", "MA", "MI", "MN", "MS", "MO", "MT", "NE", "NV", "NH", "NJ", "NM", "NY", "NC", "ND"],
-    ...["OH", "OK", "OR", "PA", "RI", "SC", "SD", "TN", "TX", "UT", "VT", "VA", "WA", "WV", "WI", "WY"],
-    ...["DC", "PR", "GU", "VI", "AS", "MP"],
-];
-/**
- * 略語として読む字面を持たない英語の機能語。単独の NOT や AND は強調で、展開するものが無い。
- * OR（Oregon、operating room）や IS（International Standard）のように略語でもある語は入れない。
- */
-const PLAIN_WORD = ["NOT", "AND"];
-const oneOf = (words) => `(?:${words.join("|")})`;
+import { escapeRegExp } from "../orthography.js";
+// 空の語彙表は「その書き方が無い」。空の選択肢 (?:) は至る所で空文字に当たるので、何にも当たらない形にする。
+const oneOf = (words) => (words.length === 0 ? "(?!)" : `(?:${words.map(escapeRegExp).join("|")})`);
 const MINUTES = String.raw `:[0-5]\d`;
 const CLOCK_12 = String.raw `(?<![\d:.])(?:1[0-2]|0?[1-9])(?:${MINUTES})?`;
 const CLOCK_24 = String.raw `(?<![\d:.])(?:[01]?\d|2[0-3])${MINUTES}`;
@@ -29,14 +9,20 @@ const ANY_MERIDIEM = String.raw `[AaPp]\.?[Mm]\.?`;
 /** 桁区切りは 3 桁ずつ揃っているときだけ金額と読む。「item 1, USD」の 1, は金額ではない。 */
 const AMOUNT = String.raw `(?:\d{1,3}(?:,\d{3})+|\d+)(?:\.\d+)?`;
 // 強調の記号は空白に置き換えてある（**3:30** PM）ので、部品の間の空白は数を問わない。
-const PATTERNS = [
-    new RegExp(String.raw `${CLOCK_12}\s*${oneOf(MERIDIEM)}`, "gu"),
-    new RegExp(String.raw `(?:${CLOCK_24}|${CLOCK_12}\s*${ANY_MERIDIEM})\s*${oneOf(TIME_ZONE)}`, "gu"),
-    new RegExp(String.raw `${oneOf(CURRENCY)}\s*[$€£¥]?${AMOUNT}`, "gu"),
-    new RegExp(String.raw `(?<![\w.,])${AMOUNT}\s*${oneOf(CURRENCY)}`, "gu"),
-    new RegExp(String.raw `,\s+${oneOf(US_STATE)}\s+\d{5}(?:-\d{4})?(?!\d)`, "gu"),
-    new RegExp(oneOf(PLAIN_WORD), "gu"),
+const patternsOf = (words) => [
+    new RegExp(String.raw `${CLOCK_12}\s*${oneOf(words.meridiem)}`, "gu"),
+    new RegExp(String.raw `(?:${CLOCK_24}|${CLOCK_12}\s*${ANY_MERIDIEM})\s*${oneOf(words.timeZones)}`, "gu"),
+    new RegExp(String.raw `${oneOf(words.currencies)}\s*[$€£¥]?${AMOUNT}`, "gu"),
+    new RegExp(String.raw `(?<![\w.,])${AMOUNT}\s*${oneOf(words.currencies)}`, "gu"),
+    new RegExp(String.raw `,\s+${oneOf(words.usStates)}\s+\d{5}(?:-\d{4})?(?!\d)`, "gu"),
+    new RegExp(oneOf(words.emphasis), "gu"),
 ];
-/** 文の中で、略語として数えない大文字を含む範囲。略語は範囲にまるごと入るときだけ外れるので、語の境目は見なくてよい（NOTE の中の NOT）。 */
-export const notAcronymSpans = (text) => PATTERNS.flatMap((pattern) => [...text.matchAll(pattern)].map((match) => ({ start: match.index, end: match.index + match[0].length })));
+/**
+ * 文の中で、略語として数えない大文字を含む範囲。語彙表から一度だけ組み立てて、文ごとに呼ぶ。
+ * 略語は範囲にまるごと入るときだけ外れるので、語の境目は見なくてよい（NOTE の中の NOT）。
+ */
+export const notAcronymSpansOf = (words) => {
+    const patterns = patternsOf(words);
+    return (text) => patterns.flatMap((pattern) => [...text.matchAll(pattern)].map((match) => ({ start: match.index, end: match.index + match[0].length })));
+};
 //# sourceMappingURL=acronym-context.js.map

package/dist/detectors/acronym-context.js.map

@@ -1 +1 @@
-{"version":3,"file":"acronym-context.js","sourceRoot":"","sources":["../../src/detectors/acronym-context.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AAIH,MAAM,QAAQ,GAAG,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC;AAE9B,kDAAkD;AAClD,MAAM,SAAS,GAAG,CAAC,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,CAAC,CAAC;AAEvI,oGAAoG;AACpG,MAAM,QAAQ,GAAG,CAAC,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,CAAC,CAAC;AAE/F,kDAAkD;AAClD,MAAM,QAAQ,GAAG;IACf,GAAG,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC;IACzG,GAAG,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC;IACzG,GAAG,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC;IACnG,GAAG,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC;CACxC,CAAC;AAEF;;;GAGG;AACH,MAAM,UAAU,GAAG,CAAC,KAAK,EAAE,KAAK,CAAC,CAAC;AAElC,MAAM,KAAK,GAAG,CAAC,KAAwB,EAAU,EAAE,CAAC,MAAM,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC;AAE7E,MAAM,OAAO,GAAG,MAAM,CAAC,GAAG,CAAA,UAAU,CAAC;AACrC,MAAM,QAAQ,GAAG,MAAM,CAAC,GAAG,CAAA,mCAAmC,OAAO,IAAI,CAAC;AAC1E,MAAM,QAAQ,GAAG,MAAM,CAAC,GAAG,CAAA,gCAAgC,OAAO,EAAE,CAAC;AACrE,6CAA6C;AAC7C,MAAM,YAAY,GAAG,MAAM,CAAC,GAAG,CAAA,kBAAkB,CAAC;AAClD,2DAA2D;AAC3D,MAAM,MAAM,GAAG,MAAM,CAAC,GAAG,CAAA,sCAAsC,CAAC;AAEhE,kDAAkD;AAClD,MAAM,QAAQ,GAAsB;IAClC,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,GAAG,QAAQ,MAAM,KAAK,CAAC,QAAQ,CAAC,EAAE,EAAE,IAAI,CAAC;IAC9D,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,MAAM,QAAQ,IAAI,QAAQ,MAAM,YAAY,OAAO,KAAK,CAAC,SAAS,CAAC,EAAE,EAAE,IAAI,CAAC;IACjG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,GAAG,KAAK,CAAC,QAAQ,CAAC,aAAa,MAAM,EAAE,EAAE,IAAI,CAAC;IACnE,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,cAAc,MAAM,MAAM,KAAK,CAAC,QAAQ,CAAC,EAAE,EAAE,IAAI,CAAC;IACvE,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,OAAO,KAAK,CAAC,QAAQ,CAAC,2BAA2B,EAAE,IAAI,CAAC;IAC7E,IAAI,MAAM,CAAC,KAAK,CAAC,UAAU,CAAC,EAAE,IAAI,CAAC;CACpC,CAAC;AAEF,8EAA8E;AAC9E,MAAM,CAAC,MAAM,eAAe,GAAG,CAAC,IAAY,EAAU,EAAE,CACtD,QAAQ,CAAC,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,GAAG,EAAE,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC,CAAC"}
+{"version":3,"file":"acronym-context.js","sourceRoot":"","sources":["../../src/detectors/acronym-context.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,MAAM,mBAAmB,CAAC;AAmBjD,2DAA2D;AAC3D,MAAM,KAAK,GAAG,CAAC,KAAwB,EAAU,EAAE,CAAC,CAAC,KAAK,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,KAAK,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;AAE/H,MAAM,OAAO,GAAG,MAAM,CAAC,GAAG,CAAA,UAAU,CAAC;AACrC,MAAM,QAAQ,GAAG,MAAM,CAAC,GAAG,CAAA,mCAAmC,OAAO,IAAI,CAAC;AAC1E,MAAM,QAAQ,GAAG,MAAM,CAAC,GAAG,CAAA,gCAAgC,OAAO,EAAE,CAAC;AACrE,6CAA6C;AAC7C,MAAM,YAAY,GAAG,MAAM,CAAC,GAAG,CAAA,kBAAkB,CAAC;AAClD,2DAA2D;AAC3D,MAAM,MAAM,GAAG,MAAM,CAAC,GAAG,CAAA,sCAAsC,CAAC;AAEhE,kDAAkD;AAClD,MAAM,UAAU,GAAG,CAAC,KAAoB,EAAqB,EAAE,CAAC;IAC9D,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,GAAG,QAAQ,MAAM,KAAK,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,EAAE,IAAI,CAAC;IACpE,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,MAAM,QAAQ,IAAI,QAAQ,MAAM,YAAY,OAAO,KAAK,CAAC,KAAK,CAAC,SAAS,CAAC,EAAE,EAAE,IAAI,CAAC;IACvG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,GAAG,KAAK,CAAC,KAAK,CAAC,UAAU,CAAC,aAAa,MAAM,EAAE,EAAE,IAAI,CAAC;IAC3E,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,cAAc,MAAM,MAAM,KAAK,CAAC,KAAK,CAAC,UAAU,CAAC,EAAE,EAAE,IAAI,CAAC;IAC/E,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,OAAO,KAAK,CAAC,KAAK,CAAC,QAAQ,CAAC,2BAA2B,EAAE,IAAI,CAAC;IACnF,IAAI,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,QAAQ,CAAC,EAAE,IAAI,CAAC;CACxC,CAAC;AAIF;;;GAGG;AACH,MAAM,CAAC,MAAM,iBAAiB,GAAG,CAAC,KAAoB,EAAmB,EAAE;IACzE,MAAM,QAAQ,GAAG,UAAU,CAAC,KAAK,CAAC,CAAC;IACnC,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,GAAG,EAAE,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC,CAAC;AAC3J,CAAC,CAAC"}

