@chaffjs/lang-en 0.3.0 → 0.6.0

package/dist/citation.d.ts

@@ -0,0 +1,6 @@
+/**
+ * The document named right after a reference, or undefined when the reference is into this document.
+ * "of this Agreement" and "of the Agreement" are this document; "of the Master Agreement" is another.
+ */
+export declare const citedDocumentAfter: (text: string, end: number) => string | undefined;
+//# sourceMappingURL=citation.d.ts.map

package/dist/citation.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"citation.d.ts","sourceRoot":"","sources":["../src/citation.ts"],"names":[],"mappings":"AA4BA;;;GAGG;AACH,eAAO,MAAM,kBAAkB,GAAI,MAAM,MAAM,EAAE,KAAK,MAAM,KAAG,MAAM,GAAG,SAQvE,CAAC"}

package/dist/citation.js

@@ -0,0 +1,43 @@
+// A reference into another document: "Section 9 of the Master Agreement" is not this document's Section 9.
+const OF = /^,? of (?:the )?/u;
+const CAPITALISED = /^[A-Z][\w'’-]*/u;
+/**
+ * How a document names itself. "Section 3 of the Agreement" in an agreement means this one.
+ * "the Act" and "the Code" are left out: a statute calls itself "this Act", and a contract's "the Act" is a law it cites.
+ */
+const SELF = new Set(["Agreement", "Contract", "Terms", "Policy"]);
+/** Words that may sit inside a title: "Code of Federal Regulations". */
+const JOINERS = new Set(["of", "and", "for", "on"]);
+const MAX_WORDS = 8;
+/** Capitalised words from the start of `rest`, allowing a joiner between two of them. Punctuation ends the title. */
+const titleWords = (rest) => {
+    const tokens = rest.split(" ");
+    const words = [];
+    for (let index = 0; index < tokens.length && words.length < MAX_WORDS; index += 1) {
+        const token = tokens[index] ?? "";
+        const word = CAPITALISED.exec(token)?.[0];
+        const joins = word === undefined && words.length > 0 && JOINERS.has(token) && CAPITALISED.test(tokens[index + 1] ?? "");
+        if (word === undefined && !joins)
+            break;
+        words.push(word ?? token);
+        if (word !== undefined && word.length < token.length)
+            break;
+    }
+    return words;
+};
+/**
+ * The document named right after a reference, or undefined when the reference is into this document.
+ * "of this Agreement" and "of the Agreement" are this document; "of the Master Agreement" is another.
+ */
+export const citedDocumentAfter = (text, end) => {
+    const rest = text.slice(end, end + 120);
+    const of = OF.exec(rest);
+    if (of === null)
+        return undefined;
+    const words = titleWords(rest.slice(of[0].length));
+    if (words.length === 0)
+        return undefined;
+    const name = words.join(" ");
+    return words.length === 1 && SELF.has(name) ? undefined : name;
+};
+//# sourceMappingURL=citation.js.map

package/dist/citation.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"citation.js","sourceRoot":"","sources":["../src/citation.ts"],"names":[],"mappings":"AAAA,2GAA2G;AAE3G,MAAM,EAAE,GAAG,mBAAmB,CAAC;AAC/B,MAAM,WAAW,GAAG,iBAAiB,CAAC;AACtC;;;GAGG;AACH,MAAM,IAAI,GAAG,IAAI,GAAG,CAAC,CAAC,WAAW,EAAE,UAAU,EAAE,OAAO,EAAE,QAAQ,CAAC,CAAC,CAAC;AACnE,wEAAwE;AACxE,MAAM,OAAO,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC,CAAC;AACpD,MAAM,SAAS,GAAG,CAAC,CAAC;AAEpB,qHAAqH;AACrH,MAAM,UAAU,GAAG,CAAC,IAAY,EAAY,EAAE;IAC5C,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IAC/B,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,MAAM,CAAC,MAAM,IAAI,KAAK,CAAC,MAAM,GAAG,SAAS,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QAClF,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC;QAClC,MAAM,IAAI,GAAG,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;QAC1C,MAAM,KAAK,GAAG,IAAI,KAAK,SAAS,IAAI,KAAK,CAAC,MAAM,GAAG,CAAC,IAAI,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,WAAW,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;QACxH,IAAI,IAAI,KAAK,SAAS,IAAI,CAAC,KAAK;YAAE,MAAM;QACxC,KAAK,CAAC,IAAI,CAAC,IAAI,IAAI,KAAK,CAAC,CAAC;QAC1B,IAAI,IAAI,KAAK,SAAS,IAAI,IAAI,CAAC,MAAM,GAAG,KAAK,CAAC,MAAM;YAAE,MAAM;IAC9D,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,CAAC,MAAM,kBAAkB,GAAG,CAAC,IAAY,EAAE,GAAW,EAAsB,EAAE;IAClF,MAAM,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,EAAE,GAAG,GAAG,GAAG,CAAC,CAAC;IACxC,MAAM,EAAE,GAAG,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IACzB,IAAI,EAAE,KAAK,IAAI;QAAE,OAAO,SAAS,CAAC;IAClC,MAAM,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC;IACnD,IAAI,KAAK,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC;IACzC,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAC7B,OAAO,KAAK,CAAC,MAAM,KAAK,CAAC,IAAI,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,IAAI,CAAC;AACjE,CAAC,CAAC"}

package/dist/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAGA,OAAO,KAAK,EAAgB,eAAe,EAA0B,MAAM,gBAAgB,CAAC;AAoB5F;;;GAGG;AACH,eAAO,MAAM,OAAO,EAAE,eA6BrB,CAAC"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAIA,OAAO,KAAK,EAAgB,eAAe,EAA0B,MAAM,gBAAgB,CAAC;AAoB5F;;;GAGG;AACH,eAAO,MAAM,OAAO,EAAE,eA8BrB,CAAC"}

package/dist/index.js

@@ -1,5 +1,6 @@
 import { split, SentenceSplitterSyntax } from "sentence-splitter";
 import { loadLexicons } from "./lexicons.js";
+import { structure } from "./structure.js";
 import { isReady, prepare, tokenize } from "./pos.js";
 // chaff からは型だけを取る。実行時の値依存を作らない。アダプタは単体で動く。
 const LATIN_LETTER = /[a-z]/giu;
@@ -46,6 +47,7 @@ export const adapter = {
         return [...source.matchAll(LATIN_LETTER)].length / total;
     },
     lexicons: loadLexicons(),
+    structure,
     segment: (text) => {
         const sentences = split(text)
             .filter((node) => node.type === SentenceSplitterSyntax.Sentence)

package/dist/index.js.map

@@ -1 +1 @@
-{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,EAAE,sBAAsB,EAAE,MAAM,mBAAmB,CAAC;AAClE,OAAO,EAAE,YAAY,EAAE,MAAM,eAAe,CAAC;AAC7C,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,QAAQ,EAAE,MAAM,UAAU,CAAC;AAGtD,2CAA2C;AAE3C,MAAM,YAAY,GAAG,UAAU,CAAC;AAChC,MAAM,SAAS,GAAG,MAAM,CAAC;AAEzB;;;GAGG;AACH,MAAM,UAAU,GAAG,CAAC,QAAkB,EAAY,EAAE;IAClD,MAAM,MAAM,GAAG,QAAQ,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;IACvC,IAAI,MAAM,KAAK,SAAS;QAAE,OAAO,QAAQ,CAAC;IAC1C,OAAO;QACL,GAAG,QAAQ;QACX,MAAM,EAAE,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,GAAG,KAAK,EAAE,IAAI,EAAE,EAAE,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,GAAG,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;KAClJ,CAAC;AACJ,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,CAAC,MAAM,OAAO,GAAoB;IACtC,IAAI,EAAE,UAAU;IAChB,EAAE,EAAE,IAAI;IACR,UAAU,EAAE,CAAC;IACb,YAAY,EAAE;QACZ,aAAa,EAAE,IAAI;QACnB,SAAS,EAAE,IAAI;QACf,6CAA6C;QAC7C,GAAG,EAAE,IAAI;QACT,KAAK,EAAE,IAAI;QACX,UAAU,EAAE,MAAM;KACnB;IACD,OAAO,EAAE,CAAC,IAAkB,EAAiB,EAAE;QAC7C,0CAA0C;QAC1C,IAAI,IAAI,CAAC,GAAG;YAAE,OAAO,EAAE,CAAC;QACxB,OAAO,OAAO,CAAC,OAAO,EAAE,CAAC;IAC3B,CAAC;IACD,MAAM,EAAE,CAAC,MAAc,EAAU,EAAE;QACjC,MAAM,KAAK,GAAG,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC;QACrD,IAAI,KAAK,KAAK,CAAC;YAAE,OAAO,CAAC,CAAC;QAC1B,OAAO,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,YAAY,CAAC,CAAC,CAAC,MAAM,GAAG,KAAK,CAAC;IAC3D,CAAC;IACD,QAAQ,EAAE,YAAY,EAAE;IACxB,OAAO,EAAE,CAAC,IAAY,EAAgB,EAAE;QACtC,MAAM,SAAS,GAAe,KAAK,CAAC,IAAI,CAAC;aACtC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,KAAK,sBAAsB,CAAC,QAAQ,CAAC;aAC/D,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,EAAE,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;QAC3H,OAAO,EAAE,SAAS,EAAE,OAAO,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,SAAS,EAAE,CAAC;IAC1E,CAAC;CACF,CAAC"}
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../src/index.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,EAAE,sBAAsB,EAAE,MAAM,mBAAmB,CAAC;AAClE,OAAO,EAAE,YAAY,EAAE,MAAM,eAAe,CAAC;AAC7C,OAAO,EAAE,SAAS,EAAE,MAAM,gBAAgB,CAAC;AAC3C,OAAO,EAAE,OAAO,EAAE,OAAO,EAAE,QAAQ,EAAE,MAAM,UAAU,CAAC;AAGtD,2CAA2C;AAE3C,MAAM,YAAY,GAAG,UAAU,CAAC;AAChC,MAAM,SAAS,GAAG,MAAM,CAAC;AAEzB;;;GAGG;AACH,MAAM,UAAU,GAAG,CAAC,QAAkB,EAAY,EAAE;IAClD,MAAM,MAAM,GAAG,QAAQ,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;IACvC,IAAI,MAAM,KAAK,SAAS;QAAE,OAAO,QAAQ,CAAC;IAC1C,OAAO;QACL,GAAG,QAAQ;QACX,MAAM,EAAE,MAAM,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,GAAG,KAAK,EAAE,IAAI,EAAE,EAAE,KAAK,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,KAAK,EAAE,GAAG,EAAE,QAAQ,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,IAAI,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;KAClJ,CAAC;AACJ,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,CAAC,MAAM,OAAO,GAAoB;IACtC,IAAI,EAAE,UAAU;IAChB,EAAE,EAAE,IAAI;IACR,UAAU,EAAE,CAAC;IACb,YAAY,EAAE;QACZ,aAAa,EAAE,IAAI;QACnB,SAAS,EAAE,IAAI;QACf,6CAA6C;QAC7C,GAAG,EAAE,IAAI;QACT,KAAK,EAAE,IAAI;QACX,UAAU,EAAE,MAAM;KACnB;IACD,OAAO,EAAE,CAAC,IAAkB,EAAiB,EAAE;QAC7C,0CAA0C;QAC1C,IAAI,IAAI,CAAC,GAAG;YAAE,OAAO,EAAE,CAAC;QACxB,OAAO,OAAO,CAAC,OAAO,EAAE,CAAC;IAC3B,CAAC;IACD,MAAM,EAAE,CAAC,MAAc,EAAU,EAAE;QACjC,MAAM,KAAK,GAAG,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC;QACrD,IAAI,KAAK,KAAK,CAAC;YAAE,OAAO,CAAC,CAAC;QAC1B,OAAO,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,YAAY,CAAC,CAAC,CAAC,MAAM,GAAG,KAAK,CAAC;IAC3D,CAAC;IACD,QAAQ,EAAE,YAAY,EAAE;IACxB,SAAS;IACT,OAAO,EAAE,CAAC,IAAY,EAAgB,EAAE;QACtC,MAAM,SAAS,GAAe,KAAK,CAAC,IAAI,CAAC;aACtC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,KAAK,sBAAsB,CAAC,QAAQ,CAAC;aAC/D,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,EAAE,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC;QAC3H,OAAO,EAAE,SAAS,EAAE,OAAO,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,SAAS,EAAE,CAAC;IAC1E,CAAC;CACF,CAAC"}

