@chaffjs/lang-en 0.5.0 → 0.7.0

package/dist/structure.js

@@ -0,0 +1,359 @@
+import { citedDocumentAfter } from "./citation.js";
+import { membersAfter } from "./reference-list.js";
+import { parseRoman } from "./roman.js";
+import { definitionScopeDepth, definitions, opensDefinitionScope } from "./definitions.js";
+import { CHAPTER_DEPTH, PART_DEPTH } from "./depth.js";
+// Contracts, specifications and statutes in English. core nests what this reads; it does not know
+// how English numbers its articles.
+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>.*)$/u;
+/**
+ * An amendment inserts a subsection between two others and numbers it "(A1)" or "(2A)". It is a subsection, written
+ * outside the sequence: it has no ordinal, so "(1)" after "(A1)" is still the first.
+ */
+const INSERTED = "\\d{1,3}[A-Z]{1,2}|[A-Z]{1,2}\\d{1,3}";
+const LETTERED = new RegExp(`^\\s{0,6}\\((?<n>[a-z]{1,4}|\\d{1,3}|${INSERTED})\\)\\s+(?<rest>\\S.*)$`, "u");
+const IS_INSERTED = new RegExp(`^(?:${INSERTED})$`, "u");
+const MULTI_ROMAN = /^(?:ii|iii|iv|vi|vii|viii|ix)$/u;
+const headed = (pattern, line, numbering, 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)),
+        numbering,
+    };
+};
+/**
+ * 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 = /^\((?<n>[A-Za-z0-9]{1,5})\)$/u.exec(open.label)?.groups?.["n"];
+    if (inner === undefined)
+        return undefined;
+    if (/^\d+$/u.test(inner) || IS_INSERTED.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) || IS_INSERTED.test(raw))
+        return "digit";
+    if (MULTI_ROMAN.test(raw))
+        return "roman";
+    if (!/^[ivx]$/u.test(raw) || followsLetter(raw, context))
+        return "letter";
+    const open = styles(context);
+    return open.includes("roman") || open.at(-1) === "letter" ? "roman" : "letter";
+};
+/**
+ * A sibling has the depth of the open item written the same way: "(b)" closes "(i)" and sits beside "(a)".
+ * A new way of numbering goes one deeper than whatever is open.
+ */
+const depthFor = (style, context) => {
+    const open = styles(context);
+    const sibling = [...context.open].reverse().find((_item, reversed) => open[context.open.length - 1 - reversed] === style);
+    return sibling?.depth ?? (context.open.at(-1)?.depth ?? 0) + 1;
+};
+const LETTER_BEFORE_A = "a".charCodeAt(0) - 1;
+/** "(b)" は 2 番目、"(ii)" も 2 番目。二文字以上の英字（"(aa)"）は並びが決まらないので付けない。 */
+const ordinalOf = (raw, style) => {
+    if (style === "digit")
+        return IS_INSERTED.test(raw) ? undefined : Number(raw);
+    if (style === "roman")
+        return parseRoman(raw);
+    return raw.length === 1 ? raw.charCodeAt(0) - LETTER_BEFORE_A : undefined;
+};
+const lettered = (line, context) => {
+    const groups = LETTERED.exec(line)?.groups;
+    const raw = groups?.["n"];
+    if (groups === undefined || raw === undefined)
+        return undefined;
+    const style = styleOf(raw, context);
+    // 番地は書かれたままの "ii" を使う。参照「Section 4.2(a)(ii)」も同じ形で書かれるので、そのまま引ける。
+    return {
+        kind: "item",
+        depth: depthFor(style, context),
+        ordinal: ordinalOf(raw, style),
+        number: raw,
+        absolute: false,
+        label: `(${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, PART_DEPTH, "pt", "Part") ??
+    chapter(CHAPTER, line, CHAPTER_DEPTH, "ch", "Chapter") ??
+    headed(ARTICLE, line, "article", (n) => `Article ${n}`) ??
+    headed(SECTION, line, "section", (n) => `Section ${n}`) ??
+    lettered(line, context);
+const REFERENCE = /(?<word>\b[Ss]ections?|\b[Aa]rticles?|§) ?(?<n>\d{1,3}(?:\.\d{1,3}){0,5}|[IVXLC]{1,7})\b/gu;
+/** "(a)", "(ii)", "(3)", and an inserted "(A1)" or "(2A)": the same labels the tree reads. */