package/dist/detectors/acronym-expansion.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"acronym-expansion.d.ts","sourceRoot":"","sources":["../../src/detectors/acronym-expansion.ts"],"names":[],"mappings":"AA4CA,iEAAiE;AACjE,MAAM,MAAM,eAAe,GAAG;IAAE,QAAQ,CAAC,OAAO,EAAE,SAAS,MAAM,EAAE,CAAC;IAAC,QAAQ,CAAC,KAAK,EAAE,SAAS,MAAM,EAAE,CAAA;CAAE,CAAC;AA6BzG,MAAM,MAAM,UAAU,GAAG,CAAC,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,EAAE,EAAE,MAAM,KAAK,OAAO,CAAC;AAEhF,+DAA+D;AAC/D,eAAO,MAAM,WAAW,GAAI,OAAO,eAAe,KAAG,UAGpD,CAAC"}
+{"version":3,"file":"acronym-expansion.d.ts","sourceRoot":"","sources":["../../src/detectors/acronym-expansion.ts"],"names":[],"mappings":"AAuEA,iEAAiE;AACjE,MAAM,MAAM,eAAe,GAAG;IAAE,QAAQ,CAAC,OAAO,EAAE,SAAS,MAAM,EAAE,CAAC;IAAC,QAAQ,CAAC,KAAK,EAAE,SAAS,MAAM,EAAE,CAAA;CAAE,CAAC;AA8BzG,MAAM,MAAM,UAAU,GAAG,CAAC,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,EAAE,EAAE,MAAM,KAAK,OAAO,CAAC;AAEhF,+DAA+D;AAC/D,eAAO,MAAM,WAAW,GAAI,OAAO,eAAe,KAAG,UAGpD,CAAC"}

package/dist/detectors/acronym-expansion.js

@@ -3,10 +3,11 @@ import { escapeRegExp } from "../orthography.js";
  * 展開は略語の**すぐ隣**にあるときだけ認める。
  * 60 文字も見ると、同じ文のどこかに括弧があるだけで「説明済み」になり、1 件も出なくなる。
  *
- * 認めるのは 4 つの形。どれも実際によく書かれる。
+ * 認めるのは次の形。どれも実際によく書かれる。
  *   CI（継続的インテグレーション）   略語のあとに括弧
  *   Continuous Integration (CI)      括弧の中が略語
  *   Tax Cuts and Jobs Act [TCJA]     角括弧の中が略語で、直前の語の頭文字と揃う
+ *   Relief Act (RA; P.L. 112-240)    括弧の最初の項目が略語で、直前の語か区切りの後の語の頭文字と揃う
  *   人事部（以下「HR」という。）      括弧の中が定義の語と略語だけ
  * 括弧と略語の間には、引用符（(“MNDA”)、（「MNDA」））と空白だけを許す。
  */
@@ -16,6 +17,10 @@ const CLOSES = new RegExp(String.raw `^${WRAP}[)）]`, "u");
 const OPENED = new RegExp(String.raw `[(（]${WRAP}$`, "u");
 const SQUARE_CLOSES = new RegExp(String.raw `^${WRAP}\]`, "u");
 const SQUARE_OPENED = new RegExp(String.raw `\[${WRAP}$`, "u");
+const ANY_OPENED = new RegExp(String.raw `[(（[]${WRAP}$`, "u");
+/** 括弧の最初の項目のあとの区切りと、括弧が閉じるまでの残り（(ARRA; P.L. 111-5)、（OP、Originator Profile））。 */
+const SEPARATED = new RegExp(String.raw `^${WRAP}[;；,，、](?<rest>[^()（）]*)`, "u");
+const QUOTES = /["“”'‘’「」『』]/gu;
 /** 括弧と略語の間の幅。空白は 1 つに畳んであるので、(“ MNDA ”) まで収まる。 */
 const NEAR = 3;
 /** 定義の語を挟む形で、括弧から略語まで・略語から括弧までに見る幅。(hereinafter referred to as “SLA”) が収まる。 */
@@ -26,15 +31,35 @@ const MARKER_GAP = String.raw `(?![A-Za-z])[\s、,]*`;
  * 角括弧は引用の印にも使う（[IANA]、[1]）。直前の語のうち大文字で始まる語の頭文字が
  * 略語と揃うときだけ展開と見なす。見る語は略語の文字数の 2 倍まで（and や of を挟むため）。
  */
+const lettersOf = (acronym) => acronym.replaceAll("&", "");
+const initialsOf = (words) => words
+    .filter((word) => /^[A-Z]/u.test(word))
+    .map((word) => word.charAt(0))
+    .join("");
 const spellsOut = (before, acronym) => {
-    const letters = acronym.replaceAll("&", "");
+    const letters = lettersOf(acronym);
     const words = before
-        .replace(SQUARE_OPENED, "")
+        .replace(ANY_OPENED, "")
         .trim()
         .split(/\s+/u)
         .slice(-letters.length * 2);
-    const initials = words.filter((word) => /^[A-Z]/u.test(word)).map((word) => word.charAt(0));
-    return initials.join("").endsWith(letters);
+    return initialsOf(words).endsWith(letters);
+};
+/**
+ * 括弧の最初の項目が略語で、区切りのあとに注記が続く形（(ARRA; P.L. 111-5)、（OP、Originator Profile））。
+ * 同じ形で列挙も書く（(MR, handbook, etc.)、(EPA, FDIC, GSA)）ので、直前の語か、区切りから括弧が閉じるまでの語の
+ * 頭文字が略語と揃うときだけ展開と見なす。区切りの後ろは閉じた括弧の中だけを見て、頭文字がちょうど略語になることを求める。
+ */
+const isSeparatedAt = (body, acronym, at) => {
+    if (!OPENED.test(body.slice(Math.max(0, at - NEAR), at)))
+        return false;
+    const separated = SEPARATED.exec(body.slice(at + acronym.length, at + acronym.length + DEFINITION_REACH));
+    if (separated === null)
+        return false;
+    const rest = separated.groups?.["rest"] ?? "";
+    const end = at + acronym.length + separated[0].length;
+    const closed = CLOSES.test(body.slice(end, end + NEAR));
+    return spellsOut(body.slice(0, at), acronym) || (closed && initialsOf(rest.replace(QUOTES, " ").trim().split(/\s+/u)) === lettersOf(acronym));
 };
 const oneOf = (words) => (words.length === 0 ? "(?!)" : words.map(escapeRegExp).join("|"));
 const definitionPatterns = (words) => ({
@@ -57,7 +82,9 @@ const isBracketedAt = (body, acronym, at) => {
     const before = body.slice(Math.max(0, at - NEAR), at);
     if (OPENS.test(after) || (OPENED.test(before) && CLOSES.test(after)))
         return true;
-    return SQUARE_OPENED.test(before) && SQUARE_CLOSES.test(after) && spellsOut(body.slice(0, at), acronym);
+    if (SQUARE_OPENED.test(before) && SQUARE_CLOSES.test(after) && spellsOut(body.slice(0, at), acronym))
+        return true;
+    return isSeparatedAt(body, acronym, at);
 };
 /** body の at にある略語が、その場で展開・定義されているか。語彙表から一度だけ組み立てて、略語ごとに呼ぶ。 */
 export const expansionAt = (words) => {