package/dist/lexicons.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"lexicons.d.ts","sourceRoot":"","sources":["../src/lexicons.ts"],"names":[],"mappings":"AAIA,OAAO,KAAK,EAAE,OAAO,EAAgB,MAAM,gBAAgB,CAAC;AAY5D;;;GAGG;AACH,eAAO,MAAM,YAAY,GAAI,MAAK,MAAY,KAAG,MAAM,CAAC,MAAM,EAAE,OAAO,CAQ7D,CAAC"}
+{"version":3,"file":"lexicons.d.ts","sourceRoot":"","sources":["../src/lexicons.ts"],"names":[],"mappings":"AAIA,OAAO,KAAK,EAAE,OAAO,EAAgB,MAAM,gBAAgB,CAAC;AAiB5D;;;GAGG;AACH,eAAO,MAAM,YAAY,GAAI,MAAK,MAAY,KAAG,MAAM,CAAC,MAAM,EAAE,OAAO,CAQ7D,CAAC"}

package/dist/lexicons.js

@@ -8,7 +8,12 @@ const toEntry = (raw) => {
     if (!isRecord(raw) || typeof raw["pattern"] !== "string")
         return undefined;
     const weight = raw["weight"];
-    return { pattern: raw["pattern"], weight: typeof weight === "number" ? weight : undefined };
+    const instead = raw["instead_of"];
+    return {
+        pattern: raw["pattern"],
+        weight: typeof weight === "number" ? weight : undefined,
+        instead_of: typeof instead === "string" ? instead : undefined,
+    };
 };
 /**
  * 語彙表はアダプタが持つ。detector は共通で、これだけが言語別。spec §11。

package/dist/lexicons.js.map

@@ -1 +1 @@
-{"version":3,"file":"lexicons.js","sourceRoot":"","sources":["../src/lexicons.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,WAAW,EAAE,MAAM,SAAS,CAAC;AACpD,OAAO,EAAE,aAAa,EAAE,MAAM,UAAU,CAAC;AACzC,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAC1C,OAAO,EAAE,KAAK,EAAE,MAAM,MAAM,CAAC;AAG7B,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,EAAE,UAAU,CAAC,CAAC;AAE5E,MAAM,QAAQ,GAAG,CAAC,KAAc,EAAoC,EAAE,CAAC,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,KAAK,IAAI,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC;AAE5I,MAAM,OAAO,GAAG,CAAC,GAAY,EAA4B,EAAE;IACzD,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,OAAO,GAAG,CAAC,SAAS,CAAC,KAAK,QAAQ;QAAE,OAAO,SAAS,CAAC;IAC3E,MAAM,MAAM,GAAG,GAAG,CAAC,QAAQ,CAAC,CAAC;IAC7B,OAAO,EAAE,OAAO,EAAE,GAAG,CAAC,SAAS,CAAC,EAAE,MAAM,EAAE,OAAO,MAAM,KAAK,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,SAAS,EAAE,CAAC;AAC9F,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,MAAc,GAAG,EAA2B,EAAE,CACzE,WAAW,CAAC,GAAG,CAAC;KACb,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC;KACxC,MAAM,CAA0B,CAAC,GAAG,EAAE,IAAI,EAAE,EAAE;IAC7C,MAAM,GAAG,GAAY,KAAK,CAAC,YAAY,CAAC,IAAI,CAAC,GAAG,EAAE,IAAI,CAAC,EAAE,MAAM,CAAC,CAAC,CAAC;IAClE,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,OAAO,GAAG,CAAC,IAAI,CAAC,KAAK,QAAQ,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;QAAE,OAAO,GAAG,CAAC;IAClG,MAAM,OAAO,GAAG,GAAG,CAAC,SAAS,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC;IACnF,OAAO,EAAE,GAAG,GAAG,EAAE,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,EAAE,OAAO,EAAE,CAAC;AAC1C,CAAC,EAAE,EAAE,CAAC,CAAC"}
+{"version":3,"file":"lexicons.js","sourceRoot":"","sources":["../src/lexicons.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,WAAW,EAAE,MAAM,SAAS,CAAC;AACpD,OAAO,EAAE,aAAa,EAAE,MAAM,UAAU,CAAC;AACzC,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAC1C,OAAO,EAAE,KAAK,EAAE,MAAM,MAAM,CAAC;AAG7B,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,EAAE,UAAU,CAAC,CAAC;AAE5E,MAAM,QAAQ,GAAG,CAAC,KAAc,EAAoC,EAAE,CAAC,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,KAAK,IAAI,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC;AAE5I,MAAM,OAAO,GAAG,CAAC,GAAY,EAA4B,EAAE;IACzD,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,OAAO,GAAG,CAAC,SAAS,CAAC,KAAK,QAAQ;QAAE,OAAO,SAAS,CAAC;IAC3E,MAAM,MAAM,GAAG,GAAG,CAAC,QAAQ,CAAC,CAAC;IAC7B,MAAM,OAAO,GAAG,GAAG,CAAC,YAAY,CAAC,CAAC;IAClC,OAAO;QACL,OAAO,EAAE,GAAG,CAAC,SAAS,CAAC;QACvB,MAAM,EAAE,OAAO,MAAM,KAAK,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,SAAS;QACvD,UAAU,EAAE,OAAO,OAAO,KAAK,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,SAAS;KAC9D,CAAC;AACJ,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,MAAc,GAAG,EAA2B,EAAE,CACzE,WAAW,CAAC,GAAG,CAAC;KACb,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC;KACxC,MAAM,CAA0B,CAAC,GAAG,EAAE,IAAI,EAAE,EAAE;IAC7C,MAAM,GAAG,GAAY,KAAK,CAAC,YAAY,CAAC,IAAI,CAAC,GAAG,EAAE,IAAI,CAAC,EAAE,MAAM,CAAC,CAAC,CAAC;IAClE,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,IAAI,OAAO,GAAG,CAAC,IAAI,CAAC,KAAK,QAAQ,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;QAAE,OAAO,GAAG,CAAC;IAClG,MAAM,OAAO,GAAG,GAAG,CAAC,SAAS,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC;IACnF,OAAO,EAAE,GAAG,GAAG,EAAE,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC,EAAE,OAAO,EAAE,CAAC;AAC1C,CAAC,EAAE,EAAE,CAAC,CAAC"}

package/dist/pos.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"pos.d.ts","sourceRoot":"","sources":["../src/pos.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,KAAK,EAAE,MAAM,gBAAgB,CAAC;AA4E5C,eAAO,MAAM,IAAI,GAAI,KAAK,MAAM,KAAG,MAAiE,CAAC;AAqBrG,eAAO,MAAM,OAAO,QAAO,IAE1B,CAAC;AAEF,eAAO,MAAM,OAAO,QAAO,OAAoC,CAAC;AAwBhE,eAAO,MAAM,QAAQ,GAAI,MAAM,MAAM,KAAG,KAAK,EAAE,GAAG,SAIjD,CAAC"}
+{"version":3,"file":"pos.d.ts","sourceRoot":"","sources":["../src/pos.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,EAAE,KAAK,EAAE,MAAM,gBAAgB,CAAC;AA4E5C,eAAO,MAAM,IAAI,GAAI,KAAK,MAAM,KAAG,MAAiE,CAAC;AA4BrG,eAAO,MAAM,OAAO,QAAO,IAE1B,CAAC;AAEF,eAAO,MAAM,OAAO,QAAO,OAAoC,CAAC;AAwBhE,eAAO,MAAM,QAAQ,GAAI,MAAM,MAAM,KAAG,KAAK,EAAE,GAAG,SAIjD,CAAC"}

package/dist/pos.js

@@ -67,14 +67,21 @@ export const upos = (tag) => (PUNCTUATION.has(tag) ? "PUNCT" : (BY_TAG[tag] ?? "
  * 受動は be + 過去分詞。過去分詞だけでは完了形（has reviewed）と見分けられないので、
  * 直前の be を見る。間に副詞が挟まる（was quickly approved）ぶんだけ遡る。
  */
-const ADVERB_TAGS = new Set(["RB", "RBR", "RBS"]);
+/**
+ * be と過去分詞の間に立てるもの。副詞と、副詞をつなぐ接続詞（was fully and carefully reviewed）。
+ *
+ * これでも拾えない形がある。`was fully and finally approved` では wink が `approved` を
+ * VBN ではなく **VBD** と付けるため、そもそも過去分詞として見えない。解析器の限界で、
+ * VBD も受動と見なすと `The team was here and approved it` まで受動になる。直さない。
+ */
+const SKIPPABLE = new Set(["RB", "RBR", "RBS", "CC", ","]);
 const BE = new Set(["be", "am", "is", "are", "was", "were", "been", "being"]);
 const isBe = (entry) => BE.has(entry.lemma ?? entry.value.toLowerCase()) || BE.has(entry.value.toLowerCase());
 const isPassive = (tagged, at) => {
     if (tagged[at]?.pos !== "VBN")
         return false;
     const before = tagged.slice(0, at).reverse();
-    const head = before.find((entry) => !ADVERB_TAGS.has(entry.pos));
+    const head = before.find((entry) => !SKIPPABLE.has(entry.pos));
     return head !== undefined && isBe(head);
 };
 const state = { ready: undefined };

