chaffjs 0.28.0 → 0.29.0

package/dist/html/html-reader-only.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"html-reader-only.js","sourceRoot":"","sources":["../../src/html/html-reader-only.ts"],"names":[],"mappings":"AAAA,oHAAoH;AACpH,6GAA6G;AAC7G,OAAO,EACL,UAAU,EACV,YAAY,EACZ,aAAa,EACb,QAAQ,EACR,SAAS,EACT,sBAAsB,EACtB,oBAAoB,GAErB,MAAM,oBAAoB,CAAC;AAE5B,yGAAyG;AACzG,sHAAsH;AACtH,gCAAgC;AAChC,MAAM,iBAAiB,GAAG,8FAA8F,CAAC;AAEzH,MAAM,aAAa,GAAG,yCAAyC,CAAC;AAEhE,MAAM,WAAW,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,oBAAoB,UAAU,oDAAoD,EAAE,IAAI,CAAC,CAAC;AAEnI,wHAAwH;AACxH,MAAM,iBAAiB,GAAG,CAAC,KAAmB,EAAW,EAAE;IACzD,MAAM,KAAK,GAAG,WAAW,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IAC9C,MAAM,KAAK,GAAG,CAAC,KAAK,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,EAAE,CAAC,CAAC,CAAC,IAAI,KAAK,EAAE,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;IAC3E,IAAI,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC;QAAE,OAAO,KAAK,CAAC;IAC3D,OAAO,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,iBAAiB,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,KAAK,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;AACxF,CAAC,CAAC;AAEF,6GAA6G;AAC7G,MAAM,aAAa,GAAG,CAAC,KAAmB,EAAW,EAAE,CACrD,CAAC,cAAc,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,2DAA2D,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;AAEzH,MAAM,SAAS,GAAG;IAChB,GAAG;IACH,KAAK;IACL,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,GAAG,YAAY;IACf,SAAS;IACT,SAAS;IACT,QAAQ;IACR,QAAQ;IACR,MAAM;IACN,MAAM;IACN,YAAY;IACZ,IAAI;IACJ,IAAI;IACJ,IAAI;IACJ,OAAO;IACP,IAAI;CACL,CAAC;AAEF,MAAM,aAAa,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,UAAU,SAAS,CAAC,IAAI,CAAC,GAAG,CAAC,yBAAyB,EAAE,KAAK,CAAC,CAAC;AAE1G,6FAA6F;AAC7F,MAAM,UAAU,GAAG,CAAC,IAAY,EAAE,KAAmB,EAA6B,EAAE;IAClF,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC;IAC1C,MAAM,SAAS,GAAG,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,aAAa,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IAC7D,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IACpC,MAAM,OAAO,GAAG,IAAI,MAAM,CAAC,aAAa,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,KAAK,IAAI,KAAK,CAAC,MAAM,CAAC;IAC1F,OAAO,CAAC,SAAS,KAAK,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,SAAS,CAAC,KAAK,GAAG,SAAS,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC,CAAC;AAC3H,CAAC,CAAC;AAEF,6FAA6F;AAC7F,MAAM,WAAW,GAAG,CAAC,IAAY,EAAE,KAAmB,EAAW,EAAE,CAAC,UAAU,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC,KAAK,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,SAAS,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC,CAAC;AAExI;;;GAGG;AACH,MAAM,CAAC,MAAM,qBAAqB,GAAG,CAAC,IAAY,EAAU,EAAE;IAC5D,MAAM,KAAK,GAAG,sBAAsB,CAAC,IAAI,EAAE,aAAa,EAAE,iBAAiB,CAAC,CAAC;IAC7E,MAAM,QAAQ,GAAG,aAAa,CAAC,KAAK,CAAC,CAAC;IACtC,MAAM,cAAc,GAAG,CAAC,MAAoB,EAAW,EAAE,CACvD,aAAa,CAAC,MAAM,CAAC,IAAI,WAAW,CAAC,KAAK,EAAE,MAAM,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,QAAQ,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC,CAAC;IAChH,OAAO,oBAAoB,CAAC,KAAK,EAAE,GAAG,EAAE,cAAc,CAAC,CAAC;AAC1D,CAAC,CAAC"}

package/dist/html/html-run-in-headings.d.ts

@@ -0,0 +1,2 @@
+export declare const withRunInHeadingsRead: (html: string) => string;
+//# sourceMappingURL=html-run-in-headings.d.ts.map

package/dist/html/html-run-in-headings.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"html-run-in-headings.d.ts","sourceRoot":"","sources":["../../src/html/html-run-in-headings.ts"],"names":[],"mappings":"AAqCA,eAAO,MAAM,qBAAqB,GAAI,MAAM,MAAM,KAAG,MAKjD,CAAC"}

package/dist/html/html-run-in-headings.js

@@ -0,0 +1,36 @@
+// A section title drawn as a bold first line of its paragraph: <p><strong>Conclusion</strong><br />In closing, …</p>.
+// Read as text the title runs into the paragraph as one line. The bold run becomes a heading one level below the
+// heading before it (h2 when none), and the rest stays the paragraph. Only a bold run that opens the paragraph, holds
+// plain text, is short, does not end as a sentence or a label (".", ":") and has a sentence (or a lead-in ending in a
+// colon) after its line break is taken for a title; a name over its role ("Jane Roe<br>Secretary") is a signature. Pure.
+import { hasSentence, plainText } from "./html-elements.js";
+const MAX_TITLE_LENGTH = 100;
+const DEEPEST_LEVEL = 6;
+const TOP_LEVEL = 1;
+// A paragraph's start, a bold run and a line break: the paragraph's text follows up to its end tag.
+const OPENING = /<p\b([^>]*)>\s*<(strong|b)\b[^>]*>([^<]*)<\/\2\s*>\s*<br\b[^>]*>/giu;
+/** The paragraph's text after from, up to its end tag; undefined when another paragraph opens first or none closes. */
+const restOfParagraph = (html, from) => {
+    const next = /<(\/?)p\b/iu.exec(html.slice(from));
+    return next?.[1] === "/" ? html.slice(from, from + next.index) : undefined;
+};
+// A sentence's end, a label's colon, or a clause's comma or semicolon: the bold run is part of the text, not a title.
+const ENDS_AS_TEXT = /[.!?:;,。！？：；、]$/u;
+/** A sentence, or a lead-in to a list ("… with the same:"): what a section holds, unlike a role under a name. */
+const isProse = (html) => hasSentence(html) || /[:：]$/u.test(plainText(html));
+const isTitle = (bold, rest) => {
+    const title = bold.trim();
+    return title !== "" && title.length <= MAX_TITLE_LENGTH && !ENDS_AS_TEXT.test(title) && rest !== undefined && isProse(rest);
+};
+/** The level of the last heading opening before at; 1 when there is none, as if under the page's title. */
+const levelBefore = (html, at) => {
+    const headings = [...html.slice(0, at).matchAll(/<h([1-6])\b/giu)];
+    return Number(headings.at(-1)?.[1] ?? TOP_LEVEL);
+};
+export const withRunInHeadingsRead = (html) => html.replace(OPENING, (opening, attributes, _tag, bold, at) => {
+    if (!isTitle(bold, restOfParagraph(html, at + opening.length)))
+        return opening;
+    const level = String(Math.min(levelBefore(html, at) + 1, DEEPEST_LEVEL));
+    return `<h${level}>${bold.trim()}</h${level}><p${attributes}>`;
+});
+//# sourceMappingURL=html-run-in-headings.js.map

package/dist/html/html-run-in-headings.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"html-run-in-headings.js","sourceRoot":"","sources":["../../src/html/html-run-in-headings.ts"],"names":[],"mappings":"AAAA,sHAAsH;AACtH,iHAAiH;AACjH,sHAAsH;AACtH,sHAAsH;AACtH,yHAAyH;AACzH,OAAO,EAAE,WAAW,EAAE,SAAS,EAAE,MAAM,oBAAoB,CAAC;AAE5D,MAAM,gBAAgB,GAAG,GAAG,CAAC;AAC7B,MAAM,aAAa,GAAG,CAAC,CAAC;AACxB,MAAM,SAAS,GAAG,CAAC,CAAC;AAEpB,oGAAoG;AACpG,MAAM,OAAO,GAAG,qEAAqE,CAAC;AAEtF,uHAAuH;AACvH,MAAM,eAAe,GAAG,CAAC,IAAY,EAAE,IAAY,EAAsB,EAAE;IACzE,MAAM,IAAI,GAAG,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC;IAClD,OAAO,IAAI,EAAE,CAAC,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,EAAE,IAAI,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;AAC7E,CAAC,CAAC;AAEF,sHAAsH;AACtH,MAAM,YAAY,GAAG,kBAAkB,CAAC;AAExC,iHAAiH;AACjH,MAAM,OAAO,GAAG,CAAC,IAAY,EAAW,EAAE,CAAC,WAAW,CAAC,IAAI,CAAC,IAAI,QAAQ,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,CAAC,CAAC;AAE/F,MAAM,OAAO,GAAG,CAAC,IAAY,EAAE,IAAwB,EAAW,EAAE;IAClE,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IAC1B,OAAO,KAAK,KAAK,EAAE,IAAI,KAAK,CAAC,MAAM,IAAI,gBAAgB,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,IAAI,KAAK,SAAS,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC;AAC9H,CAAC,CAAC;AAEF,2GAA2G;AAC3G,MAAM,WAAW,GAAG,CAAC,IAAY,EAAE,EAAU,EAAU,EAAE;IACvD,MAAM,QAAQ,GAAG,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,QAAQ,CAAC,gBAAgB,CAAC,CAAC,CAAC;IACnE,OAAO,MAAM,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,IAAI,SAAS,CAAC,CAAC;AACnD,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,qBAAqB,GAAG,CAAC,IAAY,EAAU,EAAE,CAC5D,IAAI,CAAC,OAAO,CAAC,OAAO,EAAE,CAAC,OAAe,EAAE,UAAkB,EAAE,IAAY,EAAE,IAAY,EAAE,EAAU,EAAE,EAAE;IACpG,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,eAAe,CAAC,IAAI,EAAE,EAAE,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC;QAAE,OAAO,OAAO,CAAC;IAC/E,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,WAAW,CAAC,IAAI,EAAE,EAAE,CAAC,GAAG,CAAC,EAAE,aAAa,CAAC,CAAC,CAAC;IACzE,OAAO,KAAK,KAAK,IAAI,IAAI,CAAC,IAAI,EAAE,MAAM,KAAK,MAAM,UAAU,GAAG,CAAC;AACjE,CAAC,CAAC,CAAC"}