package/dist/detectors/acronym-expansion.js.map

@@ -1 +1 @@
-{"version":3,"file":"acronym-expansion.js","sourceRoot":"","sources":["../../src/detectors/acronym-expansion.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,MAAM,mBAAmB,CAAC;AAEjD;;;;;;;;;;GAUG;AACH,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAA,iBAAiB,CAAC;AACzC,MAAM,KAAK,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,IAAI,IAAI,MAAM,EAAE,GAAG,CAAC,CAAC;AACxD,MAAM,MAAM,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,IAAI,IAAI,MAAM,EAAE,GAAG,CAAC,CAAC;AACzD,MAAM,MAAM,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,OAAO,IAAI,GAAG,EAAE,GAAG,CAAC,CAAC;AACzD,MAAM,aAAa,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,IAAI,IAAI,IAAI,EAAE,GAAG,CAAC,CAAC;AAC9D,MAAM,aAAa,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,KAAK,IAAI,GAAG,EAAE,GAAG,CAAC,CAAC;AAE9D,kDAAkD;AAClD,MAAM,IAAI,GAAG,CAAC,CAAC;AAEf,+EAA+E;AAC/E,MAAM,gBAAgB,GAAG,EAAE,CAAC;AAE5B,4DAA4D;AAC5D,MAAM,UAAU,GAAG,MAAM,CAAC,GAAG,CAAA,qBAAqB,CAAC;AAEnD;;;GAGG;AACH,MAAM,SAAS,GAAG,CAAC,MAAc,EAAE,OAAe,EAAW,EAAE;IAC7D,MAAM,OAAO,GAAG,OAAO,CAAC,UAAU,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;IAC5C,MAAM,KAAK,GAAG,MAAM;SACjB,OAAO,CAAC,aAAa,EAAE,EAAE,CAAC;SAC1B,IAAI,EAAE;SACN,KAAK,CAAC,MAAM,CAAC;SACb,KAAK,CAAC,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC9B,MAAM,QAAQ,GAAG,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC;IAC5F,OAAO,QAAQ,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC;AAC7C,CAAC,CAAC;AAOF,MAAM,KAAK,GAAG,CAAC,KAAwB,EAAU,EAAE,CAAC,CAAC,KAAK,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;AAEtH,MAAM,kBAAkB,GAAG,CAAC,KAAsB,EAAsB,EAAE,CAAC,CAAC;IAC1E,MAAM,EAAE,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,2BAA2B,KAAK,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,UAAU,MAAM,IAAI,gBAAgB,EAAE,IAAI,CAAC;IAC3H,MAAM,EAAE,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,IAAI,IAAI,cAAc,KAAK,CAAC,KAAK,CAAC,KAAK,CAAC,WAAW,IAAI,MAAM,EAAE,IAAI,CAAC;CAClG,CAAC,CAAC;AAEH;;;GAGG;AACH,MAAM,WAAW,GAAG,CAAC,QAA4B,EAAE,IAAY,EAAE,OAAe,EAAE,EAAU,EAAW,EAAE;IACvG,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,GAAG,gBAAgB,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC;IACxF,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,GAAG,OAAO,CAAC,MAAM,EAAE,EAAE,GAAG,OAAO,CAAC,MAAM,GAAG,gBAAgB,CAAC,CAAC,CAAC;IAC7G,IAAI,MAAM,KAAK,IAAI,IAAI,MAAM,KAAK,IAAI;QAAE,OAAO,KAAK,CAAC;IACrD,OAAO,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC,KAAK,EAAE,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,KAAK,EAAE,CAAC;AAC7F,CAAC,CAAC;AAEF,MAAM,aAAa,GAAG,CAAC,IAAY,EAAE,OAAe,EAAE,EAAU,EAAW,EAAE;IAC3E,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,EAAE,GAAG,OAAO,CAAC,MAAM,EAAE,EAAE,GAAG,OAAO,CAAC,MAAM,GAAG,IAAI,CAAC,CAAC;IAC1E,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,GAAG,IAAI,CAAC,EAAE,EAAE,CAAC,CAAC;IACtD,IAAI,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QAAE,OAAO,IAAI,CAAC;IAClF,OAAO,aAAa,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,OAAO,CAAC,CAAC;AAC1G,CAAC,CAAC;AAIF,+DAA+D;AAC/D,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,KAAsB,EAAc,EAAE;IAChE,MAAM,QAAQ,GAAG,kBAAkB,CAAC,KAAK,CAAC,CAAC;IAC3C,OAAO,CAAC,IAAI,EAAE,OAAO,EAAE,EAAE,EAAE,EAAE,CAAC,aAAa,CAAC,IAAI,EAAE,OAAO,EAAE,EAAE,CAAC,IAAI,WAAW,CAAC,QAAQ,EAAE,IAAI,EAAE,OAAO,EAAE,EAAE,CAAC,CAAC;AAC7G,CAAC,CAAC"}
+{"version":3,"file":"acronym-expansion.js","sourceRoot":"","sources":["../../src/detectors/acronym-expansion.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,MAAM,mBAAmB,CAAC;AAEjD;;;;;;;;;;;GAWG;AACH,MAAM,IAAI,GAAG,MAAM,CAAC,GAAG,CAAA,iBAAiB,CAAC;AACzC,MAAM,KAAK,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,IAAI,IAAI,MAAM,EAAE,GAAG,CAAC,CAAC;AACxD,MAAM,MAAM,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,IAAI,IAAI,MAAM,EAAE,GAAG,CAAC,CAAC;AACzD,MAAM,MAAM,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,OAAO,IAAI,GAAG,EAAE,GAAG,CAAC,CAAC;AACzD,MAAM,aAAa,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,IAAI,IAAI,IAAI,EAAE,GAAG,CAAC,CAAC;AAC9D,MAAM,aAAa,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,KAAK,IAAI,GAAG,EAAE,GAAG,CAAC,CAAC;AAC9D,MAAM,UAAU,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,QAAQ,IAAI,GAAG,EAAE,GAAG,CAAC,CAAC;AAC9D,gFAAgF;AAChF,MAAM,SAAS,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,IAAI,IAAI,0BAA0B,EAAE,GAAG,CAAC,CAAC;AAChF,MAAM,MAAM,GAAG,gBAAgB,CAAC;AAEhC,kDAAkD;AAClD,MAAM,IAAI,GAAG,CAAC,CAAC;AAEf,+EAA+E;AAC/E,MAAM,gBAAgB,GAAG,EAAE,CAAC;AAE5B,4DAA4D;AAC5D,MAAM,UAAU,GAAG,MAAM,CAAC,GAAG,CAAA,qBAAqB,CAAC;AAEnD;;;GAGG;AACH,MAAM,SAAS,GAAG,CAAC,OAAe,EAAU,EAAE,CAAC,OAAO,CAAC,UAAU,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;AAE3E,MAAM,UAAU,GAAG,CAAC,KAAwB,EAAU,EAAE,CACtD,KAAK;KACF,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;KACtC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;KAC7B,IAAI,CAAC,EAAE,CAAC,CAAC;AAEd,MAAM,SAAS,GAAG,CAAC,MAAc,EAAE,OAAe,EAAW,EAAE;IAC7D,MAAM,OAAO,GAAG,SAAS,CAAC,OAAO,CAAC,CAAC;IACnC,MAAM,KAAK,GAAG,MAAM;SACjB,OAAO,CAAC,UAAU,EAAE,EAAE,CAAC;SACvB,IAAI,EAAE;SACN,KAAK,CAAC,MAAM,CAAC;SACb,KAAK,CAAC,CAAC,OAAO,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC9B,OAAO,UAAU,CAAC,KAAK,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC;AAC7C,CAAC,CAAC;AAEF;;;;GAIG;AACH,MAAM,aAAa,GAAG,CAAC,IAAY,EAAE,OAAe,EAAE,EAAU,EAAW,EAAE;IAC3E,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,GAAG,IAAI,CAAC,EAAE,EAAE,CAAC,CAAC;QAAE,OAAO,KAAK,CAAC;IACvE,MAAM,SAAS,GAAG,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,GAAG,OAAO,CAAC,MAAM,EAAE,EAAE,GAAG,OAAO,CAAC,MAAM,GAAG,gBAAgB,CAAC,CAAC,CAAC;IAC1G,IAAI,SAAS,KAAK,IAAI;QAAE,OAAO,KAAK,CAAC;IACrC,MAAM,IAAI,GAAG,SAAS,CAAC,MAAM,EAAE,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC;IAC9C,MAAM,GAAG,GAAG,EAAE,GAAG,OAAO,CAAC,MAAM,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IACtD,MAAM,MAAM,GAAG,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,EAAE,GAAG,GAAG,IAAI,CAAC,CAAC,CAAC;IACxD,OAAO,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,OAAO,CAAC,IAAI,CAAC,MAAM,IAAI,UAAU,CAAC,IAAI,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,KAAK,SAAS,CAAC,OAAO,CAAC,CAAC,CAAC;AAChJ,CAAC,CAAC;AAOF,MAAM,KAAK,GAAG,CAAC,KAAwB,EAAU,EAAE,CAAC,CAAC,KAAK,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC;AAEtH,MAAM,kBAAkB,GAAG,CAAC,KAAsB,EAAsB,EAAE,CAAC,CAAC;IAC1E,MAAM,EAAE,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,2BAA2B,KAAK,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,UAAU,MAAM,IAAI,gBAAgB,EAAE,IAAI,CAAC;IAC3H,MAAM,EAAE,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,IAAI,IAAI,cAAc,KAAK,CAAC,KAAK,CAAC,KAAK,CAAC,WAAW,IAAI,MAAM,EAAE,IAAI,CAAC;CAClG,CAAC,CAAC;AAEH;;;GAGG;AACH,MAAM,WAAW,GAAG,CAAC,QAA4B,EAAE,IAAY,EAAE,OAAe,EAAE,EAAU,EAAW,EAAE;IACvG,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,GAAG,gBAAgB,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC;IACxF,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,GAAG,OAAO,CAAC,MAAM,EAAE,EAAE,GAAG,OAAO,CAAC,MAAM,GAAG,gBAAgB,CAAC,CAAC,CAAC;IAC7G,IAAI,MAAM,KAAK,IAAI,IAAI,MAAM,KAAK,IAAI;QAAE,OAAO,KAAK,CAAC;IACrD,OAAO,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC,KAAK,EAAE,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,KAAK,EAAE,CAAC;AAC7F,CAAC,CAAC;AAEF,MAAM,aAAa,GAAG,CAAC,IAAY,EAAE,OAAe,EAAE,EAAU,EAAW,EAAE;IAC3E,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,EAAE,GAAG,OAAO,CAAC,MAAM,EAAE,EAAE,GAAG,OAAO,CAAC,MAAM,GAAG,IAAI,CAAC,CAAC;IAC1E,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,GAAG,IAAI,CAAC,EAAE,EAAE,CAAC,CAAC;IACtD,IAAI,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QAAE,OAAO,IAAI,CAAC;IAClF,IAAI,aAAa,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,SAAS,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,OAAO,CAAC;QAAE,OAAO,IAAI,CAAC;IAClH,OAAO,aAAa,CAAC,IAAI,EAAE,OAAO,EAAE,EAAE,CAAC,CAAC;AAC1C,CAAC,CAAC;AAIF,+DAA+D;AAC/D,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,KAAsB,EAAc,EAAE;IAChE,MAAM,QAAQ,GAAG,kBAAkB,CAAC,KAAK,CAAC,CAAC;IAC3C,OAAO,CAAC,IAAI,EAAE,OAAO,EAAE,EAAE,EAAE,EAAE,CAAC,aAAa,CAAC,IAAI,EAAE,OAAO,EAAE,EAAE,CAAC,IAAI,WAAW,CAAC,QAAQ,EAAE,IAAI,EAAE,OAAO,EAAE,EAAE,CAAC,CAAC;AAC7G,CAAC,CAAC"}

package/dist/detectors/char-shape.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"char-shape.d.ts","sourceRoot":"","sources":["../../src/detectors/char-shape.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAA6C,MAAM,cAAc,CAAC;AAgDxF;;;GAGG;AACH,eAAO,MAAM,QAAQ,EAAE,QAWhB,CAAC;AAIR;;;GAGG;AACH,eAAO,MAAM,SAAS,EAAE,QAWjB,CAAC"}
+{"version":3,"file":"char-shape.d.ts","sourceRoot":"","sources":["../../src/detectors/char-shape.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAAsC,MAAM,cAAc,CAAC;AA4BjF;;;GAGG;AACH,eAAO,MAAM,QAAQ,EAAE,QAatB,CAAC;AAIF;;;GAGG;AACH,eAAO,MAAM,SAAS,EAAE,QAWjB,CAAC"}