+const SUBDIVISION = new RegExp(`^\\((?<p>[a-z0-9]{1,4}|${INSERTED})\\)`, "u");
+/** The longest label, with its parentheses: "(ZZ999)". */
+const MAX_SUBDIVISION_LENGTH = 7;
+/** "(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 + MAX_SUBDIVISION_LENGTH))?.groups?.["p"];
+        if (part === undefined)
+            break;
+        parts.push(part);
+        end += part.length + 2;
+    }
+    return { parts, end };
+};
+/** Parentheses opened and not yet closed in `between`. */
+const PAREN_STEP = { "(": 1, ")": -1 };
+const advance = (gloss, text, to) => {
+    for (let index = gloss.scanned; index < to; index += 1) {
+        gloss.depth += PAREN_STEP[text[index] ?? ""] ?? 0;
+        while ((gloss.anchors.at(-1)?.depth ?? -Infinity) > gloss.depth)
+            gloss.anchors.pop();
+    }
+    gloss.scanned = Math.max(gloss.scanned, to);
+};
+const glossedDocument = (gloss) => {
+    const anchor = gloss.anchors.at(-1);
+    return anchor !== undefined && gloss.depth > anchor.depth ? anchor.document : undefined;
+};
+/**
+ * "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) => {
+    const gloss = { depth: 0, scanned: 0, anchors: [] };
+    return [...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);
+        advance(gloss, text, match.index);
+        const cited = citedDocumentAfter(text, end);
+        const document = cited ?? glossedDocument(gloss);
+        gloss.scanned = Math.max(gloss.scanned, end);
+        if (cited !== undefined)
+            gloss.anchors.push({ document: cited, depth: gloss.depth });
+        const numbering = /^[Aa]/u.test(match.groups?.["word"] ?? "") ? "article" : "section";
+        const shared = { numbering, ...(document === undefined ? {} : { document }) };
+        const first = { start: match.index, end, attrs: { target: [main, ...parts].join("."), label: text.slice(match.index, end), ...shared } };
+        const [plural, roman] = [/s$/u.test(match.groups?.["word"] ?? ""), /^[IVXLC]+$/u.test(match.groups?.["n"] ?? "")];
+        return [first, ...membersAfter(text, end, [main, ...parts], shared, plural, roman)];
+    });
+};
+/** 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,
+    opensDefinitionScope,
+    definitionScopeDepth,
+};
+//# 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;AACnD,OAAO,EAAE,YAAY,EAAE,MAAM,qBAAqB,CAAC;AACnD,OAAO,EAAE,UAAU,EAAE,MAAM,YAAY,CAAC;AACxC,OAAO,EAAE,oBAAoB,EAAE,WAAW,EAAE,oBAAoB,EAAE,MAAM,kBAAkB,CAAC;AAC3F,OAAO,EAAE,aAAa,EAAE,UAAU,EAAE,MAAM,YAAY,CAAC;AAEvD,kGAAkG;AAClG,oCAAoC;AAEpC,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;;;GAGG;AACH,MAAM,QAAQ,GAAG,uCAAuC,CAAC;AACzD,MAAM,QAAQ,GAAG,IAAI,MAAM,CAAC,wCAAwC,QAAQ,yBAAyB,EAAE,GAAG,CAAC,CAAC;AAC5G,MAAM,WAAW,GAAG,IAAI,MAAM,CAAC,OAAO,QAAQ,IAAI,EAAE,GAAG,CAAC,CAAC;AACzD,MAAM,WAAW,GAAG,iCAAiC,CAAC;AAEtD,MAAM,MAAM,GAAG,CAAC,OAAe,EAAE,IAAY,EAAE,SAAiB,EAAE,KAA4B,EAA4B,EAAE;IAC1H,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;QAC7B,SAAS;KACV,CAAC;AACJ,CAAC,CAAC;AAIF;;;GAGG;AACH,MAAM,YAAY,GAAG,CAAC,IAAkB,EAAE,KAAa,EAAE,GAA4B,EAAqB,EAAE;IAC1G,MAAM,KAAK,GAAG,+BAA+B,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC;IAC9E,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,SAAS,CAAC;IAC1C,IAAI,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC;QAAE,OAAO,OAAO,CAAC;IACpE,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,IAAI,WAAW,CAAC,IAAI,CAAC,GAAG,CAAC;QAAE,OAAO,OAAO,CAAC;IAChE,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,WAAW,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;IAC9E,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,UAAU,