package/dist/html/html-tables.d.ts

@@ -0,0 +1,3 @@
+/** Every table read from the inside out: a table of sentences as blocks, any other table replaced by a space. */
+export declare const withTablesRead: (html: string) => string;
+//# sourceMappingURL=html-tables.d.ts.map

package/dist/html/html-tables.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"html-tables.d.ts","sourceRoot":"","sources":["../../src/html/html-tables.ts"],"names":[],"mappings":"AA0EA,iHAAiH;AACjH,eAAO,MAAM,cAAc,GAAI,MAAM,MAAM,KAAG,MAA+E,CAAC"}

package/dist/html/html-tables.js

@@ -0,0 +1,58 @@
+// Tables as the HTML converter reads them. A table whose rows are sentences (a glossary of terms and what they mean,
+// a list of routes and how each one works) keeps each cell as a block of its own, so its sentences are read as prose;
+// its header rows go. Any other table (figures, a grid of links) goes whole, as before. Nested tables are read from
+// the inside out. Pure; a regular-expression reading, not a parser.
+import { plainText, untilStable } from "./html-elements.js";
+const INNERMOST_TABLE = /<table\b[^>]*>((?:(?!<table\b)[\s\S])*?)<\/table\s*>/giu;
+// A row and a cell may leave out their closing tags: each ends where the next one opens.
+const ROW_OPENING = /<tr\b[^>]*>/iu;
+const ROW_END = /<\/(?:tr|thead|tbody|tfoot)\s*>[\s\S]*$/iu;
+const CELL_OPENING = /<(t[dh])\b[^>]*>/giu;
+const CELL_END = /<\/t[dh]\s*>[\s\S]*$/iu;
+const cellsOf = (row) => {
+    const openings = [...row.matchAll(CELL_OPENING)];
+    return openings.map((opening, index) => {
+        const end = openings[index + 1]?.index ?? row.length;
+        const html = row.slice(opening.index + opening[0].length, end).replace(CELL_END, "");
+        return { header: opening[1]?.toLowerCase() === "th", html };
+    });
+};
+const rowsOf = (table) => table
+    .split(ROW_OPENING)
+    .slice(1)
+    .map((row) => cellsOf(row.replace(ROW_END, "")));
+const hasText = (cell) => plainText(cell.html) !== "";
+// A cell ends a sentence with its last mark. A point inside it ("H.Con.Res. 218") or after a short capitalised word
+// or a letter after a point ("Inc.", "U.S.") is an abbreviation.
+const ENDS_SENTENCE = /(?:[。．！？!?]|(?<!\b[A-Z][A-Za-z]{0,3}|\.[A-Za-z])\.)[)\]"'”’」』）]*$/u;
+const holdsSentence = (cell) => ENDS_SENTENCE.test(plainText(cell.html));
+// A cell that opens with a number and holds no word of the Latin alphabet: a row number ("1", "2."), a figure or a
+// range ("0.3 以上 1.6未満", "１未満"), a section ("302(c)"). "401(k) plan" and "24/7 desk" are names.
+const isNumberLabel = (text) => /^\p{Nd}/u.test(text) && !/[A-Za-z]{2,}/u.test(text);
+/**
+ * A cell kept as a block: one with a sentence, or a label in words. A number label is read by no prose rule, and alone
+ * on a line it would read as a section number.
+ */
+const isKept = (cell) => holdsSentence(cell) || (hasText(cell) && !isNumberLabel(plainText(cell.html)));
+/** A row of nothing but header cells labels the columns. */
+const isHeaderRow = (row) => row.length > 0 && row.every((cell) => cell.header);
+/** At least half of the rows with text in them, header rows aside, hold a sentence. */
+const isProse = (rows) => {
+    const written = rows.filter((row) => !isHeaderRow(row) && row.some(hasText));
+    const sentences = written.filter((row) => row.some(holdsSentence)).length;
+    return sentences > 0 && sentences * 2 >= written.length;
+};
+const CAPTION = /<caption\b[^>]*>([\s\S]*?)<\/caption\s*>/iu;
+/** Each cell with text as a block, the caption first. */
+const asBlocks = (table, rows) => {
+    const caption = CAPTION.exec(table)?.[1];
+    const cells = rows.filter((row) => !isHeaderRow(row)).flatMap((row) => row.filter(isKept).map((cell) => cell.html));
+    return ["<div>", ...[caption ?? "", ...cells].filter((html) => html !== "").map((html) => `<div>${html}</div>`), "</div>"].join("");
+};
+const readTable = (_whole, inside) => {
+    const rows = rowsOf(inside);
+    return isProse(rows) ? asBlocks(inside, rows) : " ";
+};
+/** Every table read from the inside out: a table of sentences as blocks, any other table replaced by a space. */
+export const withTablesRead = (html) => untilStable(html, (text) => text.replace(INNERMOST_TABLE, readTable));
+//# sourceMappingURL=html-tables.js.map

package/dist/html/html-tables.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"html-tables.js","sourceRoot":"","sources":["../../src/html/html-tables.ts"],"names":[],"mappings":"AAAA,qHAAqH;AACrH,sHAAsH;AACtH,oHAAoH;AACpH,oEAAoE;AACpE,OAAO,EAAE,SAAS,EAAE,WAAW,EAAE,MAAM,oBAAoB,CAAC;AAK5D,MAAM,eAAe,GAAG,yDAAyD,CAAC;AAElF,yFAAyF;AACzF,MAAM,WAAW,GAAG,eAAe,CAAC;AACpC,MAAM,OAAO,GAAG,2CAA2C,CAAC;AAC5D,MAAM,YAAY,GAAG,qBAAqB,CAAC;AAC3C,MAAM,QAAQ,GAAG,wBAAwB,CAAC;AAE1C,MAAM,OAAO,GAAG,CAAC,GAAW,EAAO,EAAE;IACnC,MAAM,QAAQ,GAAG,CAAC,GAAG,GAAG,CAAC,QAAQ,CAAC,YAAY,CAAC,CAAC,CAAC;IACjD,OAAO,QAAQ,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,KAAK,EAAE,EAAE;QACrC,MAAM,GAAG,GAAG,QAAQ,CAAC,KAAK,GAAG,CAAC,CAAC,EAAE,KAAK,IAAI,GAAG,CAAC,MAAM,CAAC;QACrD,MAAM,IAAI,GAAG,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,KAAK,GAAG,OAAO,CAAC,CAAC,CAAC,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;QACrF,OAAO,EAAE,MAAM,EAAE,OAAO,CAAC,CAAC,CAAC,EAAE,WAAW,EAAE,KAAK,IAAI,EAAE,IAAI,EAAE,CAAC;IAC9D,CAAC,CAAC,CAAC;AACL,CAAC,CAAC;AAEF,MAAM,MAAM,GAAG,CAAC,KAAa,EAAS,EAAE,CACtC,KAAK;KACF,KAAK,CAAC,WAAW,CAAC;KAClB,KAAK,CAAC,CAAC,CAAC;KACR,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC;AAErD,MAAM,OAAO,GAAG,CAAC,IAAU,EAAW,EAAE,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,EAAE,CAAC;AAErE,oHAAoH;AACpH,iEAAiE;AACjE,MAAM,aAAa,GAAG,oEAAoE,CAAC;AAE3F,MAAM,aAAa,GAAG,CAAC,IAAU,EAAW,EAAE,CAAC,aAAa,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;AAExF,mHAAmH;AACnH,gGAAgG;AAChG,MAAM,aAAa,GAAG,CAAC,IAAY,EAAW,EAAE,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,eAAe,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAEtG;;;GAGG;AACH,MAAM,MAAM,GAAG,CAAC,IAAU,EAAW,EAAE,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,aAAa,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;AAEvH,4DAA4D;AAC5D,MAAM,WAAW,GAAG,CAAC,GAAQ,EAAW,EAAE,CAAC,GAAG,CAAC,MAAM,GAAG,CAAC,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;AAE9F,uFAAuF;AACvF,MAAM,OAAO,GAAG,CAAC,IAAoB,EAAW,EAAE;IAChD,MAAM,OAAO,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CAAC;IAC7E,MAAM,SAAS,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC,CAAC,MAAM,CAAC;IAC1E,OAAO,SAAS,GAAG,CAAC,IAAI,SAAS,GAAG,CAAC,IAAI,OAAO,CAAC,MAAM,CAAC;AAC1D,CAAC,CAAC;AAEF,MAAM,OAAO,GAAG,4CAA4C,CAAC;AAE7D,yDAAyD;AACzD,MAAM,QAAQ,GAAG,CAAC,KAAa,EAAE,IAAoB,EAAU,EAAE;IAC/D,MAAM,OAAO,GAAG,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IACzC,MAAM,KAAK,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,GAAG,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;IACpH,OAAO,CAAC,OAAO,EAAE,GAAG,CAAC,OAAO,IAAI,EAAE,EAAE,GAAG,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,KAAK,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,QAAQ,IAAI,QAAQ,CAAC,EAAE,QAAQ,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AACtI,CAAC,CAAC;AAEF,MAAM,SAAS,GAAG,CAAC,MAAc,EAAE,MAAc,EAAU,EAAE;IAC3D,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,CAAC;IAC5B,OAAO,OAAO,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC;AACtD,CAAC,CAAC;AAEF,iHAAiH;AACjH,MAAM,CAAC,MAAM,cAAc,GAAG,CAAC,IAAY,EAAU,EAAE,CAAC,WAAW,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,eAAe,EAAE,SAAS,CAAC,CAAC,CAAC"}

package/dist/html/html-widgets.d.ts

@@ -0,0 +1,3 @@
+/** Controls and their text: tooltips, unspoken data, media players, scripted link buttons and paging bars. */
+export declare const withoutWidgets: (html: string) => string;
+//# sourceMappingURL=html-widgets.d.ts.map

package/dist/html/html-widgets.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"html-widgets.d.ts","sourceRoot":"","sources":["../../src/html/html-widgets.ts"],"names":[],"mappings":"AAgHA,8GAA8G;AAC9G,eAAO,MAAM,cAAc,GAAI,MAAM,MAAM,KAAG,MAA2G,CAAC"}