package/dist/detectors/char-shape.js

@@ -1,58 +1,44 @@
 import { compacted, placeOf } from "./gram-place.js";
 import { maskAddresses } from "../address-chain.js";
+import { isAddressRun } from "./place-run.js";
 import { proseText } from "../measure.js";
 /**
  * 文字の並びを見る検出。漢字の連なりが住所かどうかだけは、形態素解析の地名と数で決める。
  * 覆った箇所は空白になっているので、proseText で潰してから数える。
  */
 const KANJI_RUN = /[一-鿿]+/gu;
-/** 地名（Geo）と、地名に付く単位（GeoUnit: 都・県・市・区・町）。 */
-const isGeoName = (token) => token?.features?.["NameType"] === "Geo";
-const isGeo = (token) => isGeoName(token) || token?.features?.["NameType"] === "GeoUnit";
-const isNumber = (token) => token?.features?.["NumType"] === "Card";
-/**
- * 住所の語: 地名と地名の単位（NameType=Geo / GeoUnit）、数（NumType=Card）、数のすぐ後ろの助数詞（丁目、NounType=Class）。
- * 辞書に無い町名（新橋）は、地名の後ろで数の前に来るときだけ住所の一部とする。
- */
-const isPlacePart = (tokens, index) => isGeo(tokens[index]) ||
-    isNumber(tokens[index]) ||
-    (isNumber(tokens[index - 1]) && tokens[index]?.features?.["NounType"] === "Class") ||
-    (isGeo(tokens[index - 1]) && isNumber(tokens[index + 1]));
-/**
- * 漢字の連なりが住所か（東京都港区新橋二丁目）。形態素解析が地名と読んだ語で始まり、住所の語だけでできている。
- * 住所は決まった形で、ひらがなを挟んで書けない。組織名（日本経済団体連合会）は地名の後ろに普通の語が続くので住所ではない。品詞が無ければ判定しない。
- */
-const isPlaceName = (sentence, run) => {
+/** 漢字の連なりが住所か（東京都港区新橋二丁目）。形態素解析の地名と数で決める。品詞が無ければ判定しない。 */
+const isPlaceName = (sentence, run, topUnits) => {
     const tokens = sentence.tokens;
     const place = tokens === undefined ? undefined : placeOf(compacted(sentence.text, "word"), run);
     if (tokens === undefined || place === undefined)
         return false;
     const [start, end] = [sentence.span.start + place.start, sentence.span.start + place.end];
     const covering = tokens.flatMap((token, index) => (token.span.start < end && start < token.span.end ? [index] : []));
-    const first = covering[0];
-    // 地名が単位を挟まずに続けば（東京大阪名古屋福岡）、住所ではなく地名の並び。
-    const listed = covering.some((index) => isGeoName(tokens[index]) && isGeoName(tokens[index + 1]));
-    return first !== undefined && isGeo(tokens[first]) && !listed && covering.every((index) => isPlacePart(tokens, index));
+    return isAddressRun(tokens, covering, topUnits);
 };
-const longestKanji = (sentence, profile) => [...maskAddresses(proseText(sentence), profile).matchAll(KANJI_RUN)]
+const longestKanji = (sentence, profile, topUnits) => [...maskAddresses(proseText(sentence), profile).matchAll(KANJI_RUN)]
     .map((match) => match[0])
-    .filter((run) => !isPlaceName(sentence, run))
+    .filter((run) => !isPlaceName(sentence, run, topUnits))
     .reduce((longest, run) => (run.length > longest.length ? run : longest), "");
 /**
  * 漢字が続くと、どこで語が切れるのか読み手が探すことになる。
  * 「情報処理推進機構認定試験」は 12 字。ひらがなを 1 つ挟むだけで読める。
  */
-export const kanjiRun = (doc, options) => doc.sentences
-    .map((sentence) => ({ sentence, run: longestKanji(sentence, doc.profile) }))
-    .filter(({ run }) => run.length > options.limit)
-    .map(({ sentence, run }) => ({
-    rule: "max-kanji-continuous",
-    severity: "warning",
-    line: 0,
-    column: 0,
-    quote: sentence.text.trim(),
-    values: { word: run, count: run.length, limit: options.limit, offset: sentence.span.start },
-}));
+export const kanjiRun = (doc, options) => {
+    const topUnits = new Set((doc.lexicons["prefecture-unit"] ?? []).map((entry) => entry.pattern));
+    return doc.sentences
+        .map((sentence) => ({ sentence, run: longestKanji(sentence, doc.profile, topUnits) }))
+        .filter(({ run }) => run.length > options.limit)
+        .map(({ sentence, run }) => ({
+        rule: "max-kanji-continuous",
+        severity: "warning",
+        line: 0,
+        column: 0,
+        quote: sentence.text.trim(),
+        values: { word: run, count: run.length, limit: options.limit, offset: sentence.span.start },
+    }));
+};
 const MIDDLE_DOT = /・/gu;
 /**
  * 中黒で並べると、どこまでが 1 つの項目か分からなくなる。

package/dist/detectors/char-shape.js.map

@@ -1 +1 @@
-{"version":3,"file":"char-shape.js","sourceRoot":"","sources":["../../src/detectors/char-shape.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,SAAS,EAAE,OAAO,EAAE,MAAM,iBAAiB,CAAC;AACrD,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACpD,OAAO,EAAE,SAAS,EAAE,MAAM,eAAe,CAAC;AAE1C;;;GAGG;AACH,MAAM,SAAS,GAAG,UAAU,CAAC;AAE7B,4CAA4C;AAC5C,MAAM,SAAS,GAAG,CAAC,KAAwB,EAAW,EAAE,CAAC,KAAK,EAAE,QAAQ,EAAE,CAAC,UAAU,CAAC,KAAK,KAAK,CAAC;AACjG,MAAM,KAAK,GAAG,CAAC,KAAwB,EAAW,EAAE,CAAC,SAAS,CAAC,KAAK,CAAC,IAAI,KAAK,EAAE,QAAQ,EAAE,CAAC,UAAU,CAAC,KAAK,SAAS,CAAC;AACrH,MAAM,QAAQ,GAAG,CAAC,KAAwB,EAAW,EAAE,CAAC,KAAK,EAAE,QAAQ,EAAE,CAAC,SAAS,CAAC,KAAK,MAAM,CAAC;AAEhG;;;GAGG;AACH,MAAM,WAAW,GAAG,CAAC,MAAwB,EAAE,KAAa,EAAW,EAAE,CACvE,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;IACpB,QAAQ,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;IACvB,CAAC,QAAQ,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,MAAM,CAAC,KAAK,CAAC,EAAE,QAAQ,EAAE,CAAC,UAAU,CAAC,KAAK,OAAO,CAAC;IAClF,CAAC,KAAK,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,QAAQ,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;AAE5D;;;GAGG;AACH,MAAM,WAAW,GAAG,CAAC,QAAkB,EAAE,GAAW,EAAW,EAAE;IAC/D,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,CAAC;IAC/B,MAAM,KAAK,GAAG,MAAM,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,SAAS,CAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC,EAAE,GAAG,CAAC,CAAC;IAChG,IAAI,MAAM,KAAK,SAAS,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,KAAK,CAAC;IAC9D,MAAM,CAAC,KAAK,EAAE,GAAG,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC;IAC1F,MAAM,QAAQ,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,GAAG,IAAI,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;IACrH,MAAM,KAAK,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;IAC1B,wCAAwC;IACxC,MAAM,MAAM,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,SAAS,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,IAAI,SAAS,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;IAClG,OAAO,KAAK,KAAK,SAAS,IAAI,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,MAAM,IAAI,QAAQ,CAAC,KAAK,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,WAAW,CAAC,MAAM,EAAE,KAAK,CAAC,CAAC,CAAC;AACzH,CAAC,CAAC;AAEF,MAAM,YAAY,GAAG,CAAC,QAAkB,EAAE,OAAoC,EAAU,EAAE,CACxF,CAAC,GAAG,aAAa,CAAC,SAAS,CAAC,QAAQ,CAAC,EAAE,OAAO,CAAC,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC;KACjE,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;KACxB,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,WAAW,CAAC,QAAQ,EAAE,GAAG,CAAC,CAAC;KAC5C,MAAM,CAAC,CAAC,OAAO,EAAE,GAAG,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,OAAO,CAAC,EAAE,EAAE,CAAC,CAAC;AAEjF;;;GAGG;AACH,MAAM,CAAC,MAAM,QAAQ,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE,CAC5D,GAAG,CAAC,SAAS;KACV,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,GAAG,EAAE,YAAY,CAAC,QAAQ,EAAE,GAAG,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC;KAC3E,MAAM,CAAC,CAAC,EAAE,GAAG,EAAE,EAAE,EAAE,CAAC,GAAG,CAAC,MAAM,GAAG,OAAO,CAAC,KAAK,CAAC;KAC/C,GAAG,CAAC,CAAC,EAAE,QAAQ,EAAE,GAAG,EAAE,EAAE,EAAE,CAAC,CAAC;IAC3B,IAAI,EAAE,sBAAsB;IAC5B,QAAQ,EAAE,SAAS;IACnB,IAAI,EAAE,CAAC;IACP,MAAM,EAAE,CAAC;IACT,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE;IAC3B,MAAM,EAAE,EAAE,IAAI,EAAE,GAAG,EAAE,KAAK,EAAE,GAAG,CAAC,MAAM,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,EAAE;CAC5F,CAAC,CAAC,CAAC;AAER,MAAM,UAAU,GAAG,KAAK,CAAC;AAEzB;;;GAGG;AACH,MAAM,CAAC,MAAM,SAAS,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE,CAC7D,GAAG,CAAC,SAAS;KACV,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAC,GAAG,SAAS,CAAC,QAAQ,CAAC,CAAC,QAAQ,CAAC,UAAU,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC;KAC9F,MAAM,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC;KAC5C,GAAG,CAAC,CAAC,EAAE,QAAQ,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,CAAC;IAC7B,IAAI,EAAE,sBAAsB;IAC5B,QAAQ,EAAE,MAAM;IAChB,IAAI,EAAE,CAAC;IACP,MAAM,EAAE,CAAC;IACT,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE;IAC3B,MAAM,EAAE,EAAE,KAAK,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,EAAE;CACrE,CAAC,CAAC,CAAC"}
+{"version":3,"file":"char-shape.js","sourceRoot":"","sources":["../../src/detectors/char-shape.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,SAAS,EAAE,OAAO,EAAE,MAAM,iBAAiB,CAAC;AACrD,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACpD,OAAO,EAAE,YAAY,EAAE,MAAM,gBAAgB,CAAC;AAC9C,OAAO,EAAE,SAAS,EAAE,MAAM,eAAe,CAAC;AAE1C;;;GAGG;AACH,MAAM,SAAS,GAAG,UAAU,CAAC;AAE7B,0DAA0D;AAC1D,MAAM,WAAW,GAAG,CAAC,QAAkB,EAAE,GAAW,EAAE,QAA6B,EAAW,EAAE;IAC9F,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,CAAC;IAC/B,MAAM,KAAK,GAAG,MAAM,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,SAAS,CAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC,EAAE,GAAG,CAAC,CAAC;IAChG,IAAI,MAAM,KAAK,SAAS,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,KAAK,CAAC;IAC9D,MAAM,CAAC,KAAK,EAAE,GAAG,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC;IAC1F,MAAM,QAAQ,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,GAAG,IAAI,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;IACrH,OAAO,YAAY,CAAC,MAAM,EAAE,QAAQ,EAAE,QAAQ,CAAC,CAAC;AAClD,CAAC,CAAC;AAEF,MAAM,YAAY,GAAG,CAAC,QAAkB,EAAE,OAAoC,EAAE,QAA6B,EAAU,EAAE,CACvH,CAAC,GAAG,aAAa,CAAC,SAAS,CAAC,QAAQ,CAAC,EAAE,OAAO,CAAC,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC;KACjE,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;KACxB,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,WAAW,CAAC,QAAQ,EAAE,GAAG,EAAE,QAAQ,CAAC,CAAC;KACtD,MAAM,CAAC,CAAC,OAAO,EAAE,GAAG,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,OAAO,CAAC,EAAE,EAAE,CAAC,CAAC;AAEjF;;;GAGG;AACH,MAAM,CAAC,MAAM,QAAQ,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IAC5D,MAAM,QAAQ,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,CAAC,QAAQ,CAAC,iBAAiB,CAAC,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC;IAChG,OAAO,GAAG,CAAC,SAAS;SACjB,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,GAAG,EAAE,YAAY,CAAC,QAAQ,EAAE,GAAG,CAAC,OAAO,EAAE,QAAQ,CAAC,EAAE,CAAC,CAAC;SACrF,MAAM,CAAC,CAAC,EAAE,GAAG,EAAE,EAAE,EAAE,CAAC,GAAG,CAAC,MAAM,GAAG,OAAO,CAAC,KAAK,CAAC;SAC/C,GAAG,CAAC,CAAC,EAAE,QAAQ,EAAE,GAAG,EAAE,EAAE,EAAE,CAAC,CAAC;QAC3B,IAAI,EAAE,sBAAsB;QAC5B,QAAQ,EAAE,SAAS;QACnB,IAAI,EAAE,CAAC;QACP,MAAM,EAAE,CAAC;QACT,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE;QAC3B,MAAM,EAAE,EAAE,IAAI,EAAE,GAAG,EAAE,KAAK,EAAE,GAAG,CAAC,MAAM,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,EAAE;KAC5F,CAAC,CAAC,CAAC;AACR,CAAC,CAAC;AAEF,MAAM,UAAU,GAAG,KAAK,CAAC;AAEzB;;;GAGG;AACH,MAAM,CAAC,MAAM,SAAS,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE,CAC7D,GAAG,CAAC,SAAS;KACV,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAC,GAAG,SAAS,CAAC,QAAQ,CAAC,CAAC,QAAQ,CAAC,UAAU,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,CAAC;KAC9F,MAAM,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC;KAC5C,GAAG,CAAC,CAAC,EAAE,QAAQ,EAAE,KAAK,EAAE,EAAE,EAAE,CAAC,CAAC;IAC7B,IAAI,EAAE,sBAAsB;IAC5B,QAAQ,EAAE,MAAM;IAChB,IAAI,EAAE,CAAC;IACP,MAAM,EAAE,CAAC;IACT,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE;IAC3B,MAAM,EAAE,EAAE,KAAK,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,EAAE;CACrE,CAAC,CAAC,CAAC"}