package/dist/pos.js.map

@@ -1 +1 @@
-{"version":3,"file":"pos.js","sourceRoot":"","sources":["../src/pos.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,aAAa,EAAE,MAAM,aAAa,CAAC;AAG5C,MAAM,OAAO,GAAG,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AAO/C,MAAM,QAAQ,GAAG,CAAC,KAAc,EAAoC,EAAE,CAAC,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,KAAK,IAAI,CAAC;AAEnH,MAAM,UAAU,GAAG,CAAC,KAAc,EAAqD,EAAE,CAAC,OAAO,KAAK,KAAK,UAAU,CAAC;AAEtH,MAAM,QAAQ,GAAG,CAAC,KAAc,EAAmB,EAAE,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,OAAO,KAAK,CAAC,OAAO,CAAC,KAAK,QAAQ,IAAI,OAAO,KAAK,CAAC,KAAK,CAAC,KAAK,QAAQ,CAAC;AAEhJ,MAAM,OAAO,GAAG,CAAC,KAAc,EAAsB,EAAE,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;AAE5F,kEAAkE;AAClE,MAAM,UAAU,GAAG,CAAC,KAA8B,EAAE,IAAY,EAAE,IAAwB,EAAW,EAAE;IACrG,MAAM,MAAM,GAAY,KAAK,CAAC,IAAI,CAAC,CAAC;IACpC,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,qBAAqB,IAAI,WAAW,CAAC,CAAC;IAC/E,OAAO,OAAO,CAAC,KAAK,CAAC,MAAM,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC;AAC5C,CAAC,CAAC;AAEF,MAAM,KAAK,GAAG,GAAW,EAAE;IACzB,MAAM,OAAO,GAAY,OAAO,CAAC,iBAAiB,CAAC,CAAC;IACpD,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,oCAAoC,CAAC,CAAC;IAChF,MAAM,MAAM,GAAY,OAAO,EAAE,CAAC;IAClC,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,oCAAoC,CAAC,CAAC;IAC7E,OAAO,MAAM,CAAC;AAChB,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,MAAM,GAAqC;IAC/C,EAAE,EAAE,OAAO;IACX,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,MAAM;IACV,EAAE,EAAE,GAAG;IACP,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,KAAK;IACT,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,MAAM;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,OAAO;IACZ,IAAI,EAAE,OAAO;IACb,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,IAAI,EAAE,MAAM;IACZ,EAAE,EAAE,KAAK;IACT,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,KAAK;IACT,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,MAAM;IACV,EAAE,EAAE,MAAM;IACV,EAAE,EAAE,MAAM;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,MAAM;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,KAAK;CACX,CAAC;AAEF,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAE7E,MAAM,CAAC,MAAM,IAAI,GAAG,CAAC,GAAW,EAAU,EAAE,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC;AAErG;;;GAGG;AACH,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,KAAK,CAAC,CAAC,CAAC;AAElD,MAAM,EAAE,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,CAAC,CAAC,CAAC;AAE9E,MAAM,IAAI,GAAG,CAAC,KAAa,EAAW,EAAE,CAAC,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,IAAI,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,IAAI,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC;AAE/H,MAAM,SAAS,GAAG,CAAC,MAAyB,EAAE,EAAU,EAAW,EAAE;IACnE,IAAI,MAAM,CAAC,EAAE,CAAC,EAAE,GAAG,KAAK,KAAK;QAAE,OAAO,KAAK,CAAC;IAC5C,MAAM,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC;IAC7C,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC;IACjE,OAAO,IAAI,KAAK,SAAS,IAAI,IAAI,CAAC,IAAI,CAAC,CAAC;AAC1C,CAAC,CAAC;AAEF,MAAM,KAAK,GAAkC,EAAE,KAAK,EAAE,SAAS,EAAE,CAAC;AAElE,MAAM,CAAC,MAAM,OAAO,GAAG,GAAS,EAAE;IAChC,KAAK,CAAC,KAAK,KAAK,KAAK,EAAE,CAAC;AAC1B,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,OAAO,GAAG,GAAY,EAAE,CAAC,KAAK,CAAC,KAAK,KAAK,SAAS,CAAC;AAEhE;;;GAGG;AACH,MAAM,MAAM,GAAG,CAAC,IAAY,EAAE,MAAyB,EAAW,EAAE,CAClE,MAAM,CAAC,MAAM,CACX,CAAC,GAAG,EAAE,KAAK,EAAE,EAAE,EAAE,EAAE;IACjB,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;IACpD,IAAI,KAAK,KAAK,CAAC,CAAC;QAAE,OAAO,GAAG,CAAC;IAC7B,MAAM,GAAG,GAAG,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,MAAM,CAAC;IACvC,MAAM,KAAK,GAAG;QACZ,IAAI,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE;QACpB,OAAO,EAAE,KAAK,CAAC,KAAK;QACpB,GAAG,EAAE,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC;QACpB,GAAG,CAAC,KAAK,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,CAAC;QAC5D,GAAG,CAAC,SAAS,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KAClE,CAAC;IACF,OAAO,EAAE,MAAM,EAAE,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,KAAK,CAAC,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC;AACzD,CAAC,EACD,EAAE,MAAM,EAAE,EAAE,EAAE,MAAM,EAAE,CAAC,EAAE,CAC1B,CAAC,MAAM,CAAC;AAEX,MAAM,CAAC,MAAM,QAAQ,GAAG,CAAC,IAAY,EAAuB,EAAE;IAC5D,MAAM,MAAM,GAAG,KAAK,CAAC,KAAK,CAAC;IAC3B,IAAI,MAAM,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAC3C,OAAO,MAAM,CAAC,IAAI,EAAE,OAAO,CAAC,UAAU,CAAC,MAAM,EAAE,aAAa,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC;AAC3F,CAAC,CAAC"}
+{"version":3,"file":"pos.js","sourceRoot":"","sources":["../src/pos.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,aAAa,EAAE,MAAM,aAAa,CAAC;AAG5C,MAAM,OAAO,GAAG,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;AAO/C,MAAM,QAAQ,GAAG,CAAC,KAAc,EAAoC,EAAE,CAAC,OAAO,KAAK,KAAK,QAAQ,IAAI,KAAK,KAAK,IAAI,CAAC;AAEnH,MAAM,UAAU,GAAG,CAAC,KAAc,EAAqD,EAAE,CAAC,OAAO,KAAK,KAAK,UAAU,CAAC;AAEtH,MAAM,QAAQ,GAAG,CAAC,KAAc,EAAmB,EAAE,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,OAAO,KAAK,CAAC,OAAO,CAAC,KAAK,QAAQ,IAAI,OAAO,KAAK,CAAC,KAAK,CAAC,KAAK,QAAQ,CAAC;AAEhJ,MAAM,OAAO,GAAG,CAAC,KAAc,EAAsB,EAAE,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;AAE5F,kEAAkE;AAClE,MAAM,UAAU,GAAG,CAAC,KAA8B,EAAE,IAAY,EAAE,IAAwB,EAAW,EAAE;IACrG,MAAM,MAAM,GAAY,KAAK,CAAC,IAAI,CAAC,CAAC;IACpC,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,qBAAqB,IAAI,WAAW,CAAC,CAAC;IAC/E,OAAO,OAAO,CAAC,KAAK,CAAC,MAAM,EAAE,KAAK,EAAE,IAAI,CAAC,CAAC;AAC5C,CAAC,CAAC;AAEF,MAAM,KAAK,GAAG,GAAW,EAAE;IACzB,MAAM,OAAO,GAAY,OAAO,CAAC,iBAAiB,CAAC,CAAC;IACpD,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,oCAAoC,CAAC,CAAC;IAChF,MAAM,MAAM,GAAY,OAAO,EAAE,CAAC;IAClC,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC;QAAE,MAAM,IAAI,KAAK,CAAC,oCAAoC,CAAC,CAAC;IAC7E,OAAO,MAAM,CAAC;AAChB,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,MAAM,GAAqC;IAC/C,EAAE,EAAE,OAAO;IACX,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,MAAM;IACV,EAAE,EAAE,GAAG;IACP,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,KAAK;IACT,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,KAAK;IACT,EAAE,EAAE,MAAM;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,OAAO;IACZ,IAAI,EAAE,OAAO;IACb,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,IAAI,EAAE,MAAM;IACZ,EAAE,EAAE,KAAK;IACT,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,KAAK;IACT,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,MAAM;IACV,EAAE,EAAE,MAAM;IACV,EAAE,EAAE,MAAM;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,KAAK;IACV,EAAE,EAAE,MAAM;IACV,GAAG,EAAE,MAAM;IACX,GAAG,EAAE,KAAK;CACX,CAAC;AAEF,MAAM,WAAW,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAE7E,MAAM,CAAC,MAAM,IAAI,GAAG,CAAC,GAAW,EAAU,EAAE,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC;AAErG;;;GAGG;AACH;;;;;;GAMG;AACH,MAAM,SAAS,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,IAAI,EAAE,GAAG,CAAC,CAAC,CAAC;AAE3D,MAAM,EAAE,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,CAAC,CAAC,CAAC;AAE9E,MAAM,IAAI,GAAG,CAAC,KAAa,EAAW,EAAE,CAAC,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,IAAI,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,IAAI,EAAE,CAAC,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE,CAAC,CAAC;AAE/H,MAAM,SAAS,GAAG,CAAC,MAAyB,EAAE,EAAU,EAAW,EAAE;IACnE,IAAI,MAAM,CAAC,EAAE,CAAC,EAAE,GAAG,KAAK,KAAK;QAAE,OAAO,KAAK,CAAC;IAC5C,MAAM,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC;IAC7C,MAAM,IAAI,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,SAAS,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC;IAC/D,OAAO,IAAI,KAAK,SAAS,IAAI,IAAI,CAAC,IAAI,CAAC,CAAC;AAC1C,CAAC,CAAC;AAEF,MAAM,KAAK,GAAkC,EAAE,KAAK,EAAE,SAAS,EAAE,CAAC;AAElE,MAAM,CAAC,MAAM,OAAO,GAAG,GAAS,EAAE;IAChC,KAAK,CAAC,KAAK,KAAK,KAAK,EAAE,CAAC;AAC1B,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,OAAO,GAAG,GAAY,EAAE,CAAC,KAAK,CAAC,KAAK,KAAK,SAAS,CAAC;AAEhE;;;GAGG;AACH,MAAM,MAAM,GAAG,CAAC,IAAY,EAAE,MAAyB,EAAW,EAAE,CAClE,MAAM,CAAC,MAAM,CACX,CAAC,GAAG,EAAE,KAAK,EAAE,EAAE,EAAE,EAAE;IACjB,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;IACpD,IAAI,KAAK,KAAK,CAAC,CAAC;QAAE,OAAO,GAAG,CAAC;IAC7B,MAAM,GAAG,GAAG,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,MAAM,CAAC;IACvC,MAAM,KAAK,GAAG;QACZ,IAAI,EAAE,EAAE,KAAK,EAAE,GAAG,EAAE;QACpB,OAAO,EAAE,KAAK,CAAC,KAAK;QACpB,GAAG,EAAE,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC;QACpB,GAAG,CAAC,KAAK,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,CAAC;QAC5D,GAAG,CAAC,SAAS,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,EAAE,KAAK,EAAE,MAAM,EAAE,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KAClE,CAAC;IACF,OAAO,EAAE,MAAM,EAAE,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,KAAK,CAAC,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC;AACzD,CAAC,EACD,EAAE,MAAM,EAAE,EAAE,EAAE,MAAM,EAAE,CAAC,EAAE,CAC1B,CAAC,MAAM,CAAC;AAEX,MAAM,CAAC,MAAM,QAAQ,GAAG,CAAC,IAAY,EAAuB,EAAE;IAC5D,MAAM,MAAM,GAAG,KAAK,CAAC,KAAK,CAAC;IAC3B,IAAI,MAAM,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAC3C,OAAO,MAAM,CAAC,IAAI,EAAE,OAAO,CAAC,UAAU,CAAC,MAAM,EAAE,aAAa,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC;AAC3F,CAAC,CAAC"}

package/dist/structure.d.ts

@@ -0,0 +1,5 @@
+import type { StructurePatterns } from "chaffjs/plugin";
+/** "IV" → 4, "xii" → 12. Undefined for anything that is not a roman numeral. */
+export declare const parseRoman: (text: string) => number | undefined;
+export declare const structure: StructurePatterns;
+//# sourceMappingURL=structure.d.ts.map

package/dist/structure.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"structure.d.ts","sourceRoot":"","sources":["../src/structure.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAA2C,iBAAiB,EAAE,MAAM,gBAAgB,CAAC;AAQjG,gFAAgF;AAChF,eAAO,MAAM,UAAU,GAAI,MAAM,MAAM,KAAG,MAAM,GAAG,SAKlD,CAAC;AA+VF,eAAO,MAAM,SAAS,EAAE,iBAAuG,CAAC"}