package/dist/html/html-widgets.js

@@ -0,0 +1,71 @@
+// Parts of a page that are controls rather than the document: a tooltip, a media player with its buttons, a block a
+// script runs when clicked, a bar of previous and next links between arrows, and text marked never to be spoken.
+// Each is told by its structure (a role, an <audio> or <video>, an event handler, the arrows), never by a site's class
+// names. Pure; a regular-expression reading.
+import { ANY_LINK, ATTRIBUTES, elementRanges, fileExtension, hasNoWords, hasSentence, isFileLink, plainText, withoutElementsOpening, withoutRanges, } from "./html-elements.js";
+/** Every element whose opening tag carries attribute (name and value as a pattern), matched to its own closing tag. */
+const withoutElementsWithAttribute = (html, attribute) => {
+    const opening = new RegExp(String.raw `<([a-z][a-z0-9-]*)${ATTRIBUTES}\s+${attribute}`, "giu");
+    const marked = new RegExp(String.raw `^<[a-z][a-z0-9-]*${ATTRIBUTES}\s+${attribute}`, "iu");
+    return withoutElementsOpening(html, opening, (range) => marked.test(range.openTag));
+};
+// A role is a list of tokens and the first one counts, as for role="main".
+const TOOLTIP_ROLE = String.raw `role\s*=\s*(?:"\s*tooltip(?:\s[^"]*)?"|'\s*tooltip(?:\s[^']*)?'|tooltip(?=[\s/>]))`;
+// speak: none in a style says the text is not to be read out: data for scripts, such as an id beside a title.
+const UNSPOKEN_STYLE = String.raw `style\s*=\s*(?:"[^"]*\bspeak\s*:\s*none\b[^"]*"|'[^']*\bspeak\s*:\s*none\b[^']*')`;
+/** A tooltip ("The code has been copied to your clipboard.") shows beside a control, and unspoken text is not read. */
+const withoutTooltipsAndUnspoken = (html) => withoutElementsWithAttribute(withoutElementsWithAttribute(html, TOOLTIP_ROLE), UNSPOKEN_STYLE);
+const MEDIA_ELEMENT = /<(audio|video)\b[^>]*>[\s\S]*?<\/\1\s*>/giu;
+const PLAYER_BLOCKS = ["div", "section", "figure"];
+// A caption is the document's even when it is not a sentence ("Interview with the mayor"), so its block is no player.
+const holdsMedia = (range) => /<(?:audio|video)\b/iu.test(range.inner) && !/<figcaption\b/iu.test(range.inner);
+/** The text around the media, without the words a browser shows only when it cannot play it. */
+const hasSentenceBesideMedia = (range) => hasSentence(range.inner.replace(MEDIA_ELEMENT, " "));
+const MEDIA_EXTENSION = /^(?:mp3|m4a|aac|wav|ogg|oga|opus|mp4|m4v|webm|mov)$/iu;
+/** A link to a file other than the media itself (a transcript in PDF, an address) is the document's. */
+const isDocumentFile = (link) => isFileLink(link) && !MEDIA_EXTENSION.test(fileExtension(link) ?? "");
+const isPlayer = (range) => holdsMedia(range) && !hasSentenceBesideMedia(range) && ![...range.inner.matchAll(ANY_LINK)].some((link) => isDocumentFile(link[0]));
+/**
+ * A block holding an <audio> or <video> and not a sentence is its player: the controls, the timer, "Embed", "share",
+ * "Direct link" to the media's own file. The outermost such block goes; a block with a sentence, a caption or a link
+ * to another file stays.
+ */
+const withoutMediaPlayers = (html) => {
+    const players = PLAYER_BLOCKS.flatMap((tag) => elementRanges(html, tag)).filter(isPlayer);
+    return withoutRanges(html, players.toSorted((left, right) => left.start - right.start || right.end - left.end));
+};
+const EVENT_HANDLER = String.raw `on[a-z]+\s*=`;
+const CONTROL_OPENING = new RegExp(String.raw `<(div|p|span|li)${ATTRIBUTES}\s+${EVENT_HANDLER}`, "giu");
+/** Links and nothing else, but no link that wraps blocks. */
+const isOnlyPlainLinks = (inner) => {
+    const links = [...inner.matchAll(ANY_LINK)].map((link) => link[0]);
+    return links.length > 0 && !links.some((link) => /<(?:div|p|h[1-6]|ul|ol)\b/iu.test(link)) && hasNoWords(plainText(inner.replace(ANY_LINK, " ")));
+};
+/**
+ * A block that runs a script when clicked (onclick) and holds only a link ("Start Quiz") is a button drawn as a link.
+ * A link with a handler of its own, inside a sentence, stays: it is only counted when followed.
+ */
+const isScriptedControl = (range) => new RegExp(String.raw `^<[a-z]+${ATTRIBUTES}\s+${EVENT_HANDLER}`, "iu").test(range.openTag) && isOnlyPlainLinks(range.inner);
+const withoutScriptedControls = (html) => withoutElementsOpening(html, CONTROL_OPENING, isScriptedControl);
+const BACK_ARROW = "[←‹«◀◁]";
+const FORWARD_ARROW = "[→›»▶▷]";
+const PAGING_TEXT = new RegExp(`^${BACK_ARROW}.*${FORWARD_ARROW}$`, "u");
+const MIN_PAGING_LINKS = 2;
+const PAGING_BLOCKS = ["div", "section", "p", "td"];
+/**
+ * Opens with a back arrow and closes with a forward one around two or more links to pages (never a file), with no sentence beside them: "←
+ * Previous | Title | Next →". Link text is not looked at, since a name's initial ("John P.") is not a sentence.
+ */
+const isPagingBar = (range) => {
+    const links = [...range.inner.matchAll(ANY_LINK)].map((link) => link[0]);
+    return (PAGING_TEXT.test(plainText(range.inner)) && links.length >= MIN_PAGING_LINKS && !links.some(isFileLink) && !hasSentence(range.inner.replace(ANY_LINK, " ")));
+};
+/** The innermost blocks that are a bar of previous and next links, as a work's header draws them over each chapter. */
+const withoutPagingBars = (html) => {
+    const bars = PAGING_BLOCKS.flatMap((tag) => elementRanges(html, tag)).filter(isPagingBar);
+    const innermost = bars.filter((outer) => !bars.some((inner) => inner !== outer && outer.start <= inner.start && inner.end <= outer.end));
+    return withoutRanges(html, innermost.toSorted((left, right) => left.start - right.start));
+};
+/** Controls and their text: tooltips, unspoken data, media players, scripted link buttons and paging bars. */
+export const withoutWidgets = (html) => withoutPagingBars(withoutScriptedControls(withoutMediaPlayers(withoutTooltipsAndUnspoken(html))));
+//# sourceMappingURL=html-widgets.js.map

package/dist/html/html-widgets.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"html-widgets.js","sourceRoot":"","sources":["../../src/html/html-widgets.ts"],"names":[],"mappings":"AAAA,oHAAoH;AACpH,iHAAiH;AACjH,uHAAuH;AACvH,6CAA6C;AAC7C,OAAO,EACL,QAAQ,EACR,UAAU,EACV,aAAa,EACb,aAAa,EACb,UAAU,EACV,WAAW,EACX,UAAU,EACV,SAAS,EACT,sBAAsB,EACtB,aAAa,GAEd,MAAM,oBAAoB,CAAC;AAE5B,uHAAuH;AACvH,MAAM,4BAA4B,GAAG,CAAC,IAAY,EAAE,SAAiB,EAAU,EAAE;IAC/E,MAAM,OAAO,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,qBAAqB,UAAU,MAAM,SAAS,EAAE,EAAE,KAAK,CAAC,CAAC;IAC9F,MAAM,MAAM,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,oBAAoB,UAAU,MAAM,SAAS,EAAE,EAAE,IAAI,CAAC,CAAC;IAC3F,OAAO,sBAAsB,CAAC,IAAI,EAAE,OAAO,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC;AACtF,CAAC,CAAC;AAEF,2EAA2E;AAC3E,MAAM,YAAY,GAAG,MAAM,CAAC,GAAG,CAAA,oFAAoF,CAAC;AAEpH,8GAA8G;AAC9G,MAAM,cAAc,GAAG,MAAM,CAAC,GAAG,CAAA,mFAAmF,CAAC;AAErH,uHAAuH;AACvH,MAAM,0BAA0B,GAAG,CAAC,IAAY,EAAU,EAAE,CAAC,4BAA4B,CAAC,4BAA4B,CAAC,IAAI,EAAE,YAAY,CAAC,EAAE,cAAc,CAAC,CAAC;AAE5J,MAAM,aAAa,GAAG,4CAA4C,CAAC;AAEnE,MAAM,aAAa,GAAG,CAAC,KAAK,EAAE,SAAS,EAAE,QAAQ,CAAC,CAAC;AAEnD,sHAAsH;AACtH,MAAM,UAAU,GAAG,CAAC,KAAmB,EAAW,EAAE,CAAC,sBAAsB,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,IAAI,CAAC,iBAAiB,CAAC,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;AAEtI,gGAAgG;AAChG,MAAM,sBAAsB,GAAG,CAAC,KAAmB,EAAW,EAAE,CAAC,WAAW,CAAC,KAAK,CAAC,KAAK,CAAC,OAAO,CAAC,aAAa,EAAE,GAAG,CAAC,CAAC,CAAC;AAEtH,MAAM,eAAe,GAAG,uDAAuD,CAAC;AAEhF,wGAAwG;AACxG,MAAM,cAAc,GAAG,CAAC,IAAY,EAAW,EAAE,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,CAAC,eAAe,CAAC,IAAI,CAAC,aAAa,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC;AAEvH,MAAM,QAAQ,GAAG,CAAC,KAAmB,EAAW,EAAE,CAChD,UAAU,CAAC,KAAK,CAAC,IAAI,CAAC,sBAAsB,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,KAAK,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,cAAc,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;AAEtI;;;;GAIG;AACH,MAAM,mBAAmB,GAAG,CAAC,IAAY,EAAU,EAAE;IACnD,MAAM,OAAO,GAAG,aAAa,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,aAAa,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;IAC1F,OAAO,aAAa,CAClB,IAAI,EACJ,OAAO,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,KAAK,IAAI,KAAK,CAAC,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,CACpF,CAAC;AACJ,CAAC,CAAC;AAEF,MAAM,aAAa,GAAG,MAAM,CAAC,GAAG,CAAA,cAAc,CAAC;AAE/C,MAAM,eAAe,GAAG,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,mBAAmB,UAAU,MAAM,aAAa,EAAE,EAAE,KAAK,CAAC,CAAC;AAExG,6DAA6D;AAC7D,MAAM,gBAAgB,GAAG,CAAC,KAAa,EAAW,EAAE;IAClD,MAAM,KAAK,GAAG,CAAC,GAAG,KAAK,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;IACnE,OAAO,KAAK,CAAC,MAAM,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,6BAA6B,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,UAAU,CAAC,SAAS,CAAC,KAAK,CAAC,OAAO,CAAC,QAAQ,EAAE,GAAG,CAAC,CAAC,CAAC,CAAC;AACpJ,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,iBAAiB,GAAG,CAAC,KAAmB,EAAW,EAAE,CACzD,IAAI,MAAM,CAAC,MAAM,CAAC,GAAG,CAAA,WAAW,UAAU,MAAM,aAAa,EAAE,EAAE,IAAI,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,gBAAgB,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;AAE9H,MAAM,uBAAuB,GAAG,CAAC,IAAY,EAAU,EAAE,CAAC,sBAAsB,CAAC,IAAI,EAAE,eAAe,EAAE,iBAAiB,CAAC,CAAC;AAE3H,MAAM,UAAU,GAAG,SAAS,CAAC;AAC7B,MAAM,aAAa,GAAG,SAAS,CAAC;AAChC,MAAM,WAAW,GAAG,IAAI,MAAM,CAAC,IAAI,UAAU,KAAK,aAAa,GAAG,EAAE,GAAG,CAAC,CAAC;AACzE,MAAM,gBAAgB,GAAG,CAAC,CAAC;AAE3B,MAAM,aAAa,GAAG,CAAC,KAAK,EAAE,SAAS,EAAE,GAAG,EAAE,IAAI,CAAC,CAAC;AAEpD;;;GAGG;AACH,MAAM,WAAW,GAAG,CAAC,KAAmB,EAAW,EAAE;IACnD,MAAM,KAAK,GAAG,CAAC,GAAG,KAAK,CAAC,KAAK,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC;IACzE,OAAO,CACL,WAAW,CAAC,IAAI,CAAC,SAAS,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,IAAI,KAAK,CAAC,MAAM,IAAI,gBAAgB,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,WAAW,CAAC,KAAK,CAAC,KAAK,CAAC,OAAO,CAAC,QAAQ,EAAE,GAAG,CAAC,CAAC,CAC5J,CAAC;AACJ,CAAC,CAAC;AAEF,uHAAuH;AACvH,MAAM,iBAAiB,GAAG,CAAC,IAAY,EAAU,EAAE;IACjD,MAAM,IAAI,GAAG,aAAa,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,aAAa,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC,CAAC,MAAM,CAAC,WAAW,CAAC,CAAC;IAC1F,MAAM,SAAS,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,KAAK,KAAK,IAAI,KAAK,CAAC,KAAK,IAAI,KAAK,CAAC,KAAK,IAAI,KAAK,CAAC,GAAG,IAAI,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC;IACzI,OAAO,aAAa,CAClB,IAAI,EACJ,SAAS,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC,CAC9D,CAAC;AACJ,CAAC,CAAC;AAEF,8GAA8G;AAC9G,MAAM,CAAC,MAAM,cAAc,GAAG,CAAC,IAAY,EAAU,EAAE,CAAC,iBAAiB,CAAC,uBAAuB,CAAC,mBAAmB,CAAC,0BAA0B,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,CAAC"}