package/dist/detectors/orthography.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"orthography.d.ts","sourceRoot":"","sources":["../../src/detectors/orthography.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAAqB,MAAM,cAAc,CAAC;AAGhE,6DAA6D;AAC7D,eAAO,MAAM,aAAa,EAAE,QAc3B,CAAC;AAOF;;;GAGG;AACH,eAAO,MAAM,YAAY,EAAE,QAyB1B,CAAC"}
+{"version":3,"file":"orthography.d.ts","sourceRoot":"","sources":["../../src/detectors/orthography.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAAqB,MAAM,cAAc,CAAC;AAIhE,6DAA6D;AAC7D,eAAO,MAAM,aAAa,EAAE,QAc3B,CAAC;AAoBF;;;GAGG;AACH,eAAO,MAAM,YAAY,EAAE,QA6B1B,CAAC"}

package/dist/detectors/orthography.js

@@ -1,4 +1,5 @@
 import { latinBoundaries, minorityStyle, occurrencesOutside } from "../orthography.js";
+import { digitRunAround, isNumberName } from "../number-name.js";
 /** チームが chaff.yaml の prefer に書いた「使わない書き方」。書いていなければ何も言わない。 */
 export const preferredTerm = (doc, options) => {
     const pairs = (options.lexicon ?? []).flatMap((entry) => (entry.instead_of === undefined ? [] : [{ avoid: entry.pattern, use: entry.instead_of }]));
@@ -16,12 +17,27 @@ export const preferredTerm = (doc, options) => {
 };
 const KIND_NAME = { letter: "英字", "before-digit": "後ろの数字", "after-digit": "前の数字" };
 const KINDS = ["letter", "before-digit", "after-digit"];
+/** 境目の隣の数字の、並びの中の 1 字の位置。数字の後ろなら左の字、数字の前なら（空白を越えた）右の字。 */
+const digitBeside = (boundary) => {
+    if (boundary.kind === "after-digit")
+        return boundary.offset - 1;
+    return boundary.spaced ? boundary.offset + 1 : boundary.offset;
+};
+/** 番号・識別子として書かれた数の境目は、空け方の好みではないので数えない（number-name.ts）。 */
+const isCounted = (sentence, boundary) => {
+    if (boundary.kind === "letter")
+        return true;
+    const run = digitRunAround(sentence.text, digitBeside(boundary));
+    return run === undefined || !isNumberName(sentence.text, run, sentence.tokens, sentence.span.start);
+};
 /**
  * 日本語と英字・数字のあいだを、空けるか詰めるか。文書の中で混ざっていたら、少ないほうを指摘する。
  * どちらが正しいかは決めない。決めるのはチームで、chaff はそろっているかだけを見る。
  */
 export const latinSpacing = (doc, options) => {
-    const located = doc.sentences.flatMap((sentence) => latinBoundaries(sentence.text).map((boundary) => ({ sentence, ...boundary })));
+    const located = doc.sentences.flatMap((sentence) => latinBoundaries(sentence.text, doc.source.slice(sentence.span.start, sentence.span.end))
+        .filter((boundary) => isCounted(sentence, boundary))
+        .map((boundary) => ({ sentence, ...boundary })));
     return KINDS.flatMap((kind) => {
         const ofKind = located.filter((entry) => entry.kind === kind);
         const minority = minorityStyle(ofKind);