package/dist/structure.js

@@ -0,0 +1,329 @@
+import { citedDocumentAfter } from "./citation.js";
+// Contracts, specifications and statutes in English. core nests what this reads; it does not know
+// how English numbers its articles.
+const ROMAN = { i: 1, v: 5, x: 10, l: 50, c: 100 };
+/** "IV" → 4, "xii" → 12. Undefined for anything that is not a roman numeral. */
+export const parseRoman = (text) => {
+    const values = [...text.toLowerCase()].map((char) => ROMAN[char]);
+    if (values.length === 0 || values.some((value) => value === undefined))
+        return undefined;
+    const known = values.filter((value) => value !== undefined);
+    return known.reduce((total, value, index) => ((known[index + 1] ?? 0) > value ? total - value : total + value), 0);
+};
+const numberOf = (text) => {
+    if (text === undefined)
+        return undefined;
+    if (/^\d{1,3}(?:\.\d{1,3}){0,5}$/u.test(text))
+        return text;
+    const roman = parseRoman(text);
+    return roman === undefined ? undefined : String(roman);
+};
+/**
+ * A heading is the number followed by nothing, punctuation, or a capitalised title.
+ * "Article 3 shall apply to…" at the start of a line is a sentence, not the heading of Article 3.
+ */
+const titleOf = (rest) => {
+    const trimmed = rest.replace(/^[.:\-–—\s]+/u, "").trim();
+    if (trimmed !== "" && /^\p{Ll}/u.test(trimmed))
+        return undefined;
+    return trimmed;
+};
+const ARTICLE = /^\s{0,3}(?:ARTICLE|Article)\s+(?<n>\d{1,3}|[IVXLC]{1,7})\b(?<rest>.*)$/u;
+const SECTION = /^\s{0,3}(?:SECTION|Section|§)\s*(?<n>\d{1,3}(?:\.\d{1,3}){0,5})\b(?<rest>.*)$/u;
+const LETTERED = /^\s{0,6}\((?<n>[a-z]{1,4}|\d{1,3})\)\s+(?<rest>\S.*)$/u;
+const MULTI_ROMAN = /^(?:ii|iii|iv|vi|vii|viii|ix)$/u;
+const headed = (pattern, line, label) => {
+    const groups = pattern.exec(line)?.groups;
+    const number = numberOf(groups?.["n"]);
+    const heading = titleOf(groups?.["rest"] ?? "");
+    if (groups === undefined || number === undefined || heading === undefined)
+        return undefined;
+    const parts = number.split(".");
+    return {
+        kind: "article",
+        depth: parts.length,
+        number,
+        absolute: true,
+        label: label(groups["n"] ?? ""),
+        heading,
+        rest: heading,
+        ordinal: Number(parts.at(-1)),
+    };
+};
+/**
+ * 開いている項目の書き方。"(ii)" はローマ数字、"(b)" は英字。一文字の "(i)" はどちらにも読めるので、
+ * 開いたときの読みを覚えておく代わりに、一つ上に英字が開いていたかで決め直す。
+ */
+const styleOfLabel = (open, index, all) => {
+    const inner = /^\((?<n>[a-z0-9]{1,4})\)$/u.exec(open.label)?.groups?.["n"];
+    if (inner === undefined)
+        return undefined;
+    if (/^\d+$/u.test(inner))
+        return "digit";
+    if (MULTI_ROMAN.test(inner))
+        return "roman";
+    const above = all[index - 1];
+    return /^[ivx]$/u.test(inner) && above !== undefined && styleOfLabel(above, index - 1, all) === "letter" && above.depth < open.depth ? "roman" : "letter";
+};
+const styles = (context) => context.open.map((open, index, all) => styleOfLabel(open, index, all));
+/** "(h)" の次の "(i)" は英字。開いている英字の次の文字なら、ローマ数字とは読まない。 */
+const followsLetter = (raw, context) => context.open.some((open, index) => styles(context)[index] === "letter" && open.number.charCodeAt(0) + 1 === raw.charCodeAt(0));
+/** "(i)" is a roman numeral right under "(a)", or when a roman list is already open; the letter i otherwise. */
+const styleOf = (raw, context) => {
+    if (/^\d+$/u.test(raw))
+        return "digit";
+    if (MULTI_ROMAN.test(raw))
+        return "roman";
+    if (!/^[ivx]$/u.test(raw) || followsLetter(raw, context))
+        return "letter";
+    const open = styles(context);
+    return open.includes("roman") || open.at(-1) === "letter" ? "roman" : "letter";
+};
+/**
+ * A sibling has the depth of the open item written the same way: "(b)" closes "(i)" and sits beside "(a)".
+ * A new way of numbering goes one deeper than whatever is open.
+ */
+const depthFor = (style, context) => {
+    const open = styles(context);
+    const sibling = [...context.open].reverse().find((_item, reversed) => open[context.open.length - 1 - reversed] === style);
+    return sibling?.depth ?? (context.open.at(-1)?.depth ?? 0) + 1;
+};
+const LETTER_BEFORE_A = "a".charCodeAt(0) - 1;
+/** "(b)" は 2 番目、"(ii)" も 2 番目。二文字以上の英字（"(aa)"）は並びが決まらないので付けない。 */
+const ordinalOf = (raw, style) => {
+    if (style === "digit")
+        return Number(raw);
+    if (style === "roman")
+        return parseRoman(raw);
+    return raw.length === 1 ? raw.charCodeAt(0) - LETTER_BEFORE_A : undefined;
+};
+const lettered = (line, context) => {
+    const groups = LETTERED.exec(line)?.groups;
+    const raw = groups?.["n"];
+    if (groups === undefined || raw === undefined)
+        return undefined;
+    const style = styleOf(raw, context);
+    // 番地は書かれたままの "ii" を使う。参照「Section 4.2(a)(ii)」も同じ形で書かれるので、そのまま引ける。
+    return {
+        kind: "item",
+        depth: depthFor(style, context),
+        ordinal: ordinalOf(raw, style),
+        number: raw,
+        absolute: false,
+        label: `(${raw})`,
+        heading: "",
+        rest: groups["rest"]?.trim() ?? "",
+    };
+};
+/** Chapters restart in each part, so a chapter continues its part's address: PART II, CHAPTER 1 → pt2.ch1. */
+const PART = /^\s{0,3}(?:PART|Part)\s+(?<n>\d{1,3}|[IVXLC]{1,7})\b(?<rest>.*)$/u;
+const CHAPTER = /^\s{0,3}(?:CHAPTER|Chapter)\s+(?<n>\d{1,3}|[IVXLC]{1,7})\b(?<rest>.*)$/u;
+const chapter = (pattern, line, depth, prefix, word) => {
+    const groups = pattern.exec(line)?.groups;
+    const number = numberOf(groups?.["n"]);
+    const heading = titleOf(groups?.["rest"] ?? "");
+    if (groups === undefined || number === undefined || heading === undefined)
+        return undefined;
+    const label = `${word} ${groups["n"] ?? ""}`;
+    return { kind: "chapter", depth, number: prefix + number, absolute: false, label, heading, rest: heading, ordinal: Number(number) };
+};
+const numbered = (line, context) => chapter(PART, line, -2, "pt", "Part") ??
+    chapter(CHAPTER, line, -1, "ch", "Chapter") ??
+    headed(ARTICLE, line, (n) => `Article ${n}`) ??
+    headed(SECTION, line, (n) => `Section ${n}`) ??
+    lettered(line, context);
+const mentions = (pattern, text, attrs) => [...text.matchAll(pattern)].flatMap((match) => {
+    const found = attrs(match.groups ?? {}, match[0]);
+    return found === undefined ? [] : [{ start: match.index, end: match.index + match[0].length, attrs: found }];
+});
+const DEFINITIONS = [
+    /["“](?<term>[^"”\n]{1,60})["”] (?:means|shall mean|refers to|has the meaning)\b/gu,
+    /\((?:the |hereinafter )?["“](?<term>[^"”\n]{1,60})["”]\)/gu,
+    /\(hereinafter referred to as ["“](?<term>[^"”\n]{1,60})["”]\)/gu,
+];
+const definitions = (text) => DEFINITIONS.flatMap((pattern) => mentions(pattern, text, (groups) => (groups["term"] === undefined ? undefined : { term: groups["term"] })));
+const REFERENCE = /\b(?:Sections?|Articles?|§) ?(?<n>\d{1,3}(?:\.\d{1,3}){0,5}|[IVXLC]{1,7})\b/gu;
+const SUBDIVISION = /^\((?<p>[a-z0-9]{1,4})\)/u;
+/** "(a)(ii)(3)" is as deep as a reference goes; more parentheses are text, not a deeper address. */
+const MAX_SUBDIVISIONS = 4;
+/**
+ * "(a)(ii)" のような続きの括弧を、正規表現を複雑にせずに一つずつ読む。
+ * 再帰にしないのは、括弧が延々と続く行でスタックを使い切らないため。
+ */
+const subdivisions = (text, from) => {
+    const parts = [];
+    let end = from;
+    while (parts.length < MAX_SUBDIVISIONS) {
+        const part = SUBDIVISION.exec(text.slice(end, end + 6))?.groups?.["p"];
+        if (part === undefined)
+            break;
+        parts.push(part);
+        end += part.length + 2;
+    }
+    return { parts, end };
+};
+/**
+ * "Section 4.2(a)" → 4.2.a, "Article III" → 3. The same addresses the tree gives.
+ * "Section 9 of the Master Agreement" carries the other document's name, and is not looked up in this tree.
+ */
+const references = (text) => [...text.matchAll(REFERENCE)].flatMap((match) => {
+    const main = numberOf(match.groups?.["n"]);
+    if (main === undefined)
+        return [];
+    const { parts, end } = subdivisions(text, match.index + match[0].length);
+    const document = citedDocumentAfter(text, end);
+    const attrs = { target: [main, ...parts].join("."), label: text.slice(match.index, end), ...(document === undefined ? {} : { document }) };
+    return [{ start: match.index, end, attrs }];
+});
+/** Longest first, and never inside a word: "shall not" is not also "shall", "mayor" is not "may". */
+const MARKERS = [
+    ["is required to", "must"],
+    ["shall not", "must-not"],
+    ["must not", "must-not"],
+    ["may not", "must-not"],
+    ["agrees to", "must"],
+    ["shall", "must"],
+    ["must", "must"],
+    ["may", "may"],
+];
+const isWordChar = (char) => char !== undefined && /[\p{L}\p{N}_]/u.test(char);
+/** Every whole-word occurrence. A loop, not recursion: a line with thousands of "may" must not exhaust the stack. */
+const wordAt = (lower, word) => {
+    const found = [];
+    for (let at = lower.indexOf(word); at !== -1; at = lower.indexOf(word, at + 1)) {
+        if (!isWordChar(lower[at - 1]) && !isWordChar(lower[at + word.length]))
+            found.push(at);
+    }
+    return found;
+};
+const SENTENCE_OPENERS = new Set([".", "!", "?", ":", ";", '"', "“", "("]);
+/** "May" in the middle of a sentence is the month: "published in May 2023". At the start it can be the modal. */
+const isMonthName = (text, at) => {
+    if (!text.startsWith("May", at))
+        return false;
+    const before = text.slice(Math.max(0, at - 4), at).trimEnd();
+    return before !== "" && !SENTENCE_OPENERS.has(before.at(-1) ?? "");
+};
+/**
+ * Longest marker first; a hit that overlaps one already kept is dropped. Overlap is checked with a mark
+ * per character, not by comparing every pair, which would go quadratic on a line with thousands of markers.
+ */
+const obligations = (text) => {
+    const lower = text.toLowerCase();
+    const taken = new Uint8Array(text.length);
+    const kept = [];
+    MARKERS.forEach(([marker, type]) => {
+        wordAt(lower, marker).forEach((start) => {
+            const end = start + marker.length;
+            if (isMonthName(text, start))
+                return;
+            if (taken.subarray(start, end).some((mark) => mark === 1))
+                return;
+            taken.fill(1, start, end);
+            kept.push({ start, end, attrs: { marker, type } });
+        });
+    });
+    return kept.sort((left, right) => left.start - right.start);
+};
+/** Find the number first, then look at what is right before (a currency) and right after (a unit). */
+const NUMBER_RUN = /\d[\d,.]{0,15}/gu;
+const UNITS = [
+    "business days",
+    "business day",
+    "calendar days",
+    "calendar day",
+    "days",
+    "day",
+    "weeks",
+    "week",
+    "months",
+    "month",
+    "years",
+    "year",
+    "hours",
+    "hour",
+    "minutes",
+    "minute",
+    "percent",
+    "times",
+    "%",
+];
+const CURRENCIES = ["USD", "EUR", "$", "€", "£"];
+/** "30." の終わりの点は文の終わり。数の一部にしない。 */
+const withoutTrailingPunctuation = (run) => {
+    let end = run.length;
+    while (end > 0 && (run[end - 1] === "." || run[end - 1] === ","))
+        end -= 1;
+    return run.slice(0, end);
+};
+/** One space or tab may sit between a number and its unit or currency. */
+const isGap = (char) => char === " " || char === "\t";
+const unitAfter = (text, from) => {
+    const gap = isGap(text[from]) ? 1 : 0;
+    const unit = UNITS.find((candidate) => text.startsWith(candidate, from + gap) && !isWordChar(text[from + gap + candidate.length]));
+    return unit;
+};
+/** The currency just before the number, allowing one space. Looks back a few characters, never at the whole line. */
+const currencyBefore = (text, at) => {
+    const end = isGap(text[at - 1]) ? at - 1 : at;
+    return CURRENCIES.find((currency) => text.startsWith(currency, end - currency.length));
+};
+const quantities = (text) => [...text.matchAll(NUMBER_RUN)].flatMap((match) => {
+    const digits = withoutTrailingPunctuation(match[0]);
+    const value = Number(digits.replace(/,/gu, ""));
+    const end = match.index + digits.length;
+    const unit = unitAfter(text, end) ?? currencyBefore(text, match.index);
+    return unit === undefined || Number.isNaN(value) || isWordChar(text[match.index - 1]) ? [] : [{ start: match.index, end, attrs: { value, unit } }];
+});
+const MONTHS = ["january", "february", "march", "april", "may", "june", "july", "august", "september", "october", "november", "december"];
+const MONTH_WORD = /\b(?<month>[A-Z][a-z]{2,8})\b/gu;
+const ISO_DATE = /\b(?<y>\d{4})-(?<m>\d{2})-(?<d>\d{2})\b/gu;
+const DAY_BEFORE = /(?<d>\d{1,2})(?:st|nd|rd|th)? $/u;
+const DAY_YEAR_AFTER = /^ (?<d>\d{1,2})(?:st|nd|rd|th)?,? (?<y>\d{4})\b/u;
+const YEAR_AFTER = /^,? (?<y>\d{4})\b/u;
+const pad = (value) => value.padStart(2, "0");
+/** "1 April 2024": the day written before the month. */
+const dayBefore = (text, at) => {
+    const found = DAY_BEFORE.exec(text.slice(Math.max(0, at - 6), at));
+    const day = found?.groups?.["d"];
+    return found === null || day === undefined ? undefined : { day, start: at - found[0].length };
+};
+/** "April 1, 2024" → 2024-04-01. */
+const monthDayYear = (text, at, end, month) => {
+    const found = DAY_YEAR_AFTER.exec(text.slice(end, end + 16));
+    if (found?.groups === undefined)
+        return undefined;
+    const value = `${found.groups["y"] ?? ""}-${pad(String(month))}-${pad(found.groups["d"] ?? "")}`;
+    return { start: at, end: end + found[0].length, attrs: { value } };
+};
+/** "1 April 2024" → 2024-04-01, "April 2024" → 2024-04. */
+const monthYear = (text, at, end, month) => {
+    const found = YEAR_AFTER.exec(text.slice(end, end + 8));
+    const year = found?.groups?.["y"];
+    if (found === null || year === undefined)
+        return undefined;
+    const before = dayBefore(text, at);
+    const value = [year, pad(String(month)), ...(before === undefined ? [] : [pad(before.day)])].join("-");
+    return { start: before?.start ?? at, end: end + found[0].length, attrs: { value } };
+};
+/**
+ * A month name alone is not a date: "May" is also the modal verb, so it counts only with a year beside it.
+ * The month is found first and its neighbours read with anchored patterns, never one long alternation.
+ */
+const namedDate = (text, match) => {
+    const month = MONTHS.indexOf((match.groups?.["month"] ?? "").toLowerCase()) + 1;
+    if (month === 0)
+        return undefined;
+    const end = match.index + match[0].length;
+    return monthDayYear(text, match.index, end, month) ?? monthYear(text, match.index, end, month);
+};
+const isoDate = (match) => ({
+    start: match.index,
+    end: match.index + match[0].length,
+    attrs: { value: `${match.groups?.["y"] ?? ""}-${match.groups?.["m"] ?? ""}-${match.groups?.["d"] ?? ""}` },
+});
+const dates = (text) => [...[...text.matchAll(MONTH_WORD)].flatMap((match) => namedDate(text, match) ?? []), ...[...text.matchAll(ISO_DATE)].map(isoDate)].sort((left, right) => left.start - right.start);
+/** "2.5 days" and "1.5 times" are amounts, not section 2.5 titled "days". */
+const countedAfter = (_number, rest) => unitAfter(` ${rest}`, 0) !== undefined;
+export const structure = { numbered, definitions, references, obligations, quantities, dates, countedAfter };
+//# sourceMappingURL=structure.js.map