EAAE,IAAI,EAAE,MAAM,CAAC;IAC7C,OAAO,CAAC,OAAO,EAAE,IAAI,EAAE,aAAa,EAAE,IAAI,EAAE,SAAS,CAAC;IACtD,MAAM,CAAC,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,WAAW,CAAC,EAAE,CAAC;IACvD,MAAM,CAAC,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,CAAC,CAAC,EAAE,EAAE,CAAC,WAAW,CAAC,EAAE,CAAC;IACvD,QAAQ,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;AAE1B,MAAM,SAAS,GAAG,4FAA4F,CAAC;AAC/G,8FAA8F;AAC9F,MAAM,WAAW,GAAG,IAAI,MAAM,CAAC,0BAA0B,QAAQ,MAAM,EAAE,GAAG,CAAC,CAAC;AAC9E,0DAA0D;AAC1D,MAAM,sBAAsB,GAAG,CAAC,CAAC;AAEjC,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,sBAAsB,CAAC,CAAC,EAAE,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC;QAC5F,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,0DAA0D;AAC1D,MAAM,UAAU,GAAqC,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,CAAC,EAAE,CAAC;AASzE,MAAM,OAAO,GAAG,CAAC,KAAY,EAAE,IAAY,EAAE,EAAU,EAAQ,EAAE;IAC/D,KAAK,IAAI,KAAK,GAAG,KAAK,CAAC,OAAO,EAAE,KAAK,GAAG,EAAE,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QACvD,KAAK,CAAC,KAAK,IAAI,UAAU,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,CAAC,IAAI,CAAC,CAAC;QAClD,OAAO,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,KAAK,IAAI,CAAC,QAAQ,CAAC,GAAG,KAAK,CAAC,KAAK;YAAE,KAAK,CAAC,OAAO,CAAC,GAAG,EAAE,CAAC;IACvF,CAAC;IACD,KAAK,CAAC,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC;AAC9C,CAAC,CAAC;AAEF,MAAM,eAAe,GAAG,CAAC,KAAY,EAAsB,EAAE;IAC3D,MAAM,MAAM,GAAG,KAAK,CAAC,OAAO,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IACpC,OAAO,MAAM,KAAK,SAAS,IAAI,KAAK,CAAC,KAAK,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC,SAAS,CAAC;AAC1F,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,UAAU,GAAG,CAAC,IAAY,EAAa,EAAE;IAC7C,MAAM,KAAK,GAAU,EAAE,KAAK,EAAE,CAAC,EAAE,OAAO,EAAE,CAAC,EAAE,OAAO,EAAE,EAAE,EAAE,CAAC;IAC3D,OAAO,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,KAAK,EAAE,EAAE;QACrD,MAAM,IAAI,GAAG,QAAQ,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;QAC3C,IAAI,IAAI,KAAK,SAAS;YAAE,OAAO,EAAE,CAAC;QAClC,MAAM,EAAE,KAAK,EAAE,GAAG,EAAE,GAAG,YAAY,CAAC,IAAI,EAAE,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC;QACzE,OAAO,CAAC,KAAK,EAAE,IAAI,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC;QAClC,MAAM,KAAK,GAAG,kBAAkB,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;QAC5C,MAAM,QAAQ,GAAG,KAAK,IAAI,eAAe,CAAC,KAAK,CAAC,CAAC;QACjD,KAAK,CAAC,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,EAAE,GAAG,CAAC,CAAC;QAC7C,IAAI,KAAK,KAAK,SAAS;YAAE,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,QAAQ,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC;QACrF,MAAM,SAAS,GAAG,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC;QACtF,MAAM,MAAM,GAAG,EAAE,SAAS,EAAE,GAAG,CAAC,QAAQ,KAAK,SAAS,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,CAAC,EAAE,CAAC;QAC9E,MAAM,KAAK,GAAG,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,GAAG,EAAE,KAAK,EAAE,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,MAAM,EAAE,EAAE,CAAC;QACzI,MAAM,CAAC,MAAM,EAAE,KAAK,CAAC,GAAG,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,EAAE,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;QAClH,OAAO,CAAC,KAAK,EAAE,GAAG,YAAY,CAAC,IAAI,EAAE,GAAG,EAAE,CAAC,IAAI,EAAE,GAAG,KAAK,CAAC,EAAE,MAAM,EAAE,MAAM,EAAE,KAAK,CAAC,CAAC,CAAC;IACtF,CAAC,CAAC,CAAC;AACL,CAAC,CAAC;AAEF,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;IAC1C,QAAQ;IACR,WAAW;IACX,UAAU;IACV,WAAW;IACX,UAAU;IACV,KAAK;IACL,YAAY;IACZ,oBAAoB;IACpB,oBAAoB;CACrB,CAAC"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@chaffjs/lang-en",
-  "version": "0.5.0",
+  "version": "0.7.0",
   "description": "English language adapter for chaff",
   "license": "MIT",
   "author": "isamu",