package/dist/detectors/orthography.js.map

@@ -1 +1 @@
-{"version":3,"file":"orthography.js","sourceRoot":"","sources":["../../src/detectors/orthography.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,eAAe,EAAE,aAAa,EAAE,kBAAkB,EAAoB,MAAM,mBAAmB,CAAC;AAEzG,6DAA6D;AAC7D,MAAM,CAAC,MAAM,aAAa,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IACjE,MAAM,KAAK,GAAG,CAAC,OAAO,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,UAAU,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,OAAO,EAAE,GAAG,EAAE,KAAK,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,CAAC;IACpJ,MAAM,IAAI,GAAG,GAAG,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE,CAC9C,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,kBAAkB,CAAC,QAAQ,CAAC,IAAI,EAAE,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CACvH,CAAC;IACF,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC,IAAI,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,KAAK;QAAE,OAAO,EAAE,CAAC;IAChE,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;QACxB,IAAI,EAAE,EAAE;QACR,QAAQ,EAAE,SAAS;QACnB,IAAI,EAAE,CAAC;QACP,MAAM,EAAE,CAAC;QACT,KAAK,EAAE,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE;QAC/B,MAAM,EAAE,EAAE,OAAO,EAAE,GAAG,CAAC,IAAI,CAAC,KAAK,EAAE,SAAS,EAAE,GAAG,CAAC,IAAI,CAAC,GAAG,EAAE,KAAK,EAAE,IAAI,CAAC,MAAM,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,GAAG,CAAC,EAAE,EAAE;KACjJ,CAAC,CAAC,CAAC;AACN,CAAC,CAAC;AAEF,MAAM,SAAS,GAA0C,EAAE,MAAM,EAAE,IAAI,EAAE,cAAc,EAAE,OAAO,EAAE,aAAa,EAAE,MAAM,EAAE,CAAC;AAC1H,MAAM,KAAK,GAA2B,CAAC,QAAQ,EAAE,cAAc,EAAE,aAAa,CAAC,CAAC;AAIhF;;;GAGG;AACH,MAAM,CAAC,MAAM,YAAY,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IAChE,MAAM,OAAO,GAAc,GAAG,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,eAAe,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,GAAG,QAAQ,EAAE,CAAC,CAAC,CAAC,CAAC;IAC9I,OAAO,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE;QAC5B,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,KAAK,IAAI,CAAC,CAAC;QAC9D,MAAM,QAAQ,GAAG,aAAa,CAAC,MAAM,CAAC,CAAC;QACvC,IAAI,QAAQ,KAAK,SAAS;YAAE,OAAO,EAAE,CAAC;QACtC,MAAM,GAAG,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,KAAK,QAAQ,CAAC,CAAC;QAChE,IAAI,GAAG,CAAC,MAAM,GAAG,OAAO,CAAC,KAAK;YAAE,OAAO,EAAE,CAAC;QAC1C,OAAO,GAAG,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;YACzB,IAAI,EAAE,EAAE;YACR,QAAQ,EAAE,SAAS;YACnB,IAAI,EAAE,CAAC;YACP,MAAM,EAAE,CAAC;YACT,KAAK,EAAE,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE;YACjC,MAAM,EAAE;gBACN,IAAI,EAAE,SAAS,CAAC,IAAI,CAAC;gBACrB,KAAK,EAAE,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,QAAQ;gBACrC,KAAK,EAAE,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK;gBAC/B,KAAK,EAAE,GAAG,CAAC,MAAM;gBACjB,EAAE,EAAE,MAAM,CAAC,MAAM;gBACjB,KAAK,EAAE,OAAO,CAAC,KAAK;gBACpB,MAAM,EAAE,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,MAAM;aACjD;SACF,CAAC,CAAC,CAAC;IACN,CAAC,CAAC,CAAC;AACL,CAAC,CAAC"}
+{"version":3,"file":"orthography.js","sourceRoot":"","sources":["../../src/detectors/orthography.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,eAAe,EAAE,aAAa,EAAE,kBAAkB,EAAmC,MAAM,mBAAmB,CAAC;AACxH,OAAO,EAAE,cAAc,EAAE,YAAY,EAAE,MAAM,mBAAmB,CAAC;AAEjE,6DAA6D;AAC7D,MAAM,CAAC,MAAM,aAAa,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IACjE,MAAM,KAAK,GAAG,CAAC,OAAO,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,KAAK,CAAC,UAAU,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,OAAO,EAAE,GAAG,EAAE,KAAK,CAAC,UAAU,EAAE,CAAC,CAAC,CAAC,CAAC;IACpJ,MAAM,IAAI,GAAG,GAAG,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE,CAC9C,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,kBAAkB,CAAC,QAAQ,CAAC,IAAI,EAAE,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CACvH,CAAC;IACF,IAAI,IAAI,CAAC,MAAM,KAAK,CAAC,IAAI,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,KAAK;QAAE,OAAO,EAAE,CAAC;IAChE,OAAO,IAAI,CAAC,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;QACxB,IAAI,EAAE,EAAE;QACR,QAAQ,EAAE,SAAS;QACnB,IAAI,EAAE,CAAC;QACP,MAAM,EAAE,CAAC;QACT,KAAK,EAAE,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE;QAC/B,MAAM,EAAE,EAAE,OAAO,EAAE,GAAG,CAAC,IAAI,CAAC,KAAK,EAAE,SAAS,EAAE,GAAG,CAAC,IAAI,CAAC,GAAG,EAAE,KAAK,EAAE,IAAI,CAAC,MAAM,EAAE,KAAK,EAAE,OAAO,CAAC,KAAK,EAAE,MAAM,EAAE,GAAG,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,GAAG,CAAC,EAAE,EAAE;KACjJ,CAAC,CAAC,CAAC;AACN,CAAC,CAAC;AAEF,MAAM,SAAS,GAA0C,EAAE,MAAM,EAAE,IAAI,EAAE,cAAc,EAAE,OAAO,EAAE,aAAa,EAAE,MAAM,EAAE,CAAC;AAC1H,MAAM,KAAK,GAA2B,CAAC,QAAQ,EAAE,cAAc,EAAE,aAAa,CAAC,CAAC;AAIhF,0DAA0D;AAC1D,MAAM,WAAW,GAAG,CAAC,QAAkB,EAAU,EAAE;IACjD,IAAI,QAAQ,CAAC,IAAI,KAAK,aAAa;QAAE,OAAO,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC;IAChE,OAAO,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,MAAM,CAAC;AACjE,CAAC,CAAC;AAEF,2DAA2D;AAC3D,MAAM,SAAS,GAAG,CAAC,QAAkB,EAAE,QAAkB,EAAW,EAAE;IACpE,IAAI,QAAQ,CAAC,IAAI,KAAK,QAAQ;QAAE,OAAO,IAAI,CAAC;IAC5C,MAAM,GAAG,GAAG,cAAc,CAAC,QAAQ,CAAC,IAAI,EAAE,WAAW,CAAC,QAAQ,CAAC,CAAC,CAAC;IACjE,OAAO,GAAG,KAAK,SAAS,IAAI,CAAC,YAAY,CAAC,QAAQ,CAAC,IAAI,EAAE,GAAG,EAAE,QAAQ,CAAC,MAAM,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;AACtG,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,CAAC,MAAM,YAAY,GAAa,CAAC,GAAG,EAAE,OAAO,EAAa,EAAE;IAChE,MAAM,OAAO,GAAc,GAAG,CAAC,SAAS,CAAC,OAAO,CAAC,CAAC,QAAQ,EAAE,EAAE,CAC5D,eAAe,CAAC,QAAQ,CAAC,IAAI,EAAE,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;SACrF,MAAM,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,SAAS,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC;SACnD,GAAG,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC,EAAE,QAAQ,EAAE,GAAG,QAAQ,EAAE,CAAC,CAAC,CAClD,CAAC;IACF,OAAO,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,EAAE;QAC5B,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,KAAK,IAAI,CAAC,CAAC;QAC9D,MAAM,QAAQ,GAAG,aAAa,CAAC,MAAM,CAAC,CAAC;QACvC,IAAI,QAAQ,KAAK,SAAS;YAAE,OAAO,EAAE,CAAC;QACtC,MAAM,GAAG,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,KAAK,QAAQ,CAAC,CAAC;QAChE,IAAI,GAAG,CAAC,MAAM,GAAG,OAAO,CAAC,KAAK;YAAE,OAAO,EAAE,CAAC;QAC1C,OAAO,GAAG,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;YACzB,IAAI,EAAE,EAAE;YACR,QAAQ,EAAE,SAAS;YACnB,IAAI,EAAE,CAAC;YACP,MAAM,EAAE,CAAC;YACT,KAAK,EAAE,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,EAAE;YACjC,MAAM,EAAE;gBACN,IAAI,EAAE,SAAS,CAAC,IAAI,CAAC;gBACrB,KAAK,EAAE,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,QAAQ;gBACrC,KAAK,EAAE,QAAQ,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK;gBAC/B,KAAK,EAAE,GAAG,CAAC,MAAM;gBACjB,EAAE,EAAE,MAAM,CAAC,MAAM;gBACjB,KAAK,EAAE,OAAO,CAAC,KAAK;gBACpB,MAAM,EAAE,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,MAAM;aACjD;SACF,CAAC,CAAC,CAAC;IACN,CAAC,CAAC,CAAC;AACL,CAAC,CAAC"}

package/dist/detectors/place-run.d.ts

@@ -0,0 +1,8 @@
+import type { Token } from "../plugin.ts";
+/**
+ * 漢字の連なりの語（covering: tokens の添字）が住所か（東京都港区新橋二丁目）。地名で始まり、住所の語だけでできている。
+ * 住所は決まった形で、ひらがなを挟んで書けない。組織名（日本経済団体連合会）は地名の後ろに普通の語が続くので住所ではない。
+ * topUnits は、住所の先頭にしか来ない単位（都道府県）。言語の語彙表から渡す。
+ */
+export declare const isAddressRun: (tokens: readonly Token[], covering: readonly number[], topUnits: ReadonlySet<string>) => boolean;
+//# sourceMappingURL=place-run.d.ts.map

package/dist/detectors/place-run.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"place-run.d.ts","sourceRoot":"","sources":["../../src/detectors/place-run.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,KAAK,EAAE,MAAM,cAAc,CAAC;AAoD1C;;;;GAIG;AACH,eAAO,MAAM,YAAY,GAAI,QAAQ,SAAS,KAAK,EAAE,EAAE,UAAU,SAAS,MAAM,EAAE,EAAE,UAAU,WAAW,CAAC,MAAM,CAAC,KAAG,OAKnH,CAAC"}