package/dist/html/markup-text.d.ts

@@ -0,0 +1,5 @@
+/** In one pass, so "&amp;lt;" is "&lt;" and not "<". An unknown name is left as written. */
+export declare const decodeEntities: (text: string) => string;
+/** Trailing spaces removed, runs of blank lines made one, and a single newline at the end. */
+export declare const tidyLines: (lines: readonly string[]) => string;
+//# sourceMappingURL=markup-text.d.ts.map

package/dist/html/markup-text.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"markup-text.d.ts","sourceRoot":"","sources":["../../src/html/markup-text.ts"],"names":[],"mappings":"AAsBA,4FAA4F;AAC5F,eAAO,MAAM,cAAc,GAAI,MAAM,MAAM,KAAG,MACmE,CAAC;AAElH,8FAA8F;AAC9F,eAAO,MAAM,SAAS,GAAI,OAAO,SAAS,MAAM,EAAE,KAAG,MAOpD,CAAC"}

package/dist/html/markup-text.js

@@ -0,0 +1,32 @@
+// What the wikitext and HTML converters share: character references turned back into characters, and the blank
+// lines and trailing spaces a converter leaves behind tidied away. Pure.
+import { characterEntities } from "character-entities";
+// A named space ( ,  ) is read as a plain space, like the spacing between the page's other words.
+const isSpace = (value) => /^\p{Zs}+$/u.test(value);
+/** A name of the HTML standard, matched as written: "Ø" is Ø and "ø" is ø. */
+const decodeName = (name, whole) => {
+    const value = Object.hasOwn(characterEntities, name) ? characterEntities[name] : undefined;
+    if (value === undefined)
+        return whole;
+    return isSpace(value) ? " " : value;
+};
+const fromCodePoint = (code, whole) => (Number.isInteger(code) && code > 0 && code <= 0x10ffff ? String.fromCodePoint(code) : whole);
+const decodeOne = (whole, name) => {
+    if (name.startsWith("#x") || name.startsWith("#X"))
+        return fromCodePoint(Number.parseInt(name.slice(2), 16), whole);
+    if (name.startsWith("#"))
+        return fromCodePoint(Number.parseInt(name.slice(1), 10), whole);
+    return decodeName(name, whole);
+};
+/** In one pass, so "&amp;lt;" is "&lt;" and not "<". An unknown name is left as written. */
+export const decodeEntities = (text) => text.replace(/&(#x[0-9a-f]+|#\d+|[a-z][a-z0-9]*);/giu, (whole, name) => decodeOne(whole, name));
+/** Trailing spaces removed, runs of blank lines made one, and a single newline at the end. */
+export const tidyLines = (lines) => {
+    const text = lines
+        .map((line) => line.trimEnd())
+        .join("\n")
+        .replace(/\n{3,}/gu, "\n\n")
+        .trim();
+    return text === "" ? "" : `${text}\n`;
+};
+//# sourceMappingURL=markup-text.js.map

package/dist/html/markup-text.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"markup-text.js","sourceRoot":"","sources":["../../src/html/markup-text.ts"],"names":[],"mappings":"AAAA,+GAA+G;AAC/G,yEAAyE;AACzE,OAAO,EAAE,iBAAiB,EAAE,MAAM,oBAAoB,CAAC;AAEvD,8GAA8G;AAC9G,MAAM,OAAO,GAAG,CAAC,KAAa,EAAW,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;AAErE,4FAA4F;AAC5F,MAAM,UAAU,GAAG,CAAC,IAAY,EAAE,KAAa,EAAU,EAAE;IACzD,MAAM,KAAK,GAAG,MAAM,CAAC,MAAM,CAAC,iBAAiB,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,iBAAiB,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;IAC3F,IAAI,KAAK,KAAK,SAAS;QAAE,OAAO,KAAK,CAAC;IACtC,OAAO,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,KAAK,CAAC;AACtC,CAAC,CAAC;AAEF,MAAM,aAAa,GAAG,CAAC,IAAY,EAAE,KAAa,EAAU,EAAE,CAAC,CAAC,MAAM,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,IAAI,GAAG,CAAC,IAAI,IAAI,IAAI,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,aAAa,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC;AAE7J,MAAM,SAAS,GAAG,CAAC,KAAa,EAAE,IAAY,EAAU,EAAE;IACxD,IAAI,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC,UAAU,CAAC,IAAI,CAAC;QAAE,OAAO,aAAa,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,KAAK,CAAC,CAAC;IACpH,IAAI,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC;QAAE,OAAO,aAAa,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,KAAK,CAAC,CAAC;IAC1F,OAAO,UAAU,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;AACjC,CAAC,CAAC;AAEF,4FAA4F;AAC5F,MAAM,CAAC,MAAM,cAAc,GAAG,CAAC,IAAY,EAAU,EAAE,CACrD,IAAI,CAAC,OAAO,CAAC,wCAAwC,EAAE,CAAC,KAAa,EAAE,IAAY,EAAE,EAAE,CAAC,SAAS,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC,CAAC;AAElH,8FAA8F;AAC9F,MAAM,CAAC,MAAM,SAAS,GAAG,CAAC,KAAwB,EAAU,EAAE;IAC5D,MAAM,IAAI,GAAG,KAAK;SACf,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,OAAO,EAAE,CAAC;SAC7B,IAAI,CAAC,IAAI,CAAC;SACV,OAAO,CAAC,UAAU,EAAE,MAAM,CAAC;SAC3B,IAAI,EAAE,CAAC;IACV,OAAO,IAAI,KAAK,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,IAAI,IAAI,CAAC;AACxC,CAAC,CAAC"}

package/dist/html/page-document.d.ts

@@ -0,0 +1,9 @@
+/** The document a page reads as: the name that says how to read it (.md or .txt), and its text. */
+export type PageDocument = {
+    readonly path: string;
+    readonly text: string;
+} | {
+    readonly unsupported: string;
+};
+export declare const pageDocument: (contentType: string | null, body: string) => PageDocument;
+//# sourceMappingURL=page-document.d.ts.map

package/dist/html/page-document.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"page-document.d.ts","sourceRoot":"","sources":["../../src/html/page-document.ts"],"names":[],"mappings":"AAIA,mGAAmG;AACnG,MAAM,MAAM,YAAY,GAAG;IAAE,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAA;CAAE,GAAG;IAAE,QAAQ,CAAC,WAAW,EAAE,MAAM,CAAA;CAAE,CAAC;AAU/G,eAAO,MAAM,YAAY,GAAI,aAAa,MAAM,GAAG,IAAI,EAAE,MAAM,MAAM,KAAG,YAMvE,CAAC"}

package/dist/html/page-document.js

@@ -0,0 +1,19 @@
+// A fetched page as a document chaff reads: HTML turned into Markdown (headings become addresses), Markdown and plain
+// text as they are. Anything else (a PDF, JSON, an image) is not text a quotation can be found in. Pure.
+import { htmlToMarkdown } from "./html-markdown.js";
+const mediaTypeOf = (contentType) => (contentType ?? "").split(";")[0]?.trim().toLowerCase() ?? "";
+/** A body sent with no Content-Type is HTML when it opens like markup. */
+const LOOKS_LIKE_HTML = /^\s*(?:<!doctype html|<html[\s>]|<head[\s>]|<body[\s>])/iu;
+const HTML_TYPES = new Set(["text/html", "application/xhtml+xml"]);
+const MARKDOWN_TYPES = new Set(["text/markdown", "text/x-markdown"]);
+export const pageDocument = (contentType, body) => {
+    const type = mediaTypeOf(contentType);
+    if (HTML_TYPES.has(type) || (type === "" && LOOKS_LIKE_HTML.test(body)))
+        return { path: "page.md", text: htmlToMarkdown(body) };
+    if (MARKDOWN_TYPES.has(type))
+        return { path: "page.md", text: body };
+    if (type === "" || type.startsWith("text/"))
+        return { path: "page.txt", text: body };
+    return { unsupported: type };
+};
+//# sourceMappingURL=page-document.js.map