package/dist/structure.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"structure.js","sourceRoot":"","sources":["../src/structure.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,kBAAkB,EAAE,MAAM,eAAe,CAAC;AAEnD,kGAAkG;AAClG,oCAAoC;AAEpC,MAAM,KAAK,GAAqC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,EAAE,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC;AAErF,gFAAgF;AAChF,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,IAAY,EAAsB,EAAE;IAC7D,MAAM,MAAM,GAAG,CAAC,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC;IAClE,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC,IAAI,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,KAAK,SAAS,CAAC;QAAE,OAAO,SAAS,CAAC;IACzF,MAAM,KAAK,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC;IAC5D,OAAO,KAAK,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,KAAK,GAAG,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC;AACrH,CAAC,CAAC;AAEF,MAAM,QAAQ,GAAG,CAAC,IAAwB,EAAsB,EAAE;IAChE,IAAI,IAAI,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IACzC,IAAI,8BAA8B,CAAC,IAAI,CAAC,IAAI,CAAC;QAAE,OAAO,IAAI,CAAC;IAC3D,MAAM,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,CAAC;IAC/B,OAAO,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;AACzD,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,OAAO,GAAG,CAAC,IAAY,EAAsB,EAAE;IACnD,MAAM,OAAO,GAAG,IAAI,CAAC,OAAO,CAAC,eAAe,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;IACzD,IAAI,OAAO,KAAK,EAAE,IAAI,UAAU,CAAC,IAAI,CAAC,OAAO,CAAC;QAAE,OAAO,SAAS,CAAC;IACjE,OAAO,OAAO,CAAC;AACjB,CAAC,CAAC;AAEF,MAAM,OAAO,GAAG,yEAAyE,CAAC;AAC1F,MAAM,OAAO,GAAG,gFAAgF,CAAC;AACjG,MAAM,QAAQ,GAAG,wDAAwD,CAAC;AAC1E,MAAM,WAAW,GAAG,iCAAiC,CAAC;AAEtD,MAAM,MAAM,GAAG,CAAC,OAAe,EAAE,IAAY,EAAE,KAA4B,EAA4B,EAAE;IACvG,MAAM,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,MAAM,CAAC;IAC1C,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;IACvC,MAAM,OAAO,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC;IAChD,IAAI,MAAM,KAAK,SAAS,IAAI,MAAM,KAAK,SAAS,IAAI,OAAO,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAC5F,MAAM,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IAChC,OAAO;QACL,IAAI,EAAE,SAAS;QACf,KAAK,EAAE,KAAK,CAAC,MAAM;QACnB,MAAM;QACN,QAAQ,EAAE,IAAI;QACd,KAAK,EAAE,KAAK,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC;QAC/B,OAAO;QACP,IAAI,EAAE,OAAO;QACb,OAAO,EAAE,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;KAC9B,CAAC;AACJ,CAAC,CAAC;AAIF;;;GAGG;AACH,MAAM,YAAY,GAAG,CAAC,IAAkB,EAAE,KAAa,EAAE,GAA4B,EAAqB,EAAE;IAC1G,MAAM,KAAK,GAAG,4BAA4B,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC;IAC3E,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAC1C,IAAI,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC;QAAE,OAAO,OAAO,CAAC;IACzC,IAAI,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC;QAAE,OAAO,OAAO,CAAC;IAC5C,MAAM,KAAK,GAAG,GAAG,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;IAC7B,OAAO,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,KAAK,KAAK,SAAS,IAAI,YAAY,CAAC,KAAK,EAAE,KAAK,GAAG,CAAC,EAAE,GAAG,CAAC,KAAK,QAAQ,IAAI,KAAK,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,QAAQ,CAAC;AAC5J,CAAC,CAAC;AAEF,MAAM,MAAM,GAAG,CAAC,OAAyB,EAAyB,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,GAAG,EAAE,EAAE,CAAC,YAAY,CAAC,IAAI,EAAE,KAAK,EAAE,GAAG,CAAC,CAAC,CAAC;AAE5I,sDAAsD;AACtD,MAAM,aAAa,GAAG,CAAC,GAAW,EAAE,OAAyB,EAAW,EAAE,CACxE,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC,KAAK,CAAC,KAAK,QAAQ,IAAI,IAAI,CAAC,MAAM,CAAC,UAAU,CAAC,CAAC,CAAC,GAAG,CAAC,KAAK,GAAG,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,CAAC;AAEjI,gHAAgH;AAChH,MAAM,OAAO,GAAG,CAAC,GAAW,EAAE,OAAyB,EAAS,EAAE;IAChE,IAAI,QAAQ,CAAC,IAAI,CAAC,GAAG,CAAC;QAAE,OAAO,OAAO,CAAC;IACvC,IAAI,WAAW,CAAC,IAAI,CAAC,GAAG,CAAC;QAAE,OAAO,OAAO,CAAC;IAC1C,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,aAAa,CAAC,GAAG,EAAE,OAAO,CAAC;QAAE,OAAO,QAAQ,CAAC;IAC1E,MAAM,IAAI,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC;IAC7B,OAAO,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,IAAI,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,KAAK,QAAQ,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,QAAQ,CAAC;AACjF,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,QAAQ,GAAG,CAAC,KAAY,EAAE,OAAyB,EAAU,EAAE;IACnE,MAAM,IAAI,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC;IAC7B,MAAM,OAAO,GAAG,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,QAAQ,EAAE,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,MAAM,GAAG,CAAC,GAAG,QAAQ,CAAC,KAAK,KAAK,CAAC,CAAC;IAC1H,OAAO,OAAO,EAAE,KAAK,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC;AACjE,CAAC,CAAC;AAEF,MAAM,eAAe,GAAG,GAAG,CAAC,UAAU,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;AAE9C,kEAAkE;AAClE,MAAM,SAAS,GAAG,CAAC,GAAW,EAAE,KAAY,EAAsB,EAAE;IAClE,IAAI,KAAK,KAAK,OAAO;QAAE,OAAO,MAAM,CAAC,GAAG,CAAC,CAAC;IAC1C,IAAI,KAAK,KAAK,OAAO;QAAE,OAAO,UAAU,CAAC,GAAG,CAAC,CAAC;IAC9C,OAAO,GAAG,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC,CAAC,GAAG,eAAe,CAAC,CAAC,CAAC,SAAS,CAAC;AAC5E,CAAC,CAAC;AAEF,MAAM,QAAQ,GAAG,CAAC,IAAY,EAAE,OAAyB,EAA4B,EAAE;IACrF,MAAM,MAAM,GAAG,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,MAAM,CAAC;IAC3C,MAAM,GAAG,GAAG,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC;IAC1B,IAAI,MAAM,KAAK,SAAS,IAAI,GAAG,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAChE,MAAM,KAAK,GAAG,OAAO,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;IACpC,iEAAiE;IACjE,OAAO;QACL,IAAI,EAAE,MAAM;QACZ,KAAK,EAAE,QAAQ,CAAC,KAAK,EAAE,OAAO,CAAC;QAC/B,OAAO,EAAE,SAAS,CAAC,GAAG,EAAE,KAAK,CAAC;QAC9B,MAAM,EAAE,GAAG;QACX,QAAQ,EAAE,KAAK;QACf,KAAK,EAAE,IAAI,GAAG,GAAG;QACjB,OAAO,EAAE,EAAE;QACX,IAAI,EAAE,MAAM,CAAC,MAAM,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE;KACnC,CAAC;AACJ,CAAC,CAAC;AAEF,8GAA8G;AAC9G,MAAM,IAAI,GAAG,mEAAmE,CAAC;AACjF,MAAM,OAAO,GAAG,yEAAyE,CAAC;AAE1F,MAAM,OAAO,GAAG,CAAC,OAAe,EAAE,IAAY,EAAE,KAAa,EAAE,MAAc,EAAE,IAAY,EAA4B,EAAE;IACvH,MAAM,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,MAAM,CAAC;IAC1C,MAAM,MAAM,GAAG,QAAQ,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;IACvC,MAAM,OAAO,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC;IAChD,IAAI,MAAM,KAAK,SAAS,IAAI,MAAM,KAAK,SAAS,IAAI,OAAO,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAC5F,MAAM,KAAK,GAAG,GAAG,IAAI,IAAI,MAAM,CAAC,GAAG,CAAC,IAAI,EAAE,EAAE,CAAC;IAC7C,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,GAAG,MAAM,EAAE,QAAQ,EAAE,KAAK,EAAE,KAAK,EAAE,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,OAAO,EAAE,MAAM,CAAC,MAAM,CAAC,EAAE,CAAC;AACtI,CAAC,CAAC;AAEF,MAAM,QAAQ,GAAG,CAAC,IAAY,EAAE,OAAyB,EAA4B,EAAE,CACrF,OAAO,CAAC,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,CAAC;IACrC,OAAO,CAAC,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,SAAS,CAAC;IAC3C,MAAM,CAAC,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,WAAW,CAAC,EAAE,CAAC;IAC5C,MAAM,CAAC,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,WAAW,CAAC,EAAE,CAAC;IAC5C,QAAQ,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;AAE1B,MAAM,QAAQ,GAAG,CACf,OAAe,EACf,IAAY,EACZ,KAA4G,EACjG,EAAE,CACb,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE;IAC5C,MAAM,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,MAAM,IAAI,EAAE,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IAClD,OAAO,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,GAAG,EAAE,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,EAAE,KAAK,EAAE,KAAK,EAAE,CAAC,CAAC;AAC/G,CAAC,CAAC,CAAC;AAEL,MAAM,WAAW,GAAG;IAClB,mFAAmF;IACnF,4DAA4D;IAC5D,iEAAiE;CAClE,CAAC;AAEF,MAAM,WAAW,GAAG,CAAC,IAAY,EAAa,EAAE,CAC9C,WAAW,CAAC,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,QAAQ,CAAC,OAAO,EAAE,IAAI,EAAE,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,MAAM,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;AAE/I,MAAM,SAAS,GAAG,+EAA+E,CAAC;AAClG,MAAM,WAAW,GAAG,2BAA2B,CAAC;AAEhD,oGAAoG;AACpG,MAAM,gBAAgB,GAAG,CAAC,CAAC;AAE3B;;;GAGG;AACH,MAAM,YAAY,GAAG,CAAC,IAAY,EAAE,IAAY,EAA+D,EAAE;IAC/G,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,IAAI,GAAG,GAAG,IAAI,CAAC;IACf,OAAO,KAAK,CAAC,MAAM,GAAG,gBAAgB,EAAE,CAAC;QACvC,MAAM,IAAI,GAAG,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,EAAE,GAAG,GAAG,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC;QACvE,IAAI,IAAI,KAAK,SAAS;YAAE,MAAM;QAC9B,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACjB,GAAG,IAAI,IAAI,CAAC,MAAM,GAAG,CAAC,CAAC;IACzB,CAAC;IACD,OAAO,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC;AACxB,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,UAAU,GAAG,CAAC,IAAY,EAAa,EAAE,CAC7C,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE;IAC9C,MAAM,IAAI,GAAG,QAAQ,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;IAC3C,IAAI,IAAI,KAAK,SAAS;QAAE,OAAO,EAAE,CAAC;IAClC,MAAM,EAAE,KAAK,EAAE,GAAG,EAAE,GAAG,YAAY,CAAC,IAAI,EAAE,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC;IACzE,MAAM,QAAQ,GAAG,kBAAkB,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;IAC/C,MAAM,KAAK,GAAG,EAAE,MAAM,EAAE,CAAC,IAAI,EAAE,GAAG,KAAK,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,KAAK,EAAE,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,CAAC,EAAE,GAAG,CAAC,QAAQ,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,CAAC,EAAE,CAAC;IAC3I,OAAO,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,GAAG,EAAE,KAAK,EAAE,CAAC,CAAC;AAC9C,CAAC,CAAC,CAAC;AAEL,qGAAqG;AACrG,MAAM,OAAO,GAAgE;IAC3E,CAAC,gBAAgB,EAAE,MAAM,CAAC;IAC1B,CAAC,WAAW,EAAE,UAAU,CAAC;IACzB,CAAC,UAAU,EAAE,UAAU,CAAC;IACxB,CAAC,SAAS,EAAE,UAAU,CAAC;IACvB,CAAC,WAAW,EAAE,MAAM,CAAC;IACrB,CAAC,OAAO,EAAE,MAAM,CAAC;IACjB,CAAC,MAAM,EAAE,MAAM,CAAC;IAChB,CAAC,KAAK,EAAE,KAAK,CAAC;CACf,CAAC;AAEF,MAAM,UAAU,GAAG,CAAC,IAAwB,EAAW,EAAE,CAAC,IAAI,KAAK,SAAS,IAAI,gBAAgB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAE5G,qHAAqH;AACrH,MAAM,MAAM,GAAG,CAAC,KAAa,EAAE,IAAY,EAAY,EAAE;IACvD,MAAM,KAAK,GAAa,EAAE,CAAC;IAC3B,KAAK,IAAI,EAAE,GAAG,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,EAAE,KAAK,CAAC,CAAC,EAAE,EAAE,GAAG,KAAK,CAAC,OAAO,CAAC,IAAI,EAAE,EAAE,GAAG,CAAC,CAAC,EAAE,CAAC;QAC/E,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC;YAAE,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;IACzF,CAAC;IACD,OAAO,KAAK,CAAC;AACf,CAAC,CAAC;AAEF,MAAM,gBAAgB,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAE3E,iHAAiH;AACjH,MAAM,WAAW,GAAG,CAAC,IAAY,EAAE,EAAU,EAAW,EAAE;IACxD,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,KAAK,EAAE,EAAE,CAAC;QAAE,OAAO,KAAK,CAAC;IAC9C,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,GAAG,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC;IAC7D,OAAO,MAAM,KAAK,EAAE,IAAI,CAAC,gBAAgB,CAAC,GAAG,CAAC,MAAM,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC;AACrE,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,WAAW,GAAG,CAAC,IAAY,EAAa,EAAE;IAC9C,MAAM,KAAK,GAAG,IAAI,CAAC,WAAW,EAAE,CAAC;IACjC,MAAM,KAAK,GAAG,IAAI,UAAU,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;IAC1C,MAAM,IAAI,GAAc,EAAE,CAAC;IAC3B,OAAO,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,EAAE,IAAI,CAAC,EAAE,EAAE;QACjC,MAAM,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE;YACtC,MAAM,GAAG,GAAG,KAAK,GAAG,MAAM,CAAC,MAAM,CAAC;YAClC,IAAI,WAAW,CAAC,IAAI,EAAE,KAAK,CAAC;gBAAE,OAAO;YACrC,IAAI,KAAK,CAAC,QAAQ,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,KAAK,CAAC,CAAC;gBAAE,OAAO;YAClE,KAAK,CAAC,IAAI,CAAC,CAAC,EAAE,KAAK,EAAE,GAAG,CAAC,CAAC;YAC1B,IAAI,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,KAAK,EAAE,EAAE,MAAM,EAAE,IAAI,EAAE,EAAE,CAAC,CAAC;QACrD,CAAC,CAAC,CAAC;IACL,CAAC,CAAC,CAAC;IACH,OAAO,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC;AAC9D,CAAC,CAAC;AAEF,sGAAsG;AACtG,MAAM,UAAU,GAAG,kBAAkB,CAAC;AACtC,MAAM,KAAK,GAAG;IACZ,eAAe;IACf,cAAc;IACd,eAAe;IACf,cAAc;IACd,MAAM;IACN,KAAK;IACL,OAAO;IACP,MAAM;IACN,QAAQ;IACR,OAAO;IACP,OAAO;IACP,MAAM;IACN,OAAO;IACP,MAAM;IACN,SAAS;IACT,QAAQ;IACR,SAAS;IACT,OAAO;IACP,GAAG;CACJ,CAAC;AACF,MAAM,UAAU,GAAG,CAAC,KAAK,EAAE,KAAK,EAAE,GAAG,EAAE,GAAG,EAAE,GAAG,CAAC,CAAC;AAEjD,mCAAmC;AACnC,MAAM,0BAA0B,GAAG,CAAC,GAAW,EAAU,EAAE;IACzD,IAAI,GAAG,GAAG,GAAG,CAAC,MAAM,CAAC;IACrB,OAAO,GAAG,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,GAAG,CAAC,CAAC,KAAK,GAAG,IAAI,GAAG,CAAC,GAAG,GAAG,CAAC,CAAC,KAAK,GAAG,CAAC;QAAE,GAAG,IAAI,CAAC,CAAC;IAC3E,OAAO,GAAG,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;AAC3B,CAAC,CAAC;AAEF,0EAA0E;AAC1E,MAAM,KAAK,GAAG,CAAC,IAAwB,EAAW,EAAE,CAAC,IAAI,KAAK,GAAG,IAAI,IAAI,KAAK,IAAI,CAAC;AAEnF,MAAM,SAAS,GAAG,CAAC,IAAY,EAAE,IAAY,EAAsB,EAAE;IACnE,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IACtC,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,SAAS,EAAE,IAAI,GAAG,GAAG,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,GAAG,GAAG,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC;IACnI,OAAO,IAAI,CAAC;AACd,CAAC,CAAC;AAEF,qHAAqH;AACrH,MAAM,cAAc,GAAG,CAAC,IAAY,EAAE,EAAU,EAAsB,EAAE;IACtE,MAAM,GAAG,GAAG,KAAK,CAAC,IAAI,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IAC9C,OAAO,UAAU,CAAC,IAAI,CAAC,CAAC,QAAQ,EAAE,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,QAAQ,EAAE,GAAG,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC;AACzF,CAAC,CAAC;AAEF,MAAM,UAAU,GAAG,CAAC,IAAY,EAAa,EAAE,CAC7C,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,UAAU,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE;IAC/C,MAAM,MAAM,GAAG,0BAA0B,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC;IACpD,MAAM,KAAK,GAAG,MAAM,CAAC,MAAM,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,CAAC;IAChD,MAAM,GAAG,GAAG,KAAK,CAAC,KAAK,GAAG,MAAM,CAAC,MAAM,CAAC;IACxC,MAAM,IAAI,GAAG,SAAS,CAAC,IAAI,EAAE,GAAG,CAAC,IAAI,cAAc,CAAC,IAAI,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC;IACvE,OAAO,IAAI,KAAK,SAAS,IAAI,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,GAAG,EAAE,KAAK,EAAE,EAAE,KAAK,EAAE,IAAI,EAAE,EAAE,CAAC,CAAC;AACrJ,CAAC,CAAC,CAAC;AAEL,MAAM,MAAM,GAAG,CAAC,SAAS,EAAE,UAAU,EAAE,OAAO,EAAE,OAAO,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,QAAQ,EAAE,WAAW,EAAE,SAAS,EAAE,UAAU,EAAE,UAAU,CAAC,CAAC;AAC1I,MAAM,UAAU,GAAG,iCAAiC,CAAC;AACrD,MAAM,QAAQ,GAAG,2CAA2C,CAAC;AAC7D,MAAM,UAAU,GAAG,kCAAkC,CAAC;AACtD,MAAM,cAAc,GAAG,kDAAkD,CAAC;AAC1E,MAAM,UAAU,GAAG,oBAAoB,CAAC;AAExC,MAAM,GAAG,GAAG,CAAC,KAAa,EAAU,EAAE,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC,EAAE,GAAG,CAAC,CAAC;AAE9D,wDAAwD;AACxD,MAAM,SAAS,GAAG,CAAC,IAAY,EAAE,EAAU,EAAgE,EAAE;IAC3G,MAAM,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,EAAE,GAAG,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC;IACnE,MAAM,GAAG,GAAG,KAAK,EAAE,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC;IACjC,OAAO,KAAK,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,GAAG,EAAE,KAAK,EAAE,EAAE,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC;AAChG,CAAC,CAAC;AAEF,oCAAoC;AACpC,MAAM,YAAY,GAAG,CAAC,IAAY,EAAE,EAAU,EAAE,GAAW,EAAE,KAAa,EAAuB,EAAE;IACjG,MAAM,KAAK,GAAG,cAAc,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,EAAE,GAAG,GAAG,EAAE,CAAC,CAAC,CAAC;IAC7D,IAAI,KAAK,EAAE,MAAM,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAClD,MAAM,KAAK,GAAG,GAAG,KAAK,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,EAAE,IAAI,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,IAAI,GAAG,CAAC,KAAK,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC;IACjG,OAAO,EAAE,KAAK,EAAE,EAAE,EAAE,GAAG,EAAE,GAAG,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,KAAK,EAAE,EAAE,CAAC;AACrE,CAAC,CAAC;AAEF,2DAA2D;AAC3D,MAAM,SAAS,GAAG,CAAC,IAAY,EAAE,EAAU,EAAE,GAAW,EAAE,KAAa,EAAuB,EAAE;IAC9F,MAAM,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,EAAE,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC;IACxD,MAAM,IAAI,GAAG,KAAK,EAAE,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC;IAClC,IAAI,KAAK,KAAK,IAAI,IAAI,IAAI,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAC3D,MAAM,MAAM,GAAG,SAAS,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;IACnC,MAAM,KAAK,GAAG,CAAC,IAAI,EAAE,GAAG,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,MAAM,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IACvG,OAAO,EAAE,KAAK,EAAE,MAAM,EAAE,KAAK,IAAI,EAAE,EAAE,GAAG,EAAE,GAAG,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,EAAE,KAAK,EAAE,EAAE,KAAK,EAAE,EAAE,CAAC;AACtF,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,SAAS,GAAG,CAAC,IAAY,EAAE,KAAsB,EAAuB,EAAE;IAC9E,MAAM,KAAK,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC,CAAC,WAAW,EAAE,CAAC,GAAG,CAAC,CAAC;IAChF,IAAI,KAAK,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC;IAClC,MAAM,GAAG,GAAG,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC;IAC1C,OAAO,YAAY,CAAC,IAAI,EAAE,KAAK,CAAC,KAAK,EAAE,GAAG,EAAE,KAAK,CAAC,IAAI,SAAS,CAAC,IAAI,EAAE,KAAK,CAAC,KAAK,EAAE,GAAG,EAAE,KAAK,CAAC,CAAC;AACjG,CAAC,CAAC;AAEF,MAAM,OAAO,GAAG,CAAC,KAAsB,EAAW,EAAE,CAAC,CAAC;IACpD,KAAK,EAAE,KAAK,CAAC,KAAK;IAClB,GAAG,EAAE,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM;IAClC,KAAK,EAAE,EAAE,KAAK,EAAE,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,IAAI,EAAE,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,IAAI,EAAE,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,IAAI,EAAE,EAAE,EAAE;CAC3G,CAAC,CAAC;AAEH,MAAM,KAAK,GAAG,CAAC,IAAY,EAAa,EAAE,CACxC,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,UAAU,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,SAAS,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,EAAE,CAAC,EAAE,GAAG,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,IAAI,CACrI,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,KAAK,CAC1C,CAAC;AAEJ,6EAA6E;AAC7E,MAAM,YAAY,GAAG,CAAC,OAAe,EAAE,IAAY,EAAW,EAAE,CAAC,SAAS,CAAC,IAAI,IAAI,EAAE,EAAE,CAAC,CAAC,KAAK,SAAS,CAAC;AAExG,MAAM,CAAC,MAAM,SAAS,GAAsB,EAAE,QAAQ,EAAE,WAAW,EAAE,UAAU,EAAE,WAAW,EAAE,UAAU,EAAE,KAAK,EAAE,YAAY,EAAE,CAAC"}