package/src/citation.ts

@@ -0,0 +1,85 @@
+// A reference into another document: "Section 9 of the Master Agreement" is not this document's Section 9.
+
+const OF = /^,? of (?:the |that |those )?/u;
+/** "section 4(2)(a) (exception to liability …) of the Damages (Scotland) Act 2011": the gloss sits between the number and the name. */
+const GLOSS = /^ \([^()]{1,100}\)/u;
+const CONNECTOR = /^(?:,? (?:to|and|or)|,) /u;
+/** "9", "29(2)", "(g)", and a roman "V" for a list of Articles. The roman numeral must end the word: "VIII", not "Vendor". */
+const LISTED_NUMBERS = [/^\d{1,3}[A-Z]{0,2}(?:\([a-z0-9]{1,4}\))*/u, /^(?:\([a-z0-9]{1,4}\))+/u, /^[IVXLC]{1,7}\b/u];
+
+const listedNumber = (text: string): string | undefined => LISTED_NUMBERS.map((pattern) => pattern.exec(text)?.[0]).find((found) => found !== undefined);
+
+/**
+ * A listed number is a reference only where the list goes on or ends: "Sections 1, 2 and 9 of" or "Section 3 and 4." —
+ * not "Section 3 and 4 days", where the 4 counts days.
+ */
+const MEMBER_END = /^(?:$|[,;.:)]|\s(?:to|and|or|of)\b|\s\()/u;
+
+export type ListMember = { readonly start: number; readonly text: string };
+
+/** Every number listed after a reference, with where it ends. `start` is from the start of `rest`. */
+const listed = (rest: string): (ListMember & { readonly end: number })[] => {
+  const found: (ListMember & { end: number })[] = [];
+  let end = 0;
+  for (;;) {
+    const connector = CONNECTOR.exec(rest.slice(end))?.[0];
+    const number = connector === undefined ? undefined : listedNumber(rest.slice(end + connector.length));
+    if (connector === undefined || number === undefined) return found;
+    const start = end + connector.length;
+    end = start + number.length;
+    found.push({ start, text: number, end });
+  }
+};
+
+/** "Article 58(2)(c) to (g) and (j) of the UK GDPR": the list runs on to the name that governs all of it. */
+const listEnd = (rest: string): number => listed(rest).at(-1)?.end ?? 0;
+
+/**
+ * The members after a reference, "2" and "9" in "Sections 1, 2 and 9". After a plural ("Sections") every listed number
+ * is one; after a singular ("Section 3 and 4 days") only one where the list goes on or ends.
+ */
+export const listMembers = (rest: string, plural: boolean): ListMember[] =>
+  listed(rest)
+    .filter((member) => plural || MEMBER_END.test(rest.slice(member.end)))
+    .map(({ start, text }) => ({ start, text }));
+
+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 + 200);
+  const afterList = listEnd(rest);
+  const listed = afterList + (GLOSS.exec(rest.slice(afterList))?.[0].length ?? 0);
+  const of = OF.exec(rest.slice(listed));
+  if (of === null) return undefined;
+  const words = titleWords(rest.slice(listed + of[0].length));
+  if (words.length === 0) return undefined;
+  const name = words.join(" ");
+  return words.length === 1 && SELF.has(name) ? undefined : name;
+};

package/src/definitions.ts

@@ -0,0 +1,43 @@
+import type { Mention } from "chaffjs/plugin";
+import { CHAPTER_DEPTH, PART_DEPTH } from "./depth.ts";
+
+// Definitions in English documents, and how far a definition holds.
+
+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,
+];
+
+/**
+ * "In this Part—" and "This section applies where a person ("the seller")…": a statute defines the same word again
+ * in the next Part. A Part's definitions are compared within the Part, a section's within the section.
+ */
+const DEFINITION_SCOPE = /\b(?:In this (?:section|subsection|Part|Chapter|Schedule|Article)\b|This (?:section|Part|Chapter) defines\b)/u;
+/** "This section applies where a person ("the seller") …" names a party in parentheses for this section only. */
+const APPLIES = /\bThis section applies\b/u;
+/** "In this Part" and "This Part defines" hold in the Part; "In this Chapter" in the Chapter; the rest in the section. */
+const WIDER_SCOPE: readonly (readonly [RegExp, number])[] = [
+  [/\b(?:In this|This) Part\b/u, PART_DEPTH],
+  [/\b(?:In this|This) Chapter\b/u, CHAPTER_DEPTH],
+];
+
+export const opensDefinitionScope = (text: string): boolean => DEFINITION_SCOPE.test(text);
+
+export const definitionScopeDepth = (text: string): number | undefined =>
+  DEFINITION_SCOPE.test(text) ? WIDER_SCOPE.find(([pattern]) => pattern.test(text))?.[1] : undefined;
+
+/** "has the meaning given in section 3" points at a definition elsewhere instead of making one. */
+const POINTER = /^ has the meaning given (?:in|by)\b/u;
+const isPointer = (text: string, end: number): boolean => POINTER.test(text.slice(end - " has the meaning".length));
+
+export const definitions = (text: string): Mention[] =>
+  DEFINITIONS.flatMap((pattern) =>
+    [...text.matchAll(pattern)].flatMap((match) => {
+      const term = match.groups?.["term"];
+      if (term === undefined) return [];
+      const end = match.index + match[0].length;
+      const namesAParty = match[0].startsWith("(") && APPLIES.test(text);
+      return [{ start: match.index, end, attrs: { term, ...(isPointer(text, end) || namesAParty ? { scope: "local" } : {}) } }];
+    }),
+  );