package/dist/html/page-document.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"page-document.js","sourceRoot":"","sources":["../../src/html/page-document.ts"],"names":[],"mappings":"AAAA,sHAAsH;AACtH,yGAAyG;AACzG,OAAO,EAAE,cAAc,EAAE,MAAM,oBAAoB,CAAC;AAKpD,MAAM,WAAW,GAAG,CAAC,WAA0B,EAAU,EAAE,CAAC,CAAC,WAAW,IAAI,EAAE,CAAC,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,CAAC,WAAW,EAAE,IAAI,EAAE,CAAC;AAE1H,0EAA0E;AAC1E,MAAM,eAAe,GAAG,2DAA2D,CAAC;AAEpF,MAAM,UAAU,GAAwB,IAAI,GAAG,CAAC,CAAC,WAAW,EAAE,uBAAuB,CAAC,CAAC,CAAC;AACxF,MAAM,cAAc,GAAwB,IAAI,GAAG,CAAC,CAAC,eAAe,EAAE,iBAAiB,CAAC,CAAC,CAAC;AAE1F,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,WAA0B,EAAE,IAAY,EAAgB,EAAE;IACrF,MAAM,IAAI,GAAG,WAAW,CAAC,WAAW,CAAC,CAAC;IACtC,IAAI,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,KAAK,EAAE,IAAI,eAAe,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAAE,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,IAAI,EAAE,cAAc,CAAC,IAAI,CAAC,EAAE,CAAC;IAChI,IAAI,cAAc,CAAC,GAAG,CAAC,IAAI,CAAC;QAAE,OAAO,EAAE,IAAI,EAAE,SAAS,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;IACrE,IAAI,IAAI,KAAK,EAAE,IAAI,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC;QAAE,OAAO,EAAE,IAAI,EAAE,UAAU,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC;IACrF,OAAO,EAAE,WAAW,EAAE,IAAI,EAAE,CAAC;AAC/B,CAAC,CAAC"}

package/dist/inline-count.d.ts

@@ -0,0 +1,19 @@
+import { type CountMismatch, type CountWords } from "./announced-count.ts";
+/**
+ * 一行に並べた名前の予告（出席者（6名）：田中、鈴木、佐藤 / Attendees (6): Tanaka, Suzuki）。ラベルの後ろの括弧の数と、
+ * コロンの後ろに区切って並べた名前の数を比べる。区切りが決まらない並び（名前をつなぐ語がある、括弧が閉じていない、文が続く）と、
+ * 閉じていない並び（ほか、etc.）は数えない。言語の知識（区切り、つなぐ語、閉じていないことを言う語）は語彙表から受け取る。
+ */
+export type MemberWords = {
+    /** 名前の区切り。、 , */
+    readonly separators: readonly string[];
+    /** 名前をつなぐ語。と、and。 */
+    readonly joiners: readonly string[];
+    /** 並びが閉じていないことを言う語。ほか、etc。 */
+    readonly open: readonly string[];
+};
+/** 区切って並べた名前の数。並びとして読めないときは undefined。 */
+export declare const memberCount: (members: string, words: MemberWords) => number | undefined;
+/** 本文（コードや強調の印を空白で覆ったもの）の、一行に並べた名前の予告で、数が合わないもの。 */
+export declare const inlineCountMismatches: (text: string, words: CountWords, members: MemberWords) => CountMismatch[];
+//# sourceMappingURL=inline-count.d.ts.map

package/dist/inline-count.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"inline-count.d.ts","sourceRoot":"","sources":["../src/inline-count.ts"],"names":[],"mappings":"AACA,OAAO,EAAwB,KAAK,aAAa,EAAE,KAAK,UAAU,EAAE,MAAM,sBAAsB,CAAC;AAEjG;;;;GAIG;AACH,MAAM,MAAM,WAAW,GAAG;IACxB,iBAAiB;IACjB,QAAQ,CAAC,UAAU,EAAE,SAAS,MAAM,EAAE,CAAC;IACvC,qBAAqB;IACrB,QAAQ,CAAC,OAAO,EAAE,SAAS,MAAM,EAAE,CAAC;IACpC,8BAA8B;IAC9B,QAAQ,CAAC,IAAI,EAAE,SAAS,MAAM,EAAE,CAAC;CAClC,CAAC;AA8DF,0CAA0C;AAC1C,eAAO,MAAM,WAAW,GAAI,SAAS,MAAM,EAAE,OAAO,WAAW,KAAG,MAAM,GAAG,SAQ1E,CAAC;AAgDF,oDAAoD;AACpD,eAAO,MAAM,qBAAqB,GAAI,MAAM,MAAM,EAAE,OAAO,UAAU,EAAE,SAAS,WAAW,KAAG,aAAa,EAU1G,CAAC"}

package/dist/inline-count.js

@@ -0,0 +1,109 @@
+import { escapeRegExp } from "./orthography.js";
+import { alternation, valueOf } from "./announced-count.js";
+const MAX_LABEL_CHARS = 30;
+const MAX_MEMBER_CHARS = 40;
+const MIN_MEMBERS = 2;
+/** 一行に並べる名前の数として読む上限。括弧の年（Holland (1993): ...）を数に読まない。 */
+const MAX_ANNOUNCED = 99;
+/**
+ * 行頭（箇条書きの印の後ろでもよい）のラベル、括弧の中の数（数える語は付いても付かなくてもよい）、コロン。
+ * 括弧の中に数以外の語があるもの（約6名、6名以上、5名ほか）は形が合わず、ここで外れる。
+ */
+const headPattern = (words) => {
+    const numbers = words.numbers.length === 0 ? "\\p{Nd}+" : `\\p{Nd}+|${alternation(words.numbers)}`;
+    const counter = words.counters.length === 0 ? "" : `(?:[ \\t]?(?<counter>${alternation(words.counters)})(?![A-Za-z]))?`;
+    const label = `(?<label>[^\\s\\p{Nd}(（:：|>#*+-][^(（:：\\n]{0,${MAX_LABEL_CHARS}})`;
+    const listMark = String.raw `(?:(?:[-*+]|\d{1,9}[.)])[ \t]+)?`;
+    return new RegExp(`^[ \\t]*${listMark}${label}[(（][ \\t]*(?<phrase>(?<number>${numbers})${counter})[ \\t]*[)）][ \\t]*[:：][ \\t]*`, "diu");
+};
+const LATIN_LETTER = /[A-Za-z]/u;
+/** 英語の語は語の切れ目で照らす（and が Anderson に当たらない）。日本語の語は字面どおり。atEnd なら末尾の語だけ。 */
+const wordPattern = (word, atEnd = false) => {
+    const before = LATIN_LETTER.test(word[0] ?? "") ? "(?<![A-Za-z])" : "";
+    const after = LATIN_LETTER.test(word.at(-1) ?? "") ? "(?![A-Za-z])" : "";
+    return new RegExp(`${before}${escapeRegExp(word)}${after}${atEnd ? "\\s*$" : ""}`, "iu");
+};
+const BRACKETED = /\([^()（）]*\)|（[^()（）]*）|「[^「」]*」|\[[^[\]]*\]/gu;
+const BRACKET = /[()（）「」[\]]/u;
+/** 並びの途中で文が切れる、別のラベルが始まる。名前の並びでなく文なので数えない。 */
+const NOT_A_NAME_LIST = /[。．.:：;；]/u;
+const TRAILING_END = /[。．.]\s*$/u;
+/** 括弧の中の注記（議長）を除いた並び。括弧が閉じていなければ undefined。 */
+const withoutNotes = (members) => {
+    const stripped = members.replace(BRACKETED, "");
+    return BRACKET.test(stripped) ? undefined : stripped;
+};
+const hasAny = (text, words) => words.some((word) => wordPattern(word).test(text));
+/** 閉じていないことを言う語は名前の後ろに付く（田中ほか、etc）。名前の中の字（等々力）は数える。 */
+const endsOpen = (member, open) => open.some((word) => wordPattern(word, true).test(member));
+/** 最後の名前の前のつなぐ語（, and Ito）は並びの終わりの印で、名前の一部ではない。 */
+const withoutLeadingJoiner = (member, joiners) => {
+    const leading = joiners.find((joiner) => new RegExp(`^${escapeRegExp(joiner)}\\s+`, "iu").test(member));
+    return leading === undefined ? member : member.slice(leading.length).trimStart();
+};
+const splitMembers = (text, separators) => text.split(new RegExp(separators.map(escapeRegExp).join("|"), "u")).map((part) => part.trim());
+const withLastJoinerOff = (parts, joiners) => {
+    const last = parts.at(-1);
+    return last === undefined ? [...parts] : [...parts.slice(0, -1), withoutLeadingJoiner(last, joiners)];
+};
+const isNameLike = (member, words) => member !== "" && member.length <= MAX_MEMBER_CHARS && !hasAny(member, words.joiners);
+/** 区切って並べた名前の数。並びとして読めないときは undefined。 */
+export const memberCount = (members, words) => {
+    if (words.separators.length === 0)
+        return undefined;
+    const notes = withoutNotes(members.trim().replace(TRAILING_END, ""));
+    if (notes === undefined || NOT_A_NAME_LIST.test(notes))
+        return undefined;
+    const parts = splitMembers(notes, words.separators);
+    if (parts.some((part) => endsOpen(part, words.open)))
+        return undefined;
+    const names = withLastJoinerOff(parts, words.joiners);
+    return names.length >= MIN_MEMBERS && names.every((name) => isNameLike(name, words)) ? names.length : undefined;
+};
+/** 次の行に並びが続いてよい行。空行、箇条書き、見出し、引用、表の行は別のもの。 */
+const CONTINUATION = /^[ \t]*[^\s\-*+#>|]/u;
+const endsWithSeparator = (text, words) => words.separators.some((separator) => text.trimEnd().endsWith(separator));
+/** 区切りで終わる行は、次の行に並びが続く。 */
+const membersFrom = (lines, next, soFar, words) => {
+    const line = lines[next] ?? "";
+    return endsWithSeparator(soFar, words) && CONTINUATION.test(line) ? membersFrom(lines, next + 1, soFar + line, words) : soFar;
+};
+/**
+ * ラベルは短い名前（出席者、Attendees）。区切りを含むもの（Dunn, G.E. (1960):）は文献の著者で、番号で終わるもの（Step 2 (3):）は
+ * 番号の付いた項目。どちらも括弧の数は並べた名前の数ではない。
+ */
+const isLabel = (label, members) => !/\p{Nd}\s*$/u.test(label) && !members.separators.some((separator) => label.includes(separator));
+/** 数える語の無い数（Attendees (6)）は、ラベル自身が数えるもの（attendees、items）を言うときだけ数。Appendix (3) は番号。 */
+const labelCounts = (label, counters) => counters.some((counter) => wordPattern(counter, true).test(label));
+const linesOf = (text) => {
+    const starts = [0, ...[...text.matchAll(/\n/gu)].map((match) => match.index + 1)];
+    return text.split("\n").map((line, index) => ({ text: line, start: starts[index] ?? 0 }));
+};
+/** 数える語が数に付いているか、ラベルが数えるものを言っている（Attendees (6)）ときだけ、括弧の数を名前の数に読む。 */
+const isCount = (announced, counter, label, words) => (counter !== undefined || labelCounts(label, words.counters)) && announced >= 1 && announced <= MAX_ANNOUNCED;
+const headOf = (text, head, words, members) => {
+    const match = head.exec(text);
+    const groups = match?.groups;
+    const phraseAt = match?.indices?.groups?.["phrase"]?.[0];
+    const label = groups?.["label"] ?? "";
+    if (match === null || groups === undefined || phraseAt === undefined || !isLabel(label, members))
+        return undefined;
+    const announced = valueOf(groups["number"] ?? "", words.numbers);
+    if (!isCount(announced, groups["counter"], label, words))
+        return undefined;
+    return { phrase: (groups["phrase"] ?? "").trim(), phraseAt, announced, rest: text.slice(match[0].length) };
+};
+/** 本文（コードや強調の印を空白で覆ったもの）の、一行に並べた名前の予告で、数が合わないもの。 */
+export const inlineCountMismatches = (text, words, members) => {
+    const head = headPattern(words);
+    const lines = linesOf(text);
+    const texts = lines.map((line) => line.text);
+    return lines.flatMap((line, index) => {
+        const found = headOf(line.text, head, words, members);
+        const listed = found === undefined ? undefined : memberCount(membersFrom(texts, index + 1, found.rest, members), members);
+        if (found === undefined || listed === undefined || listed === found.announced)
+            return [];
+        return [{ offset: line.start + found.phraseAt, phrase: found.phrase, announced: found.announced, listed }];
+    });
+};
+//# sourceMappingURL=inline-count.js.map