package/lexicons/contraction.yaml

@@ -0,0 +1,34 @@
+# Contractions and the long forms that say the same thing.
+#
+# Neither is wrong. The rule only asks whether one document picks one and keeps it.
+id: contraction
+language: en
+entries:
+  - pattern: "don't"
+    instead_of: do not
+  - pattern: "doesn't"
+    instead_of: does not
+  - pattern: "didn't"
+    instead_of: did not
+  - pattern: "can't"
+    instead_of: cannot
+  - pattern: "won't"
+    instead_of: will not
+  - pattern: "isn't"
+    instead_of: is not
+  - pattern: "aren't"
+    instead_of: are not
+  - pattern: "wasn't"
+    instead_of: was not
+  - pattern: "it's"
+    instead_of: it is
+  - pattern: "we're"
+    instead_of: we are
+  - pattern: "they're"
+    instead_of: they are
+  - pattern: "you're"
+    instead_of: you are
+  - pattern: "we'll"
+    instead_of: we will
+  - pattern: "we've"
+    instead_of: we have

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@chaffjs/lang-en",
-  "version": "0.3.0",
+  "version": "0.6.0",
   "description": "English language adapter for chaff",
   "license": "MIT",
   "author": "isamu",

package/src/citation.ts

@@ -0,0 +1,41 @@
+// A reference into another document: "Section 9 of the Master Agreement" is not this document's Section 9.
+
+const OF = /^,? of (?:the )?/u;
+const CAPITALISED = /^[A-Z][\w'’-]*/u;
+/**
+ * How a document names itself. "Section 3 of the Agreement" in an agreement means this one.
+ * "the Act" and "the Code" are left out: a statute calls itself "this Act", and a contract's "the Act" is a law it cites.
+ */
+const SELF = new Set(["Agreement", "Contract", "Terms", "Policy"]);
+/** Words that may sit inside a title: "Code of Federal Regulations". */
+const JOINERS = new Set(["of", "and", "for", "on"]);
+const MAX_WORDS = 8;
+
+/** Capitalised words from the start of `rest`, allowing a joiner between two of them. Punctuation ends the title. */
+const titleWords = (rest: string): string[] => {
+  const tokens = rest.split(" ");
+  const words: string[] = [];
+  for (let index = 0; index < tokens.length && words.length < MAX_WORDS; index += 1) {
+    const token = tokens[index] ?? "";
+    const word = CAPITALISED.exec(token)?.[0];
+    const joins = word === undefined && words.length > 0 && JOINERS.has(token) && CAPITALISED.test(tokens[index + 1] ?? "");
+    if (word === undefined && !joins) break;
+    words.push(word ?? token);
+    if (word !== undefined && word.length < token.length) break;
+  }
+  return words;
+};
+
+/**
+ * The document named right after a reference, or undefined when the reference is into this document.
+ * "of this Agreement" and "of the Agreement" are this document; "of the Master Agreement" is another.
+ */
+export const citedDocumentAfter = (text: string, end: number): string | undefined => {
+  const rest = text.slice(end, end + 120);
+  const of = OF.exec(rest);
+  if (of === null) return undefined;
+  const words = titleWords(rest.slice(of[0].length));
+  if (words.length === 0) return undefined;
+  const name = words.join(" ");
+  return words.length === 1 && SELF.has(name) ? undefined : name;
+};

package/src/index.ts

@@ -1,5 +1,6 @@
 import { split, SentenceSplitterSyntax } from "sentence-splitter";
 import { loadLexicons } from "./lexicons.ts";
+import { structure } from "./structure.ts";
 import { isReady, prepare, tokenize } from "./pos.ts";
 import type { AdapterNeeds, LanguageAdapter, Segmentation, Sentence } from "chaffjs/plugin";
 
@@ -48,6 +49,7 @@ export const adapter: LanguageAdapter = {
     return [...source.matchAll(LATIN_LETTER)].length / total;
   },
   lexicons: loadLexicons(),
+  structure,
   segment: (text: string): Segmentation => {
     const sentences: Sentence[] = split(text)
       .filter((node) => node.type === SentenceSplitterSyntax.Sentence)

package/src/lexicons.ts

@@ -11,7 +11,12 @@ const isRecord = (value: unknown): value is Record<string, unknown> => typeof va
 const toEntry = (raw: unknown): LexiconEntry | undefined => {
   if (!isRecord(raw) || typeof raw["pattern"] !== "string") return undefined;
   const weight = raw["weight"];
-  return { pattern: raw["pattern"], weight: typeof weight === "number" ? weight : undefined };
+  const instead = raw["instead_of"];
+  return {
+    pattern: raw["pattern"],
+    weight: typeof weight === "number" ? weight : undefined,
+    instead_of: typeof instead === "string" ? instead : undefined,
+  };
 };
 
 /**

package/src/pos.ts

@@ -81,7 +81,14 @@ export const upos = (tag: string): string => (PUNCTUATION.has(tag) ? "PUNCT" : (
  * 受動は be + 過去分詞。過去分詞だけでは完了形（has reviewed）と見分けられないので、
  * 直前の be を見る。間に副詞が挟まる（was quickly approved）ぶんだけ遡る。
  */
-const ADVERB_TAGS = new Set(["RB", "RBR", "RBS"]);
+/**
+ * be と過去分詞の間に立てるもの。副詞と、副詞をつなぐ接続詞（was fully and carefully reviewed）。
+ *
+ * これでも拾えない形がある。`was fully and finally approved` では wink が `approved` を
+ * VBN ではなく **VBD** と付けるため、そもそも過去分詞として見えない。解析器の限界で、
+ * VBD も受動と見なすと `The team was here and approved it` まで受動になる。直さない。
+ */
+const SKIPPABLE = new Set(["RB", "RBR", "RBS", "CC", ","]);
 
 const BE = new Set(["be", "am", "is", "are", "was", "were", "been", "being"]);
 
@@ -90,7 +97,7 @@ const isBe = (entry: Tagged): boolean => BE.has(entry.lemma ?? entry.value.toLow
 const isPassive = (tagged: readonly Tagged[], at: number): boolean => {
   if (tagged[at]?.pos !== "VBN") return false;
   const before = tagged.slice(0, at).reverse();
-  const head = before.find((entry) => !ADVERB_TAGS.has(entry.pos));
+  const head = before.find((entry) => !SKIPPABLE.has(entry.pos));
   return head !== undefined && isBe(head);
 };
 