package/dist/detectors/place-run.js

@@ -0,0 +1,54 @@
+/** 地名（Geo）と、地名に付く単位（GeoUnit: 都・県・郡・市・区・町・村）。 */
+const isGeoName = (token) => token?.features?.["NameType"] === "Geo";
+const isGeoUnit = (token) => token?.features?.["NameType"] === "GeoUnit";
+const isGeo = (token) => isGeoName(token) || isGeoUnit(token);
+const isNumber = (token) => token?.features?.["NumType"] === "Card";
+/**
+ * 辞書に無い町名は、形態素解析が固有名詞か 1 字ずつに割る（紀美野町は 紀＋美野、南伊勢町は 南＋伊勢、久米島町は固有名詞,一般）。
+ * 2 字以上の普通の語（政策・総務）は地名のかけらにしない。
+ */
+const isNamePiece = (token) => token !== undefined && (token.pos === "PROPN" || [...token.surface].length === 1);
+/**
+ * index から地名のかけらだけを読み進めて、連なりの終わりまでに都道府県より下の単位（郡・市・町）で閉じるか。
+ * 閉じれば、そこまでが 1 つの地名。連なりの外（「に」「の」も 1 字）までは読まない。
+ */
+const closedByUnit = (run, index) => {
+    const token = run.tokens[index];
+    if (index > run.last || token === undefined)
+        return false;
+    if (isGeoUnit(token))
+        return !run.topUnits.has(token.surface);
+    return isNamePiece(token) && closedByUnit(run, index + 1);
+};
+/**
+ * 住所の語: 地名と地名の単位（NameType=Geo / GeoUnit）、数（NumType=Card）、数のすぐ後ろの助数詞（丁目、NounType=Class）。
+ * 辞書に無い町名（新橋）は、地名の後ろで数の前に来るときだけ住所の一部とする。
+ * 割られた町名（紀美野町）は、単位の後ろで、地名のかけらだけを挟んで次の単位で閉じるときだけ住所の一部とする。
+ */
+const isPlacePart = (run, index) => {
+    const tokens = run.tokens;
+    return (isGeo(tokens[index]) ||
+        isNumber(tokens[index]) ||
+        (isNumber(tokens[index - 1]) && tokens[index]?.features?.["NounType"] === "Class") ||
+        (isGeo(tokens[index - 1]) && isNumber(tokens[index + 1])) ||
+        (isGeo(tokens[index - 1]) && isNamePiece(tokens[index]) && closedByUnit(run, index + 1)));
+};
+/**
+ * 地名が単位を挟まずに続けば地名の並び（東京大阪名古屋福岡）。ただし 2 つだけで郡・市・町が閉じれば、割られた 1 つの地名
+ * （北海道虻田郡、富士河口湖町）。3 つ続けば、単位が付いても並び（京都奈良大阪神戸市）。
+ * 名古屋大阪市は北海道札幌市と品詞が同じ形なので、ここでは分けられない。
+ */
+const isListed = (run, index) => isGeoName(run.tokens[index]) && isGeoName(run.tokens[index + 1]) && (isGeoName(run.tokens[index + 2]) || !closedByUnit(run, index + 1));
+/**
+ * 漢字の連なりの語（covering: tokens の添字）が住所か（東京都港区新橋二丁目）。地名で始まり、住所の語だけでできている。
+ * 住所は決まった形で、ひらがなを挟んで書けない。組織名（日本経済団体連合会）は地名の後ろに普通の語が続くので住所ではない。
+ * topUnits は、住所の先頭にしか来ない単位（都道府県）。言語の語彙表から渡す。
+ */
+export const isAddressRun = (tokens, covering, topUnits) => {
+    const [first, last] = [covering[0], covering.at(-1)];
+    if (first === undefined || last === undefined)
+        return false;
+    const run = { tokens, last, topUnits };
+    return isGeo(tokens[first]) && !covering.some((index) => isListed(run, index)) && covering.every((index) => isPlacePart(run, index));
+};
+//# sourceMappingURL=place-run.js.map

package/dist/detectors/place-run.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"place-run.js","sourceRoot":"","sources":["../../src/detectors/place-run.ts"],"names":[],"mappings":"AAEA,gDAAgD;AAChD,MAAM,SAAS,GAAG,CAAC,KAAwB,EAAW,EAAE,CAAC,KAAK,EAAE,QAAQ,EAAE,CAAC,UAAU,CAAC,KAAK,KAAK,CAAC;AACjG,MAAM,SAAS,GAAG,CAAC,KAAwB,EAAW,EAAE,CAAC,KAAK,EAAE,QAAQ,EAAE,CAAC,UAAU,CAAC,KAAK,SAAS,CAAC;AACrG,MAAM,KAAK,GAAG,CAAC,KAAwB,EAAW,EAAE,CAAC,SAAS,CAAC,KAAK,CAAC,IAAI,SAAS,CAAC,KAAK,CAAC,CAAC;AAC1F,MAAM,QAAQ,GAAG,CAAC,KAAwB,EAAW,EAAE,CAAC,KAAK,EAAE,QAAQ,EAAE,CAAC,SAAS,CAAC,KAAK,MAAM,CAAC;AAEhG;;;GAGG;AACH,MAAM,WAAW,GAAG,CAAC,KAAwB,EAAW,EAAE,CAAC,KAAK,KAAK,SAAS,IAAI,CAAC,KAAK,CAAC,GAAG,KAAK,OAAO,IAAI,CAAC,GAAG,KAAK,CAAC,OAAO,CAAC,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC;AAK7I;;;GAGG;AACH,MAAM,YAAY,GAAG,CAAC,GAAQ,EAAE,KAAa,EAAW,EAAE;IACxD,MAAM,KAAK,GAAG,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;IAChC,IAAI,KAAK,GAAG,GAAG,CAAC,IAAI,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,KAAK,CAAC;IAC1D,IAAI,SAAS,CAAC,KAAK,CAAC;QAAE,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IAC9D,OAAO,WAAW,CAAC,KAAK,CAAC,IAAI,YAAY,CAAC,GAAG,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC;AAC5D,CAAC,CAAC;AAEF;;;;GAIG;AACH,MAAM,WAAW,GAAG,CAAC,GAAQ,EAAE,KAAa,EAAW,EAAE;IACvD,MAAM,MAAM,GAAG,GAAG,CAAC,MAAM,CAAC;IAC1B,OAAO,CACL,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QACpB,QAAQ,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;QACvB,CAAC,QAAQ,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,MAAM,CAAC,KAAK,CAAC,EAAE,QAAQ,EAAE,CAAC,UAAU,CAAC,KAAK,OAAO,CAAC;QAClF,CAAC,KAAK,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,QAAQ,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC;QACzD,CAAC,KAAK,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,WAAW,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,IAAI,YAAY,CAAC,GAAG,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC,CACzF,CAAC;AACJ,CAAC,CAAC;AAEF;;;;GAIG;AACH,MAAM,QAAQ,GAAG,CAAC,GAAQ,EAAE,KAAa,EAAW,EAAE,CACpD,SAAS,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,IAAI,SAAS,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,YAAY,CAAC,GAAG,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC;AAE1I;;;;GAIG;AACH,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,MAAwB,EAAE,QAA2B,EAAE,QAA6B,EAAW,EAAE;IAC5H,MAAM,CAAC,KAAK,EAAE,IAAI,CAAC,GAAG,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,QAAQ,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IACrD,IAAI,KAAK,KAAK,SAAS,IAAI,IAAI,KAAK,SAAS;QAAE,OAAO,KAAK,CAAC;IAC5D,MAAM,GAAG,GAAQ,EAAE,MAAM,EAAE,IAAI,EAAE,QAAQ,EAAE,CAAC;IAC5C,OAAO,KAAK,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,QAAQ,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC,IAAI,QAAQ,CAAC,KAAK,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,WAAW,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC,CAAC;AACvI,CAAC,CAAC"}

package/dist/detectors/register.d.ts

@@ -0,0 +1,32 @@
+import type { Sentence, Span, Token } from "../plugin.ts";
+export type Register = "polite" | "plain";
+/**
+ * 文末の語の調子。preceding は文末の語より前の語。
+ * 文末に述語（動詞・形容詞・助動詞）が無ければ、ですます調でもである調でもないので undefined。
+ * 「以下の通り。」「円錐形の麦わら帽子。」のような名詞で終わる文を数えると、そのままである調の少数派になっていた。
+ * 書いた形でも原形でも当てる。「ください」の原形は「くださる」で、原形だけを見ると丁寧な文末を見落とす。
+ */
+export declare const registerOf: (ending: readonly Token[], preceding: readonly Token[], polite: readonly string[]) => Register | undefined;
+/**
+ * 「！」「？」で切れた文が、間を置かずに助詞で始まる次の文へ続くか。「作っていた！！という人」「導入しませんか？が断られた」は、
+ * 分割器が切っても一つの文で、切れ目の手前は書き手の文末ではない。「。」で切れた文は続きとは見ない。
+ */
+export declare const continuesInto: (sentence: Sentence, next: Sentence | undefined) => boolean;
+/** 位置 offset を含む一番外側の箇条書きの始まり。本文なら undefined。入れ子の項目は外側の箇条書きと一緒に見る。 */
+export declare const outermostList: (offset: number, lists: readonly Span[]) => number | undefined;
+/** group: 文が入っている箇条書き（outermostList の値）。本文なら undefined。 */
+export type Judged = {
+    readonly register: Register;
+    readonly group: number | undefined;
+};
+/** 混ざった文と、その文が属する本文または箇条書きの中の少数派の数。 */
+export type Slip<T extends Judged> = {
+    readonly entry: T;
+    readonly count: number;
+};
+/**
+ * 調子が混ざった文を、judged の順で。本文は本文どうし、箇条書きは 1 つずつ、その中で揃っているかを見る。
+ * ですます調の本文に常体の箇条書きを置くのはよくある書き方で、箇条書きが丸ごと揃っていれば混在ではない。
+ */
+export declare const slipsOf: <T extends Judged>(judged: readonly T[], limit: number) => Slip<T>[];
+//# sourceMappingURL=register.d.ts.map

package/dist/detectors/register.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"register.d.ts","sourceRoot":"","sources":["../../src/detectors/register.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,QAAQ,EAAE,IAAI,EAAE,KAAK,EAAE,MAAM,cAAc,CAAC;AAE1D,MAAM,MAAM,QAAQ,GAAG,QAAQ,GAAG,OAAO,CAAC;AAqB1C;;;;;GAKG;AACH,eAAO,MAAM,UAAU,GAAI,QAAQ,SAAS,KAAK,EAAE,EAAE,WAAW,SAAS,KAAK,EAAE,EAAE,QAAQ,SAAS,MAAM,EAAE,KAAG,QAAQ,GAAG,SAKxH,CAAC;AAIF;;;GAGG;AACH,eAAO,MAAM,aAAa,GAAI,UAAU,QAAQ,EAAE,MAAM,QAAQ,GAAG,SAAS,KAAG,OACkD,CAAC;AAElI,sEAAsE;AACtE,eAAO,MAAM,aAAa,GAAI,QAAQ,MAAM,EAAE,OAAO,SAAS,IAAI,EAAE,KAAG,MAAM,GAAG,SAG6C,CAAC;AAE9H,2DAA2D;AAC3D,MAAM,MAAM,MAAM,GAAG;IAAE,QAAQ,CAAC,QAAQ,EAAE,QAAQ,CAAC;IAAC,QAAQ,CAAC,KAAK,EAAE,MAAM,GAAG,SAAS,CAAA;CAAE,CAAC;AAEzF,uCAAuC;AACvC,MAAM,MAAM,IAAI,CAAC,CAAC,SAAS,MAAM,IAAI;IAAE,QAAQ,CAAC,KAAK,EAAE,CAAC,CAAC;IAAC,QAAQ,CAAC,KAAK,EAAE,MAAM,CAAA;CAAE,CAAC;AAoBnF;;;GAGG;AACH,eAAO,MAAM,OAAO,GAAI,CAAC,SAAS,MAAM,EAAE,QAAQ,SAAS,CAAC,EAAE,EAAE,OAAO,MAAM,KAAG,IAAI,CAAC,CAAC,CAAC,EAGtF,CAAC"}