package/dist/inline-count.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"inline-count.js","sourceRoot":"","sources":["../src/inline-count.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,YAAY,EAAE,MAAM,kBAAkB,CAAC;AAChD,OAAO,EAAE,WAAW,EAAE,OAAO,EAAuC,MAAM,sBAAsB,CAAC;AAgBjG,MAAM,eAAe,GAAG,EAAE,CAAC;AAC3B,MAAM,gBAAgB,GAAG,EAAE,CAAC;AAC5B,MAAM,WAAW,GAAG,CAAC,CAAC;AACtB,0DAA0D;AAC1D,MAAM,aAAa,GAAG,EAAE,CAAC;AAEzB;;;GAGG;AACH,MAAM,WAAW,GAAG,CAAC,KAAiB,EAAU,EAAE;IAChD,MAAM,OAAO,GAAG,KAAK,CAAC,OAAO,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,YAAY,WAAW,CAAC,KAAK,CAAC,OAAO,CAAC,EAAE,CAAC;IACnG,MAAM,OAAO,GAAG,KAAK,CAAC,QAAQ,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,wBAAwB,WAAW,CAAC,KAAK,CAAC,QAAQ,CAAC,iBAAiB,CAAC;IACxH,MAAM,KAAK,GAAG,gDAAgD,eAAe,IAAI,CAAC;IAClF,MAAM,QAAQ,GAAG,MAAM,CAAC,GAAG,CAAA,kCAAkC,CAAC;IAC9D,OAAO,IAAI,MAAM,CAAC,WAAW,QAAQ,GAAG,KAAK,kCAAkC,OAAO,IAAI,OAAO,gCAAgC,EAAE,KAAK,CAAC,CAAC;AAC5I,CAAC,CAAC;AAEF,MAAM,YAAY,GAAG,WAAW,CAAC;AAEjC,wEAAwE;AACxE,MAAM,WAAW,GAAG,CAAC,IAAY,EAAE,KAAK,GAAG,KAAK,EAAU,EAAE;IAC1D,MAAM,MAAM,GAAG,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,eAAe,CAAC,CAAC,CAAC,EAAE,CAAC;IACvE,MAAM,KAAK,GAAG,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,cAAc,CAAC,CAAC,CAAC,EAAE,CAAC;IACzE,OAAO,IAAI,MAAM,CAAC,GAAG,MAAM,GAAG,YAAY,CAAC,IAAI,CAAC,GAAG,KAAK,GAAG,KAAK,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,EAAE,IAAI,CAAC,CAAC;AAC3F,CAAC,CAAC;AAEF,MAAM,SAAS,GAAG,gDAAgD,CAAC;AACnE,MAAM,OAAO,GAAG,cAAc,CAAC;AAC/B,8CAA8C;AAC9C,MAAM,eAAe,GAAG,YAAY,CAAC;AACrC,MAAM,YAAY,GAAG,YAAY,CAAC;AAElC,+CAA+C;AAC/C,MAAM,YAAY,GAAG,CAAC,OAAe,EAAsB,EAAE;IAC3D,MAAM,QAAQ,GAAG,OAAO,CAAC,OAAO,CAAC,SAAS,EAAE,EAAE,CAAC,CAAC;IAChD,OAAO,OAAO,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,QAAQ,CAAC;AACvD,CAAC,CAAC;AAEF,MAAM,MAAM,GAAG,CAAC,IAAY,EAAE,KAAwB,EAAW,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC;AAEvH,uDAAuD;AACvD,MAAM,QAAQ,GAAG,CAAC,MAAc,EAAE,IAAuB,EAAW,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,WAAW,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC;AAEjI,mDAAmD;AACnD,MAAM,oBAAoB,GAAG,CAAC,MAAc,EAAE,OAA0B,EAAU,EAAE;IAClF,MAAM,OAAO,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,IAAI,MAAM,CAAC,IAAI,YAAY,CAAC,MAAM,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC;IACxG,OAAO,OAAO,KAAK,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,CAAC,SAAS,EAAE,CAAC;AACnF,CAAC,CAAC;AAEF,MAAM,YAAY,GAAG,CAAC,IAAY,EAAE,UAA6B,EAAY,EAAE,CAC7E,IAAI,CAAC,KAAK,CAAC,IAAI,MAAM,CAAC,UAAU,CAAC,GAAG,CAAC,YAAY,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC,CAAC;AAEjG,MAAM,iBAAiB,GAAG,CAAC,KAAwB,EAAE,OAA0B,EAAY,EAAE;IAC3F,MAAM,IAAI,GAAG,KAAK,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC;IAC1B,OAAO,IAAI,KAAK,SAAS,CAAC,CAAC,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,GAAG,KAAK,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,EAAE,oBAAoB,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC,CAAC;AACxG,CAAC,CAAC;AAEF,MAAM,UAAU,GAAG,CAAC,MAAc,EAAE,KAAkB,EAAW,EAAE,CAAC,MAAM,KAAK,EAAE,IAAI,MAAM,CAAC,MAAM,IAAI,gBAAgB,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,KAAK,CAAC,OAAO,CAAC,CAAC;AAEzJ,0CAA0C;AAC1C,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,OAAe,EAAE,KAAkB,EAAsB,EAAE;IACrF,IAAI,KAAK,CAAC,UAAU,CAAC,MAAM,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC;IACpD,MAAM,KAAK,GAAG,YAAY,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC,OAAO,CAAC,YAAY,EAAE,EAAE,CAAC,CAAC,CAAC;IACrE,IAAI,KAAK,KAAK,SAAS,IAAI,eAAe,CAAC,IAAI,CAAC,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC;IACzE,MAAM,KAAK,GAAG,YAAY,CAAC,KAAK,EAAE,KAAK,CAAC,UAAU,CAAC,CAAC;IACpD,IAAI,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,QAAQ,CAAC,IAAI,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC;QAAE,OAAO,SAAS,CAAC;IACvE,MAAM,KAAK,GAAG,iBAAiB,CAAC,KAAK,EAAE,KAAK,CAAC,OAAO,CAAC,CAAC;IACtD,OAAO,KAAK,CAAC,MAAM,IAAI,WAAW,IAAI,KAAK,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,UAAU,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,SAAS,CAAC;AAClH,CAAC,CAAC;AAEF,6CAA6C;AAC7C,MAAM,YAAY,GAAG,sBAAsB,CAAC;AAE5C,MAAM,iBAAiB,GAAG,CAAC,IAAY,EAAE,KAAkB,EAAW,EAAE,CAAC,KAAK,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,IAAI,CAAC,OAAO,EAAE,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,CAAC;AAElJ,2BAA2B;AAC3B,MAAM,WAAW,GAAG,CAAC,KAAwB,EAAE,IAAY,EAAE,KAAa,EAAE,KAAkB,EAAU,EAAE;IACxG,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;IAC/B,OAAO,iBAAiB,CAAC,KAAK,EAAE,KAAK,CAAC,IAAI,YAAY,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,WAAW,CAAC,KAAK,EAAE,IAAI,GAAG,CAAC,EAAE,KAAK,GAAG,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC;AAChI,CAAC,CAAC;AAEF;;;GAGG;AACH,MAAM,OAAO,GAAG,CAAC,KAAa,EAAE,OAAoB,EAAW,EAAE,CAC/D,CAAC,aAAa,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,KAAK,CAAC,QAAQ,CAAC,SAAS,CAAC,CAAC,CAAC;AAEnG,sFAAsF;AACtF,MAAM,WAAW,GAAG,CAAC,KAAa,EAAE,QAA2B,EAAW,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,WAAW,CAAC,OAAO,EAAE,IAAI,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC;AAIhJ,MAAM,OAAO,GAAG,CAAC,IAAY,EAAU,EAAE;IACvC,MAAM,MAAM,GAAG,CAAC,CAAC,EAAE,GAAG,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC,CAAC;IAClF,OAAO,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC;AAC5F,CAAC,CAAC;AAKF,qEAAqE;AACrE,MAAM,OAAO,GAAG,CAAC,SAAiB,EAAE,OAA2B,EAAE,KAAa,EAAE,KAAiB,EAAW,EAAE,CAC5G,CAAC,OAAO,KAAK,SAAS,IAAI,WAAW,CAAC,KAAK,EAAE,KAAK,CAAC,QAAQ,CAAC,CAAC,IAAI,SAAS,IAAI,CAAC,IAAI,SAAS,IAAI,aAAa,CAAC;AAEhH,MAAM,MAAM,GAAG,CAAC,IAAY,EAAE,IAAY,EAAE,KAAiB,EAAE,OAAoB,EAAoB,EAAE;IACvG,MAAM,KAAK,GAAG,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC9B,MAAM,MAAM,GAAG,KAAK,EAAE,MAAM,CAAC;IAC7B,MAAM,QAAQ,GAAG,KAAK,EAAE,OAAO,EAAE,MAAM,EAAE,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC;IACzD,MAAM,KAAK,GAAG,MAAM,EAAE,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC;IACtC,IAAI,KAAK,KAAK,IAAI,IAAI,MAAM,KAAK,SAAS,IAAI,QAAQ,KAAK,SAAS,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,OAAO,CAAC;QAAE,OAAO,SAAS,CAAC;IACnH,MAAM,SAAS,GAAG,OAAO,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,EAAE,EAAE,KAAK,CAAC,OAAO,CAAC,CAAC;IACjE,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE,MAAM,CAAC,SAAS,CAAC,EAAE,KAAK,EAAE,KAAK,CAAC;QAAE,OAAO,SAAS,CAAC;IAC3E,OAAO,EAAE,MAAM,EAAE,CAAC,MAAM,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,EAAE,QAAQ,EAAE,SAAS,EAAE,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,EAAE,CAAC;AAC7G,CAAC,CAAC;AAEF,oDAAoD;AACpD,MAAM,CAAC,MAAM,qBAAqB,GAAG,CAAC,IAAY,EAAE,KAAiB,EAAE,OAAoB,EAAmB,EAAE;IAC9G,MAAM,IAAI,GAAG,WAAW,CAAC,KAAK,CAAC,CAAC;IAChC,MAAM,KAAK,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;IAC5B,MAAM,KAAK,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;IAC7C,OAAO,KAAK,CAAC,OAAO,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE;QACnC,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,IAAI,EAAE,IAAI,EAAE,KAAK,EAAE,OAAO,CAAC,CAAC;QACtD,MAAM,MAAM,GAAG,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,WAAW,CAAC,WAAW,CAAC,KAAK,EAAE,KAAK,GAAG,CAAC,EAAE,KAAK,CAAC,IAAI,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,CAAC;QAC1H,IAAI,KAAK,KAAK,SAAS,IAAI,MAAM,KAAK,SAAS,IAAI,MAAM,KAAK,KAAK,CAAC,SAAS;YAAE,OAAO,EAAE,CAAC;QACzF,OAAO,CAAC,EAAE,MAAM,EAAE,IAAI,CAAC,KAAK,GAAG,KAAK,CAAC,QAAQ,EAAE,MAAM,EAAE,KAAK,CAAC,MAAM,EAAE,SAAS,EAAE,KAAK,CAAC,SAAS,EAAE,MAAM,EAAE,CAAC,CAAC;IAC7G,CAAC,CAAC,CAAC;AACL,CAAC,CAAC"}