package/src/structure.ts

@@ -0,0 +1,366 @@
+import type { Mention, NumberedLine, NumberingContext, StructurePatterns } from "chaffjs/plugin";
+import { citedDocumentAfter } from "./citation.ts";
+
+// Contracts, specifications and statutes in English. core nests what this reads; it does not know
+// how English numbers its articles.
+
+const ROMAN: Readonly<Record<string, number>> = { i: 1, v: 5, x: 10, l: 50, c: 100 };
+
+/** "IV" → 4, "xii" → 12. Undefined for anything that is not a roman numeral. */
+export const parseRoman = (text: string): number | undefined => {
+  const values = [...text.toLowerCase()].map((char) => ROMAN[char]);
+  if (values.length === 0 || values.some((value) => value === undefined)) return undefined;
+  const known = values.filter((value) => value !== undefined);
+  return known.reduce((total, value, index) => ((known[index + 1] ?? 0) > value ? total - value : total + value), 0);
+};
+
+const numberOf = (text: string | undefined): string | undefined => {
+  if (text === undefined) return undefined;
+  if (/^\d{1,3}(?:\.\d{1,3}){0,5}$/u.test(text)) return text;
+  const roman = parseRoman(text);
+  return roman === undefined ? undefined : String(roman);
+};
+
+/**
+ * A heading is the number followed by nothing, punctuation, or a capitalised title.
+ * "Article 3 shall apply to…" at the start of a line is a sentence, not the heading of Article 3.
+ */
+const titleOf = (rest: string): string | undefined => {
+  const trimmed = rest.replace(/^[.:\-–—\s]+/u, "").trim();
+  if (trimmed !== "" && /^\p{Ll}/u.test(trimmed)) return undefined;
+  return trimmed;
+};
+
+const ARTICLE = /^\s{0,3}(?:ARTICLE|Article)\s+(?<n>\d{1,3}|[IVXLC]{1,7})\b(?<rest>.*)$/u;
+const SECTION = /^\s{0,3}(?:SECTION|Section|§)\s*(?<n>\d{1,3}(?:\.\d{1,3}){0,5})\b(?<rest>.*)$/u;
+const LETTERED = /^\s{0,6}\((?<n>[a-z]{1,4}|\d{1,3})\)\s+(?<rest>\S.*)$/u;
+const MULTI_ROMAN = /^(?:ii|iii|iv|vi|vii|viii|ix)$/u;
+
+const headed = (pattern: RegExp, line: string, label: (n: string) => string): NumberedLine | undefined => {
+  const groups = pattern.exec(line)?.groups;
+  const number = numberOf(groups?.["n"]);
+  const heading = titleOf(groups?.["rest"] ?? "");
+  if (groups === undefined || number === undefined || heading === undefined) return undefined;
+  const parts = number.split(".");
+  return {
+    kind: "article",
+    depth: parts.length,
+    number,
+    absolute: true,
+    label: label(groups["n"] ?? ""),
+    heading,
+    rest: heading,
+    ordinal: Number(parts.at(-1)),
+  };
+};
+
+type Style = "letter" | "roman" | "digit";
+
+/**
+ * 開いている項目の書き方。"(ii)" はローマ数字、"(b)" は英字。一文字の "(i)" はどちらにも読めるので、
+ * 開いたときの読みを覚えておく代わりに、一つ上に英字が開いていたかで決め直す。
+ */
+const styleOfLabel = (open: NumberedLine, index: number, all: readonly NumberedLine[]): Style | undefined => {
+  const inner = /^\((?<n>[a-z0-9]{1,4})\)$/u.exec(open.label)?.groups?.["n"];
+  if (inner === undefined) return undefined;
+  if (/^\d+$/u.test(inner)) return "digit";
+  if (MULTI_ROMAN.test(inner)) return "roman";
+  const above = all[index - 1];
+  return /^[ivx]$/u.test(inner) && above !== undefined && styleOfLabel(above, index - 1, all) === "letter" && above.depth < open.depth ? "roman" : "letter";
+};
+
+const styles = (context: NumberingContext): (Style | undefined)[] => context.open.map((open, index, all) => styleOfLabel(open, index, all));
+
+/** "(h)" の次の "(i)" は英字。開いている英字の次の文字なら、ローマ数字とは読まない。 */
+const followsLetter = (raw: string, context: NumberingContext): boolean =>
+  context.open.some((open, index) => styles(context)[index] === "letter" && open.number.charCodeAt(0) + 1 === raw.charCodeAt(0));
+
+/** "(i)" is a roman numeral right under "(a)", or when a roman list is already open; the letter i otherwise. */
+const styleOf = (raw: string, context: NumberingContext): Style => {
+  if (/^\d+$/u.test(raw)) return "digit";
+  if (MULTI_ROMAN.test(raw)) return "roman";
+  if (!/^[ivx]$/u.test(raw) || followsLetter(raw, context)) return "letter";
+  const open = styles(context);
+  return open.includes("roman") || open.at(-1) === "letter" ? "roman" : "letter";
+};
+
+/**
+ * A sibling has the depth of the open item written the same way: "(b)" closes "(i)" and sits beside "(a)".
+ * A new way of numbering goes one deeper than whatever is open.
+ */
+const depthFor = (style: Style, context: NumberingContext): number => {
+  const open = styles(context);
+  const sibling = [...context.open].reverse().find((_item, reversed) => open[context.open.length - 1 - reversed] === style);
+  return sibling?.depth ?? (context.open.at(-1)?.depth ?? 0) + 1;
+};
+
+const LETTER_BEFORE_A = "a".charCodeAt(0) - 1;
+
+/** "(b)" は 2 番目、"(ii)" も 2 番目。二文字以上の英字（"(aa)"）は並びが決まらないので付けない。 */
+const ordinalOf = (raw: string, style: Style): number | undefined => {
+  if (style === "digit") return Number(raw);
+  if (style === "roman") return parseRoman(raw);
+  return raw.length === 1 ? raw.charCodeAt(0) - LETTER_BEFORE_A : undefined;
+};
+
+const lettered = (line: string, context: NumberingContext): NumberedLine | undefined => {
+  const groups = LETTERED.exec(line)?.groups;
+  const raw = groups?.["n"];
+  if (groups === undefined || raw === undefined) return undefined;
+  const style = styleOf(raw, context);
+  // 番地は書かれたままの "ii" を使う。参照「Section 4.2(a)(ii)」も同じ形で書かれるので、そのまま引ける。
+  return {
+    kind: "item",
+    depth: depthFor(style, context),
+    ordinal: ordinalOf(raw, style),
+    number: raw,
+    absolute: false,
+    label: `(${raw})`,
+    heading: "",
+    rest: groups["rest"]?.trim() ?? "",
+  };
+};
+
+/** Chapters restart in each part, so a chapter continues its part's address: PART II, CHAPTER 1 → pt2.ch1. */
+const PART = /^\s{0,3}(?:PART|Part)\s+(?<n>\d{1,3}|[IVXLC]{1,7})\b(?<rest>.*)$/u;
+const CHAPTER = /^\s{0,3}(?:CHAPTER|Chapter)\s+(?<n>\d{1,3}|[IVXLC]{1,7})\b(?<rest>.*)$/u;
+
+const chapter = (pattern: RegExp, line: string, depth: number, prefix: string, word: string): NumberedLine | undefined => {
+  const groups = pattern.exec(line)?.groups;
+  const number = numberOf(groups?.["n"]);
+  const heading = titleOf(groups?.["rest"] ?? "");
+  if (groups === undefined || number === undefined || heading === undefined) return undefined;
+  const label = `${word} ${groups["n"] ?? ""}`;
+  return { kind: "chapter", depth, number: prefix + number, absolute: false, label, heading, rest: heading, ordinal: Number(number) };
+};
+
+const numbered = (line: string, context: NumberingContext): NumberedLine | undefined =>
+  chapter(PART, line, -2, "pt", "Part") ??
+  chapter(CHAPTER, line, -1, "ch", "Chapter") ??
+  headed(ARTICLE, line, (n) => `Article ${n}`) ??
+  headed(SECTION, line, (n) => `Section ${n}`) ??
+  lettered(line, context);
+
+const mentions = (
+  pattern: RegExp,
+  text: string,
+  attrs: (groups: Readonly<Record<string, string | undefined>>, whole: string) => Mention["attrs"] | undefined,
+): Mention[] =>
+  [...text.matchAll(pattern)].flatMap((match) => {
+    const found = attrs(match.groups ?? {}, match[0]);
+    return found === undefined ? [] : [{ start: match.index, end: match.index + match[0].length, attrs: found }];
+  });
+
+const DEFINITIONS = [
+  /["“](?<term>[^"”\n]{1,60})["”] (?:means|shall mean|refers to|has the meaning)\b/gu,
+  /\((?:the |hereinafter )?["“](?<term>[^"”\n]{1,60})["”]\)/gu,
+  /\(hereinafter referred to as ["“](?<term>[^"”\n]{1,60})["”]\)/gu,
+];
+
+const definitions = (text: string): Mention[] =>
+  DEFINITIONS.flatMap((pattern) => mentions(pattern, text, (groups) => (groups["term"] === undefined ? undefined : { term: groups["term"] })));
+
+const REFERENCE = /\b(?:Sections?|Articles?|§) ?(?<n>\d{1,3}(?:\.\d{1,3}){0,5}|[IVXLC]{1,7})\b/gu;
+const SUBDIVISION = /^\((?<p>[a-z0-9]{1,4})\)/u;
+
+/** "(a)(ii)(3)" is as deep as a reference goes; more parentheses are text, not a deeper address. */
+const MAX_SUBDIVISIONS = 4;
+
+/**
+ * "(a)(ii)" のような続きの括弧を、正規表現を複雑にせずに一つずつ読む。
+ * 再帰にしないのは、括弧が延々と続く行でスタックを使い切らないため。
+ */
+const subdivisions = (text: string, from: number): { readonly parts: readonly string[]; readonly end: number } => {
+  const parts: string[] = [];
+  let end = from;
+  while (parts.length < MAX_SUBDIVISIONS) {
+    const part = SUBDIVISION.exec(text.slice(end, end + 6))?.groups?.["p"];
+    if (part === undefined) break;
+    parts.push(part);
+    end += part.length + 2;
+  }
+  return { parts, end };
+};
+
+/**
+ * "Section 4.2(a)" → 4.2.a, "Article III" → 3. The same addresses the tree gives.
+ * "Section 9 of the Master Agreement" carries the other document's name, and is not looked up in this tree.
+ */
+const references = (text: string): Mention[] =>
+  [...text.matchAll(REFERENCE)].flatMap((match) => {
+    const main = numberOf(match.groups?.["n"]);
+    if (main === undefined) return [];
+    const { parts, end } = subdivisions(text, match.index + match[0].length);
+    const document = citedDocumentAfter(text, end);
+    const attrs = { target: [main, ...parts].join("."), label: text.slice(match.index, end), ...(document === undefined ? {} : { document }) };
+    return [{ start: match.index, end, attrs }];
+  });
+
+/** Longest first, and never inside a word: "shall not" is not also "shall", "mayor" is not "may". */
+const MARKERS: readonly (readonly [string, "must" | "must-not" | "may"])[] = [
+  ["is required to", "must"],
+  ["shall not", "must-not"],
+  ["must not", "must-not"],
+  ["may not", "must-not"],
+  ["agrees to", "must"],
+  ["shall", "must"],
+  ["must", "must"],
+  ["may", "may"],
+];
+
+const isWordChar = (char: string | undefined): boolean => char !== undefined && /[\p{L}\p{N}_]/u.test(char);
+
+/** Every whole-word occurrence. A loop, not recursion: a line with thousands of "may" must not exhaust the stack. */
+const wordAt = (lower: string, word: string): number[] => {
+  const found: number[] = [];
+  for (let at = lower.indexOf(word); at !== -1; at = lower.indexOf(word, at + 1)) {
+    if (!isWordChar(lower[at - 1]) && !isWordChar(lower[at + word.length])) found.push(at);
+  }
+  return found;
+};
+
+const SENTENCE_OPENERS = new Set([".", "!", "?", ":", ";", '"', "“", "("]);
+
+/** "May" in the middle of a sentence is the month: "published in May 2023". At the start it can be the modal. */
+const isMonthName = (text: string, at: number): boolean => {
+  if (!text.startsWith("May", at)) return false;
+  const before = text.slice(Math.max(0, at - 4), at).trimEnd();
+  return before !== "" && !SENTENCE_OPENERS.has(before.at(-1) ?? "");
+};
+
+/**
+ * Longest marker first; a hit that overlaps one already kept is dropped. Overlap is checked with a mark
+ * per character, not by comparing every pair, which would go quadratic on a line with thousands of markers.
+ */
+const obligations = (text: string): Mention[] => {
+  const lower = text.toLowerCase();
+  const taken = new Uint8Array(text.length);
+  const kept: Mention[] = [];
+  MARKERS.forEach(([marker, type]) => {
+    wordAt(lower, marker).forEach((start) => {
+      const end = start + marker.length;
+      if (isMonthName(text, start)) return;
+      if (taken.subarray(start, end).some((mark) => mark === 1)) return;
+      taken.fill(1, start, end);
+      kept.push({ start, end, attrs: { marker, type } });
+    });
+  });
+  return kept.sort((left, right) => left.start - right.start);
+};
+
+/** Find the number first, then look at what is right before (a currency) and right after (a unit). */
+const NUMBER_RUN = /\d[\d,.]{0,15}/gu;
+const UNITS = [
+  "business days",
+  "business day",
+  "calendar days",
+  "calendar day",
+  "days",
+  "day",
+  "weeks",
+  "week",
+  "months",
+  "month",
+  "years",
+  "year",
+  "hours",
+  "hour",
+  "minutes",
+  "minute",
+  "percent",
+  "times",
+  "%",
+];
+const CURRENCIES = ["USD", "EUR", "$", "€", "£"];
+
+/** "30." の終わりの点は文の終わり。数の一部にしない。 */
+const withoutTrailingPunctuation = (run: string): string => {
+  let end = run.length;
+  while (end > 0 && (run[end - 1] === "." || run[end - 1] === ",")) end -= 1;
+  return run.slice(0, end);
+};
+
+/** One space or tab may sit between a number and its unit or currency. */
+const isGap = (char: string | undefined): boolean => char === " " || char === "\t";
+
+const unitAfter = (text: string, from: number): string | undefined => {
+  const gap = isGap(text[from]) ? 1 : 0;
+  const unit = UNITS.find((candidate) => text.startsWith(candidate, from + gap) && !isWordChar(text[from + gap + candidate.length]));
+  return unit;
+};
+
+/** The currency just before the number, allowing one space. Looks back a few characters, never at the whole line. */
+const currencyBefore = (text: string, at: number): string | undefined => {
+  const end = isGap(text[at - 1]) ? at - 1 : at;
+  return CURRENCIES.find((currency) => text.startsWith(currency, end - currency.length));
+};
+
+const quantities = (text: string): Mention[] =>
+  [...text.matchAll(NUMBER_RUN)].flatMap((match) => {
+    const digits = withoutTrailingPunctuation(match[0]);
+    const value = Number(digits.replace(/,/gu, ""));
+    const end = match.index + digits.length;
+    const unit = unitAfter(text, end) ?? currencyBefore(text, match.index);
+    return unit === undefined || Number.isNaN(value) || isWordChar(text[match.index - 1]) ? [] : [{ start: match.index, end, attrs: { value, unit } }];
+  });
+
+const MONTHS = ["january", "february", "march", "april", "may", "june", "july", "august", "september", "october", "november", "december"];
+const MONTH_WORD = /\b(?<month>[A-Z][a-z]{2,8})\b/gu;
+const ISO_DATE = /\b(?<y>\d{4})-(?<m>\d{2})-(?<d>\d{2})\b/gu;
+const DAY_BEFORE = /(?<d>\d{1,2})(?:st|nd|rd|th)? $/u;
+const DAY_YEAR_AFTER = /^ (?<d>\d{1,2})(?:st|nd|rd|th)?,? (?<y>\d{4})\b/u;
+const YEAR_AFTER = /^,? (?<y>\d{4})\b/u;
+
+const pad = (value: string): string => value.padStart(2, "0");
+
+/** "1 April 2024": the day written before the month. */
+const dayBefore = (text: string, at: number): { readonly day: string; readonly start: number } | undefined => {
+  const found = DAY_BEFORE.exec(text.slice(Math.max(0, at - 6), at));
+  const day = found?.groups?.["d"];
+  return found === null || day === undefined ? undefined : { day, start: at - found[0].length };
+};
+
+/** "April 1, 2024" → 2024-04-01. */
+const monthDayYear = (text: string, at: number, end: number, month: number): Mention | undefined => {
+  const found = DAY_YEAR_AFTER.exec(text.slice(end, end + 16));
+  if (found?.groups === undefined) return undefined;
+  const value = `${found.groups["y"] ?? ""}-${pad(String(month))}-${pad(found.groups["d"] ?? "")}`;
+  return { start: at, end: end + found[0].length, attrs: { value } };
+};
+
+/** "1 April 2024" → 2024-04-01, "April 2024" → 2024-04. */
+const monthYear = (text: string, at: number, end: number, month: number): Mention | undefined => {
+  const found = YEAR_AFTER.exec(text.slice(end, end + 8));
+  const year = found?.groups?.["y"];
+  if (found === null || year === undefined) return undefined;
+  const before = dayBefore(text, at);
+  const value = [year, pad(String(month)), ...(before === undefined ? [] : [pad(before.day)])].join("-");
+  return { start: before?.start ?? at, end: end + found[0].length, attrs: { value } };
+};
+
+/**
+ * A month name alone is not a date: "May" is also the modal verb, so it counts only with a year beside it.
+ * The month is found first and its neighbours read with anchored patterns, never one long alternation.
+ */
+const namedDate = (text: string, match: RegExpExecArray): Mention | undefined => {
+  const month = MONTHS.indexOf((match.groups?.["month"] ?? "").toLowerCase()) + 1;
+  if (month === 0) return undefined;
+  const end = match.index + match[0].length;
+  return monthDayYear(text, match.index, end, month) ?? monthYear(text, match.index, end, month);
+};
+
+const isoDate = (match: RegExpExecArray): Mention => ({
+  start: match.index,
+  end: match.index + match[0].length,
+  attrs: { value: `${match.groups?.["y"] ?? ""}-${match.groups?.["m"] ?? ""}-${match.groups?.["d"] ?? ""}` },
+});
+
+const dates = (text: string): Mention[] =>
+  [...[...text.matchAll(MONTH_WORD)].flatMap((match) => namedDate(text, match) ?? []), ...[...text.matchAll(ISO_DATE)].map(isoDate)].sort(
+    (left, right) => left.start - right.start,
+  );
+
+/** "2.5 days" and "1.5 times" are amounts, not section 2.5 titled "days". */
+const countedAfter = (_number: string, rest: string): boolean => unitAfter(` ${rest}`, 0) !== undefined;
+
+export const structure: StructurePatterns = { numbered, definitions, references, obligations, quantities, dates, countedAfter };