package/dist/detectors/register.js

@@ -0,0 +1,62 @@
+const PREDICATE = new Set(["VERB", "ADJ", "AUX"]);
+const isPredicate = (token) => PREDICATE.has(token.pos);
+const isDependent = (token) => token?.features?.["NounType"] === "Dependent";
+/** preceding の後ろの述語の連なり: 助動詞の連なりと、その前の語一つ。「おかけしましたこと」なら「し・まし・た」。 */
+const chainAtEnd = (preceding) => {
+    const head = preceding.findLastIndex((token) => token.pos !== "AUX");
+    return preceding.slice(Math.max(head, 0));
+};
+/**
+ * 非自立名詞で終わる文末は、手前の述語の連なりまで見る。「予約できること。」「おかけしましたこと。」は手前の述語が調子を持つ。
+ * 要件や規程のである調は「〜こと。」で書く。「以下のとおり。」は手前も述語でないので、調子を持たないまま。
+ */
+const withPreceding = (ending, preceding) => isDependent(ending[0]) ? [...chainAtEnd(preceding), ...ending] : ending;
+/**
+ * 文末の語の調子。preceding は文末の語より前の語。
+ * 文末に述語（動詞・形容詞・助動詞）が無ければ、ですます調でもである調でもないので undefined。
+ * 「以下の通り。」「円錐形の麦わら帽子。」のような名詞で終わる文を数えると、そのままである調の少数派になっていた。
+ * 書いた形でも原形でも当てる。「ください」の原形は「くださる」で、原形だけを見ると丁寧な文末を見落とす。
+ */
+export const registerOf = (ending, preceding, polite) => {
+    const judged = withPreceding(ending, preceding);
+    if (!judged.some(isPredicate))
+        return undefined;
+    const isPolite = judged.some((token) => polite.includes(token.surface) || (token.lemma !== undefined && polite.includes(token.lemma)));
+    return isPolite ? "polite" : "plain";
+};
+const EXCLAIMED = /[！？!?]$/u;
+/**
+ * 「！」「？」で切れた文が、間を置かずに助詞で始まる次の文へ続くか。「作っていた！！という人」「導入しませんか？が断られた」は、
+ * 分割器が切っても一つの文で、切れ目の手前は書き手の文末ではない。「。」で切れた文は続きとは見ない。
+ */
+export const continuesInto = (sentence, next) => EXCLAIMED.test(sentence.text) && next !== undefined && next.span.start === sentence.span.end && next.tokens?.[0]?.pos === "ADP";
+/** 位置 offset を含む一番外側の箇条書きの始まり。本文なら undefined。入れ子の項目は外側の箇条書きと一緒に見る。 */
+export const outermostList = (offset, lists) => lists
+    .filter((list) => offset >= list.start && offset < list.end)
+    .reduce((outer, list) => (outer === undefined || list.start < outer ? list.start : outer), undefined);
+const countIn = (judged, register) => judged.filter((entry) => entry.register === register).length;
+/** 少ないほう。同数なら文書全体で少ないほう、それも同数なら丁寧体。 */
+const minorityOf = (group, whole) => {
+    const difference = countIn(group, "polite") - countIn(group, "plain");
+    if (difference !== 0)
+        return difference < 0 ? "polite" : "plain";
+    return countIn(whole, "polite") <= countIn(whole, "plain") ? "polite" : "plain";
+};
+const slipsInGroup = (judged, group, limit) => {
+    const members = judged.filter((entry) => entry.group === group);
+    const minority = minorityOf(members, judged);
+    const slips = members.filter((entry) => entry.register === minority);
+    // 片方しか無ければ揃っている。少数派が閾値を超えて多ければ、混在ではなく別の文体。
+    if (slips.length === 0 || slips.length > limit)
+        return [];
+    return slips.map((entry) => ({ entry, count: slips.length }));
+};
+/**
+ * 調子が混ざった文を、judged の順で。本文は本文どうし、箇条書きは 1 つずつ、その中で揃っているかを見る。
+ * ですます調の本文に常体の箇条書きを置くのはよくある書き方で、箇条書きが丸ごと揃っていれば混在ではない。
+ */
+export const slipsOf = (judged, limit) => {
+    const slips = [...new Set(judged.map((entry) => entry.group))].flatMap((group) => slipsInGroup(judged, group, limit));
+    return judged.flatMap((entry) => slips.filter((slip) => slip.entry === entry));
+};
+//# sourceMappingURL=register.js.map

package/dist/detectors/register.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"register.js","sourceRoot":"","sources":["../../src/detectors/register.ts"],"names":[],"mappings":"AAIA,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,KAAK,EAAE,KAAK,CAAC,CAAC,CAAC;AAElD,MAAM,WAAW,GAAG,CAAC,KAAY,EAAW,EAAE,CAAC,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;AAExE,MAAM,WAAW,GAAG,CAAC,KAAwB,EAAW,EAAE,CAAC,KAAK,EAAE,QAAQ,EAAE,CAAC,UAAU,CAAC,KAAK,WAAW,CAAC;AAEzG,oEAAoE;AACpE,MAAM,UAAU,GAAG,CAAC,SAA2B,EAAoB,EAAE;IACnE,MAAM,IAAI,GAAG,SAAS,CAAC,aAAa,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,GAAG,KAAK,KAAK,CAAC,CAAC;IACrE,OAAO,SAAS,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC;AAC5C,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,aAAa,GAAG,CAAC,MAAwB,EAAE,SAA2B,EAAoB,EAAE,CAChG,WAAW,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,UAAU,CAAC,SAAS,CAAC,EAAE,GAAG,MAAM,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;AAE1E;;;;;GAKG;AACH,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,MAAwB,EAAE,SAA2B,EAAE,MAAyB,EAAwB,EAAE;IACnI,MAAM,MAAM,GAAG,aAAa,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC;IAChD,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC;QAAE,OAAO,SAAS,CAAC;IAChD,MAAM,QAAQ,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,KAAK,SAAS,IAAI,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;IACvI,OAAO,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC;AACvC,CAAC,CAAC;AAEF,MAAM,SAAS,GAAG,UAAU,CAAC;AAE7B;;;GAGG;AACH,MAAM,CAAC,MAAM,aAAa,GAAG,CAAC,QAAkB,EAAE,IAA0B,EAAW,EAAE,CACvF,SAAS,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,IAAI,KAAK,SAAS,IAAI,IAAI,CAAC,IAAI,CAAC,KAAK,KAAK,QAAQ,CAAC,IAAI,CAAC,GAAG,IAAI,IAAI,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC,EAAE,GAAG,KAAK,KAAK,CAAC;AAElI,sEAAsE;AACtE,MAAM,CAAC,MAAM,aAAa,GAAG,CAAC,MAAc,EAAE,KAAsB,EAAsB,EAAE,CAC1F,KAAK;KACF,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,IAAI,IAAI,CAAC,KAAK,IAAI,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC;KAC3D,MAAM,CAAqB,CAAC,KAAK,EAAE,IAAI,EAAE,EAAE,CAAC,CAAC,KAAK,KAAK,SAAS,IAAI,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,EAAE,SAAS,CAAC,CAAC;AAQ9H,MAAM,OAAO,GAAG,CAAC,MAAyB,EAAE,QAAkB,EAAU,EAAE,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,QAAQ,KAAK,QAAQ,CAAC,CAAC,MAAM,CAAC;AAExI,uCAAuC;AACvC,MAAM,UAAU,GAAG,CAAC,KAAwB,EAAE,KAAwB,EAAY,EAAE;IAClF,MAAM,UAAU,GAAG,OAAO,CAAC,KAAK,EAAE,QAAQ,CAAC,GAAG,OAAO,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;IACtE,IAAI,UAAU,KAAK,CAAC;QAAE,OAAO,UAAU,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC;IACjE,OAAO,OAAO,CAAC,KAAK,EAAE,QAAQ,CAAC,IAAI,OAAO,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC;AAClF,CAAC,CAAC;AAEF,MAAM,YAAY,GAAG,CAAmB,MAAoB,EAAE,KAAyB,EAAE,KAAa,EAAa,EAAE;IACnH,MAAM,OAAO,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,KAAK,KAAK,KAAK,CAAC,CAAC;IAChE,MAAM,QAAQ,GAAG,UAAU,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC;IAC7C,MAAM,KAAK,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,QAAQ,KAAK,QAAQ,CAAC,CAAC;IACrE,2CAA2C;IAC3C,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC,IAAI,KAAK,CAAC,MAAM,GAAG,KAAK;QAAE,OAAO,EAAE,CAAC;IAC1D,OAAO,KAAK,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,CAAC,MAAM,EAAE,CAAC,CAAC,CAAC;AAChE,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,CAAC,MAAM,OAAO,GAAG,CAAmB,MAAoB,EAAE,KAAa,EAAa,EAAE;IAC1F,MAAM,KAAK,GAAG,CAAC,GAAG,IAAI,GAAG,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,YAAY,CAAC,MAAM,EAAE,KAAK,EAAE,KAAK,CAAC,CAAC,CAAC;IACtH,OAAO,MAAM,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,KAAK,KAAK,KAAK,CAAC,CAAC,CAAC;AACjF,CAAC,CAAC"}

package/dist/detectors/sentence-ending.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"sentence-ending.d.ts","sourceRoot":"","sources":["../../src/detectors/sentence-ending.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,QAAQ,EAAqB,MAAM,cAAc,CAAC;AAqBhE,eAAO,MAAM,cAAc,EAAE,QAkB5B,CAAC"}
+{"version":3,"file":"sentence-ending.d.ts","sourceRoot":"","sources":["../../src/detectors/sentence-ending.ts"],"names":[],"mappings":"AAEA,OAAO,KAAK,EAAE,QAAQ,EAAqB,MAAM,cAAc,CAAC;AAkBhE,eAAO,MAAM,cAAc,EAAE,QAgB5B,CAAC"}