package/src/depth.ts

@@ -0,0 +1,3 @@
+/** Parts hold chapters, chapters hold sections (depth 1). The tree nests by these. */
+export const PART_DEPTH = -2;
+export const CHAPTER_DEPTH = -1;

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/reference-list.ts

@@ -0,0 +1,66 @@
+import type { Mention } from "chaffjs/plugin";
+import { listMembers } from "./citation.ts";
+import { parseRoman } from "./roman.ts";
+
+// The later members of a reference list: "Sections 1, 2 and 9" names Section 9 too.
+
+/**
+ * A member is a number, as in the reference itself ("45A" is not read there either), a roman numeral in a list that
+ * began with one ("Articles IV, V and VI"), or only parentheses.
+ */
+const MEMBER = /^(?:(?<main>\d{1,3})|(?<roman>[IVXLC]{1,7}))?(?<parens>(?:\([a-z0-9]{1,4}\))*)$/u;
+const PAREN = /\(([a-z0-9]{1,4})\)/gu;
+
+const isDigits = (part: string): boolean => /^\d+$/u.test(part);
+
+/**
+ * The part of the address before that a parenthesised member stands beside: the last one written the same way
+ * (digits with digits, letters with letters). 45.3.b and (5) → 45.5; 58.2.c and (g) → 58.2.g. The main number stays.
+ */
+const replaceLast = (previous: readonly string[] | undefined, parens: readonly string[]): string[] | undefined => {
+  const first = parens[0];
+  if (previous === undefined || first === undefined) return undefined;
+  const at = previous.findLastIndex((part, index) => index > 0 && isDigits(part) === isDigits(first));
+  return at === -1 ? undefined : [...previous.slice(0, at), ...parens];
+};
+
+/** The main number of a member: digits as written, a roman numeral as its value, but only in a roman list. */
+const mainOf = (groups: Readonly<Record<string, string | undefined>>, romanList: boolean): string | undefined => {
+  const roman = groups["roman"];
+  if (roman === undefined) return groups["main"];
+  const value = romanList ? parseRoman(roman) : undefined;
+  return value === undefined ? undefined : String(value);
+};
+
+/** A member's address, from the member before it when it is only parentheses. */
+const addressOf = (member: string, previous: readonly string[] | undefined, romanList: boolean): string[] | undefined => {
+  const groups = MEMBER.exec(member)?.groups ?? {};
+  const parens = [...(groups["parens"] ?? "").matchAll(PAREN)].map((part) => part[1] ?? "");
+  const written = groups["main"] ?? groups["roman"];
+  if (written === undefined) return replaceLast(previous, parens);
+  const main = mainOf(groups, romanList);
+  return main === undefined ? undefined : [main, ...parens];
+};
+
+/**
+ * "Sections 1, 2 and 9" → 2 and 9 as references too, with the list's numbering and document (`shared`).
+ * `first` is the address of the reference that opens the list, and `end` where it ends in `text`.
+ */
+export const membersAfter = (
+  text: string,
+  end: number,
+  first: readonly string[],
+  shared: Readonly<Record<string, string>>,
+  plural: boolean,
+  romanList: boolean,
+): Mention[] => {
+  const mentions: Mention[] = [];
+  let previous: readonly string[] | undefined = first;
+  listMembers(text.slice(end), plural).forEach((member) => {
+    previous = addressOf(member.text, previous, romanList);
+    if (previous === undefined) return;
+    const start = end + member.start;
+    mentions.push({ start, end: start + member.text.length, attrs: { target: previous.join("."), label: member.text, ...shared } });
+  });
+  return mentions;
+};

package/src/roman.ts

@@ -0,0 +1,9 @@
+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);
+};