package/dist/kana-spelling.d.ts

@@ -3,6 +3,11 @@ export declare const toKatakana: (text: string) => string;
 export declare const isKanaOnly: (text: string) => boolean;
 /** The kanji of a spelling, in order: 引越 for both 引っ越し and 引越し, nothing for できる. */
 export declare const kanjiSkeleton: (text: string) => string;
+/**
+ * Whether a spelling is the kanji alone of a word the others spell with okurigana: 取扱 beside 取扱い, 締切 beside 締切り. Official
+ * Japanese drops okurigana at the head of a compound (取扱事業者) and keeps it on the word alone, so the two are one convention.
+ */
+export declare const dropsOkurigana: (spelling: string, spellings: readonly string[]) => boolean;
 /**
  * The reading of the dictionary form, from the reading of the written form: 下さい (クダサイ) with lemma 下さる reads クダサル.
  * The written form and the lemma differ only in their kana ending, so that ending is swapped in the reading. When the ending is

package/dist/kana-spelling.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"kana-spelling.d.ts","sourceRoot":"","sources":["../src/kana-spelling.ts"],"names":[],"mappings":"AASA,6DAA6D;AAC7D,eAAO,MAAM,UAAU,GAAI,MAAM,MAAM,KAAG,MAAmH,CAAC;AAE9J,eAAO,MAAM,UAAU,GAAI,MAAM,MAAM,KAAG,OAA+B,CAAC;AAE1E,oFAAoF;AACpF,eAAO,MAAM,aAAa,GAAI,MAAM,MAAM,KAAG,MAAqE,CAAC;AAUnH;;;;GAIG;AACH,eAAO,MAAM,YAAY,GAAI,SAAS,MAAM,EAAE,OAAO,MAAM,EAAE,SAAS,MAAM,KAAG,MAAM,GAAG,SAOvF,CAAC;AAKF,oHAAoH;AACpH,eAAO,MAAM,WAAW,GAAI,SAAS,MAAM,KAAG,MAKrB,CAAC"}
+{"version":3,"file":"kana-spelling.d.ts","sourceRoot":"","sources":["../src/kana-spelling.ts"],"names":[],"mappings":"AASA,6DAA6D;AAC7D,eAAO,MAAM,UAAU,GAAI,MAAM,MAAM,KAAG,MAAmH,CAAC;AAE9J,eAAO,MAAM,UAAU,GAAI,MAAM,MAAM,KAAG,OAA+B,CAAC;AAE1E,oFAAoF;AACpF,eAAO,MAAM,aAAa,GAAI,MAAM,MAAM,KAAG,MAAqE,CAAC;AAEnH;;;GAGG;AACH,eAAO,MAAM,cAAc,GAAI,UAAU,MAAM,EAAE,WAAW,SAAS,MAAM,EAAE,KAAG,OAC+D,CAAC;AAUhJ;;;;GAIG;AACH,eAAO,MAAM,YAAY,GAAI,SAAS,MAAM,EAAE,OAAO,MAAM,EAAE,SAAS,MAAM,KAAG,MAAM,GAAG,SAOvF,CAAC;AAKF,oHAAoH;AACpH,eAAO,MAAM,WAAW,GAAI,SAAS,MAAM,KAAG,MAKrB,CAAC"}

package/dist/kana-spelling.js

@@ -10,6 +10,11 @@ export const toKatakana = (text) => text.replace(HIRAGANA, (char) => String.from
 export const isKanaOnly = (text) => KANA_ONLY.test(text);
 /** The kanji of a spelling, in order: 引越 for both 引っ越し and 引越し, nothing for できる. */
 export const kanjiSkeleton = (text) => [...text.matchAll(KANJI)].map((match) => match[0]).join("");
+/**
+ * Whether a spelling is the kanji alone of a word the others spell with okurigana: 取扱 beside 取扱い, 締切 beside 締切り. Official
+ * Japanese drops okurigana at the head of a compound (取扱事業者) and keeps it on the word alone, so the two are one convention.
+ */
+export const dropsOkurigana = (spelling, spellings) => spelling !== "" && kanjiSkeleton(spelling) === spelling && spellings.some((other) => other !== spelling && kanjiSkeleton(other) === spelling);
 /** The length, in UTF-16 units, of the start both strings share, compared character by character (𠮟 is one character, two units). */
 const sharedPrefixLength = (left, right) => {
     const [leftChars, rightChars] = [[...left], [...right]];

package/dist/kana-spelling.js.map

@@ -1 +1 @@
-{"version":3,"file":"kana-spelling.js","sourceRoot":"","sources":["../src/kana-spelling.ts"],"names":[],"mappings":"AAAA,yHAAyH;AACzH,mHAAmH;AACnH,yEAAyE;AAEzE,MAAM,oBAAoB,GAAG,IAAI,CAAC;AAClC,MAAM,QAAQ,GAAG,SAAS,CAAC;AAC3B,MAAM,SAAS,GAAG,eAAe,CAAC;AAClC,MAAM,KAAK,GAAG,qBAAqB,CAAC;AAEpC,6DAA6D;AAC7D,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,IAAY,EAAU,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,CAAC,aAAa,CAAC,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,oBAAoB,CAAC,CAAC,CAAC;AAE9J,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,IAAY,EAAW,EAAE,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAE1E,oFAAoF;AACpF,MAAM,CAAC,MAAM,aAAa,GAAG,CAAC,IAAY,EAAU,EAAE,CAAC,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AAEnH,sIAAsI;AACtI,MAAM,kBAAkB,GAAG,CAAC,IAAY,EAAE,KAAa,EAAU,EAAE;IACjE,MAAM,CAAC,SAAS,EAAE,UAAU,CAAC,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC;IACxD,MAAM,OAAO,GAAG,SAAS,CAAC,SAAS,CAAC,CAAC,IAAI,EAAE,EAAE,EAAE,EAAE,CAAC,UAAU,CAAC,EAAE,CAAC,KAAK,IAAI,CAAC,CAAC;IAC3E,MAAM,MAAM,GAAG,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,UAAU,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC;IACpG,OAAO,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,MAAM,CAAC;AAChC,CAAC,CAAC;AAEF;;;;GAIG;AACH,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,OAAe,EAAE,KAAa,EAAE,OAAe,EAAsB,EAAE;IAClG,MAAM,MAAM,GAAG,kBAAkB,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;IAClD,MAAM,CAAC,UAAU,EAAE,QAAQ,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,KAAK,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC;IAC5E,IAAI,CAAC,UAAU,KAAK,EAAE,IAAI,CAAC,UAAU,CAAC,UAAU,CAAC,CAAC,IAAI,CAAC,QAAQ,KAAK,EAAE,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC,CAAC;QAAE,OAAO,SAAS,CAAC;IACnH,MAAM,MAAM,GAAG,UAAU,CAAC,UAAU,CAAC,CAAC;IACtC,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAC;QAAE,OAAO,SAAS,CAAC;IAChD,OAAO,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC,GAAG,UAAU,CAAC,QAAQ,CAAC,CAAC;AACjF,CAAC,CAAC;AAEF,MAAM,cAAc,GAAqC,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC;AAC5H,MAAM,EAAE,GAAqC,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC;AAE5F,oHAAoH;AACpH,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,OAAe,EAAU,EAAE,CACrD,OAAO;KACJ,OAAO,CAAC,YAAY,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,EAAE,CAAC,KAAK,CAAC,IAAI,KAAK,CAAC;KACpD,OAAO,CAAC,cAAc,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,cAAc,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC;KAC/D,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC;KAClB,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC"}
+{"version":3,"file":"kana-spelling.js","sourceRoot":"","sources":["../src/kana-spelling.ts"],"names":[],"mappings":"AAAA,yHAAyH;AACzH,mHAAmH;AACnH,yEAAyE;AAEzE,MAAM,oBAAoB,GAAG,IAAI,CAAC;AAClC,MAAM,QAAQ,GAAG,SAAS,CAAC;AAC3B,MAAM,SAAS,GAAG,eAAe,CAAC;AAClC,MAAM,KAAK,GAAG,qBAAqB,CAAC;AAEpC,6DAA6D;AAC7D,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,IAAY,EAAU,EAAE,CAAC,IAAI,CAAC,OAAO,CAAC,QAAQ,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,MAAM,CAAC,aAAa,CAAC,CAAC,IAAI,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,oBAAoB,CAAC,CAAC,CAAC;AAE9J,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,IAAY,EAAW,EAAE,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;AAE1E,oFAAoF;AACpF,MAAM,CAAC,MAAM,aAAa,GAAG,CAAC,IAAY,EAAU,EAAE,CAAC,CAAC,GAAG,IAAI,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;AAEnH;;;GAGG;AACH,MAAM,CAAC,MAAM,cAAc,GAAG,CAAC,QAAgB,EAAE,SAA4B,EAAW,EAAE,CACxF,QAAQ,KAAK,EAAE,IAAI,aAAa,CAAC,QAAQ,CAAC,KAAK,QAAQ,IAAI,SAAS,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,KAAK,QAAQ,IAAI,aAAa,CAAC,KAAK,CAAC,KAAK,QAAQ,CAAC,CAAC;AAEhJ,sIAAsI;AACtI,MAAM,kBAAkB,GAAG,CAAC,IAAY,EAAE,KAAa,EAAU,EAAE;IACjE,MAAM,CAAC,SAAS,EAAE,UAAU,CAAC,GAAG,CAAC,CAAC,GAAG,IAAI,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,CAAC,CAAC;IACxD,MAAM,OAAO,GAAG,SAAS,CAAC,SAAS,CAAC,CAAC,IAAI,EAAE,EAAE,EAAE,EAAE,CAAC,UAAU,CAAC,EAAE,CAAC,KAAK,IAAI,CAAC,CAAC;IAC3E,MAAM,MAAM,GAAG,OAAO,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,UAAU,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC;IACpG,OAAO,MAAM,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,MAAM,CAAC;AAChC,CAAC,CAAC;AAEF;;;;GAIG;AACH,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,OAAe,EAAE,KAAa,EAAE,OAAe,EAAsB,EAAE;IAClG,MAAM,MAAM,GAAG,kBAAkB,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC;IAClD,MAAM,CAAC,UAAU,EAAE,QAAQ,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,KAAK,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC;IAC5E,IAAI,CAAC,UAAU,KAAK,EAAE,IAAI,CAAC,UAAU,CAAC,UAAU,CAAC,CAAC,IAAI,CAAC,QAAQ,KAAK,EAAE,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC,CAAC;QAAE,OAAO,SAAS,CAAC;IACnH,MAAM,MAAM,GAAG,UAAU,CAAC,UAAU,CAAC,CAAC;IACtC,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAC;QAAE,OAAO,SAAS,CAAC;IAChD,OAAO,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,OAAO,CAAC,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC,GAAG,UAAU,CAAC,QAAQ,CAAC,CAAC;AACjF,CAAC,CAAC;AAEF,MAAM,cAAc,GAAqC,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC;AAC5H,MAAM,EAAE,GAAqC,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,EAAE,EAAE,GAAG,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC;AAE5F,oHAAoH;AACpH,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,OAAe,EAAU,EAAE,CACrD,OAAO;KACJ,OAAO,CAAC,YAAY,EAAE,CAAC,KAAK,EAAE,EAAE,CAAC,EAAE,CAAC,KAAK,CAAC,IAAI,KAAK,CAAC;KACpD,OAAO,CAAC,cAAc,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,cAAc,CAAC,IAAI,CAAC,IAAI,IAAI,CAAC;KAC/D,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC;KAClB,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC"}

package/dist/location-part.d.ts

@@ -0,0 +1,13 @@
+import type { Span, Token } from "./plugin.ts";
+/** units は階・部屋の番号の単位（語彙表 floor-unit）、levels は番号の前に書いて階の名前の一部になる語（語彙表 floor-level）。 */
+export type FloorWords = {
+    readonly units: ReadonlySet<string>;
+    readonly levels: ReadonlySet<string>;
+};
+/**
+ * 「本社 5階 第2会議室」「研究所 3階 セミナー室」のように、名前の語の後ろに空白で区切って書いた階・部屋の番号。
+ * その前の空白は所在の組を分ける区切りで、日本語と数字のあいだの空け方の好みではない。「地下 1階」「地上 5階」の空白は階の名前の中の空け方なので、区切りと読まない。
+ * 番号は数字だけで、単位の後ろで組が閉じる（空白・読点・文末・英字が続く）ものだけ。tokens は文書全体の座標で、base は text の先頭の位置。
+ */
+export declare const isSpacedLocationPart: (text: string, run: Span, tokens: readonly Token[] | undefined, base: number, words: FloorWords) => boolean;
+//# sourceMappingURL=location-part.d.ts.map

package/dist/location-part.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"location-part.d.ts","sourceRoot":"","sources":["../src/location-part.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,IAAI,EAAE,KAAK,EAAE,MAAM,aAAa,CAAC;AAU/C,sFAAsF;AACtF,MAAM,MAAM,UAAU,GAAG;IAAE,QAAQ,CAAC,KAAK,EAAE,WAAW,CAAC,MAAM,CAAC,CAAC;IAAC,QAAQ,CAAC,MAAM,EAAE,WAAW,CAAC,MAAM,CAAC,CAAA;CAAE,CAAC;AAMvG;;;;GAIG;AACH,eAAO,MAAM,oBAAoB,GAAI,MAAM,MAAM,EAAE,KAAK,IAAI,EAAE,QAAQ,SAAS,KAAK,EAAE,GAAG,SAAS,EAAE,MAAM,MAAM,EAAE,OAAO,UAAU,KAAG,OAOrI,CAAC"}

package/dist/location-part.js

@@ -0,0 +1,24 @@
+/** 空白の前の、建物や場所の名前になる語の品詞。助詞の後ろ（「会場は 5階」）は名前の区切りではない。 */
+const NAME_POS = new Set(["NOUN", "PROPN"]);
+/** 階・部屋の番号の後ろに続けば、番号が語の一部（5階建て、5階です）で、所在の組を閉じていない字。 */
+const WORD_CHAR = /[\p{Script=Han}\p{Script=Hiragana}\p{Script=Katakana}]/u;
+const SPACES = new Set([" ", "　"]);
+/** run のすぐ後ろに書いた、units のうち最も長い語。無ければ undefined。 */
+const unitAfter = (text, run, units) => [...units].filter((unit) => text.startsWith(unit, run.end)).toSorted((left, right) => right.length - left.length)[0];
+/**
+ * 「本社 5階 第2会議室」「研究所 3階 セミナー室」のように、名前の語の後ろに空白で区切って書いた階・部屋の番号。
+ * その前の空白は所在の組を分ける区切りで、日本語と数字のあいだの空け方の好みではない。「地下 1階」「地上 5階」の空白は階の名前の中の空け方なので、区切りと読まない。
+ * 番号は数字だけで、単位の後ろで組が閉じる（空白・読点・文末・英字が続く）ものだけ。tokens は文書全体の座標で、base は text の先頭の位置。
+ */
+export const isSpacedLocationPart = (text, run, tokens, base, words) => {
+    if (tokens === undefined || !SPACES.has(text[run.start - 1] ?? ""))
+        return false;
+    if (!/^\d+$/u.test(text.slice(run.start, run.end)))
+        return false;
+    const unit = unitAfter(text, run, words.units);
+    if (unit === undefined || WORD_CHAR.test(text[run.end + unit.length] ?? ""))
+        return false;
+    const named = tokens.find((token) => token.span.end === base + run.start - 1);
+    return named !== undefined && NAME_POS.has(named.pos) && !words.levels.has(named.surface);
+};
+//# sourceMappingURL=location-part.js.map

package/dist/location-part.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"location-part.js","sourceRoot":"","sources":["../src/location-part.ts"],"names":[],"mappings":"AAEA,wDAAwD;AACxD,MAAM,QAAQ,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC,CAAC;AAE5C,uDAAuD;AACvD,MAAM,SAAS,GAAG,yDAAyD,CAAC;AAE5E,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,CAAC,GAAG,EAAE,GAAG,CAAC,CAAC,CAAC;AAKnC,mDAAmD;AACnD,MAAM,SAAS,GAAG,CAAC,IAAY,EAAE,GAAS,EAAE,KAA0B,EAAsB,EAAE,CAC5F,CAAC,GAAG,KAAK,CAAC,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,IAAI,EAAE,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,CAAC;AAEvH;;;;GAIG;AACH,MAAM,CAAC,MAAM,oBAAoB,GAAG,CAAC,IAAY,EAAE,GAAS,EAAE,MAAoC,EAAE,IAAY,EAAE,KAAiB,EAAW,EAAE;IAC9I,IAAI,MAAM,KAAK,SAAS,IAAI,CAAC,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,KAAK,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC;QAAE,OAAO,KAAK,CAAC;IACjF,IAAI,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,KAAK,EAAE,GAAG,CAAC,GAAG,CAAC,CAAC;QAAE,OAAO,KAAK,CAAC;IACjE,MAAM,IAAI,GAAG,SAAS,CAAC,IAAI,EAAE,GAAG,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC;IAC/C,IAAI,IAAI,KAAK,SAAS,IAAI,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,GAAG,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC;QAAE,OAAO,KAAK,CAAC;IAC1F,MAAM,KAAK,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,KAAK,CAAC,IAAI,CAAC,GAAG,KAAK,IAAI,GAAG,GAAG,CAAC,KAAK,GAAG,CAAC,CAAC,CAAC;IAC9E,OAAO,KAAK,KAAK,SAAS,IAAI,QAAQ,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,GAAG,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;AAC5F,CAAC,CAAC"}

package/dist/nado-closed-list.d.ts

@@ -0,0 +1,20 @@
+import type { LexiconEntry, Token } from "./plugin.ts";
+export type ClosedListWords = {
+    readonly open: readonly string[];
+    /** A limit right after the open word (等に限り, etc. only). */
+    readonly limitsAfter: readonly string[];
+    /** A limit before the list it governs (only A, B, etc.). */
+    readonly limitsBefore: readonly string[];
+    /** A phrase holding a limit's letters that closes nothing (に限らず, not only). */
+    readonly notLimits: readonly string[];
+};
+/** An open word on a closed list: where it is in the sentence, the word and the limit. */
+export type OpenOnClosed = {
+    readonly offset: number;
+    readonly open: string;
+    readonly limit: string;
+};
+export declare const closedListWordsOf: (entries: readonly LexiconEntry[]) => ClosedListWords;
+/** The open words of one sentence that a limit closes, by their offset in the sentence. */
+export declare const openWordsOnClosedLists: (text: string, tokens: readonly Token[], words: ClosedListWords) => OpenOnClosed[];
+//# sourceMappingURL=nado-closed-list.d.ts.map