kordoc 4.23.0 → 4.24.0

package/dist/{chunk-W67KQWMU.js → chunk-G4HWEM6Q.js}

@@ -10,14 +10,14 @@ import {
   ocrSourcePath,
   parseCharacterDict,
   resolveOcrModels
-} from "./chunk-A5HW2Z5T.js";
+} from "./chunk-PX6X4YOR.js";
 import {
   validateOcrOptions
-} from "./chunk-2URKNOOQ.js";
+} from "./chunk-6YDPTTIY.js";
 import {
   KordocError,
   OPTIONAL_DEP_INSTALL_HINT
-} from "./chunk-Y3LLNEU3.js";
+} from "./chunk-XLY5ZJRP.js";
 
 // src/ocr/ruling-lines.ts
 var INK_LUMA_MAX = 205;
@@ -2212,4 +2212,4 @@ export {
   groupOcrLines,
   ocrItemsToBlocks
 };
-//# sourceMappingURL=chunk-W67KQWMU.js.map
+//# sourceMappingURL=chunk-G4HWEM6Q.js.map

package/dist/{chunk-MSUBZT53.js → chunk-GEY6ZMXH.js}

@@ -51,10 +51,10 @@ import {
   resolveParaHeading,
   simulateWrap,
   takeTocLeaders
-} from "./chunk-4XLDFJGY.js";
+} from "./chunk-CKCJU42Y.js";
 import {
   blocksToPages
-} from "./chunk-LJSKKGWU.js";
+} from "./chunk-EN3PK6OT.js";
 import {
   bmpToPng,
   inlineImagesIntoMarkdown
@@ -76,7 +76,7 @@ import {
   tidyScriptTags,
   unframeLayoutTables,
   wrapScript
-} from "./chunk-GZFL6UOS.js";
+} from "./chunk-P6FNIU5P.js";
 import {
   MAX_COLS,
   MAX_ROWS,
@@ -99,7 +99,7 @@ import {
   noteTableHtml,
   sanitizeText,
   visibleText
-} from "./chunk-5NBUT7GX.js";
+} from "./chunk-DBPDAG7N.js";
 import {
   ANNOTATION_BLANK_RE,
   ValueCursor,
@@ -132,7 +132,7 @@ import {
   resolveUnmatched,
   scanInlineSegments,
   scanSectionXml
-} from "./chunk-S7B74SOM.js";
+} from "./chunk-Z62PHCSU.js";
 import {
   MAX_DECOMPRESS_SIZE,
   MAX_XML_DEPTH,
@@ -149,7 +149,7 @@ import {
   localName,
   rawTextContent,
   readZipEntry
-} from "./chunk-VRIHNSYF.js";
+} from "./chunk-55DOBEGD.js";
 import {
   KordocError,
   classifyError,
@@ -165,7 +165,7 @@ import {
   toArrayBuffer,
   unescapeHtml,
   unzipLimitBytes
-} from "./chunk-UIZYWW3J.js";
+} from "./chunk-Z5G6QNVH.js";
 
 // src/table/visual.ts
 import { readFile as readFile2 } from "fs/promises";
@@ -19329,7 +19329,7 @@ var MIN_SIDE = 24;
 async function ocrEmbeddedImages(blocks, options, warnings) {
   const ocr = options?.ocr;
   if (ocr !== true && ocr !== "force" && typeof ocr !== "function") return false;
-  const { parseImageDocument } = await import("./image-ocr-IYFAHFQH.js");
+  const { parseImageDocument } = await import("./image-ocr-CGJJCD4E.js");
   let inserted = false, unreadable = 0;
   let failed = null;
   const walk = async (list, depth) => {
@@ -27216,27 +27216,35 @@ function gridFromHtml(html) {
   return { rowCnt: rows.length, colCnt, cells };
 }
 function splitContact(line) {
-  if (/[|｜]/.test(line)) return line.split(/\s*[|｜]\s*/).map((s) => s.trim()).filter(Boolean);
   const out = [];
   let rest = line.trim();
   const d = /^\(?\d{4}\.\s*\d{1,2}\.\s*\d{1,2}\.?(?:\s*\([^)]{0,20}\))?\)?[,，]?/u.exec(rest);
   if (d) {
     out.push(d[0].replace(/[,，]$/, ""));
-    rest = rest.slice(d[0].length).trim();
+    rest = rest.slice(d[0].length).replace(/^\s*[|｜]?\s*/, "").trim();
   }
+  if (/[|｜]/.test(rest)) return [...out, ...rest.split(/\s*[|｜]\s*/).map((s) => s.trim()).filter(Boolean)];
   out.push(...rest.split(/(?<=☎\s{0,3}[\d-]{1,20})[\s,，]+/u).map((s) => s.trim()).filter(Boolean));
   return out;
 }
 function fillCells(xml, edits) {
-  return applySplices(xml, edits.filter((e) => e.cell.paras.length).map(({ cell: cell2, lines, post }) => {
+  return applySplices(xml, edits.filter((e) => e.cell.paras.length).map(({ cell: cell2, lines, post, raw }) => {
     const first = xml.slice(cell2.paras[0].start, cell2.paras[0].end);
+    const direct = raw?.(first);
     return {
       start: cell2.paras[0].start,
       end: cell2.paras[cell2.paras.length - 1].end,
-      replacement: (lines.length ? lines : [""]).map((l) => post ? post(fillParagraph(first, l)) : fillParagraph(first, l)).join("")
+      replacement: direct ?? (lines.length ? lines : [""]).map((l) => post ? post(fillParagraph(first, l)) : fillParagraph(first, l)).join("")
     };
   }));
 }
+function replaceCellValue(pXml, oldValue, newValue) {
+  const needle = escapeTextXml(oldValue);
+  const at = pXml.indexOf(needle);
+  if (!oldValue || at < 0 || pXml.indexOf(needle, at + 1) >= 0) return null;
+  if (pXml.lastIndexOf("<", at) > pXml.lastIndexOf(">", at)) return null;
+  return pXml.slice(0, at) + bodyText(newValue) + pXml.slice(at + needle.length);
+}
 function defaultPara(text) {
   return `<hp:p id="0" paraPrIDRef="0" styleIDRef="0" pageBreak="0" columnBreak="0" merged="0"><hp:run charPrIDRef="0">${text ? `<hp:t>${bodyText(text)}</hp:t>` : ""}</hp:run></hp:p>`;
 }
@@ -27399,11 +27407,38 @@ function titleXml(ctx, title, contact) {
 }
 function assignContacts(ctx, slots, contact) {
   const texts = matchContactPieces(slots.map((c) => c.text), contact);
-  return slots.map((cell2, i) => ({ cell: cell2, lines: [texts[i]], post: (p) => undistribute(ctx, p) }));
+  const kept = slots.filter((c, i) => contact && texts[i] && texts[i] === c.text).map((c) => c.text.trim());
+  if (kept.length) ctx.warnings.push(`template: \uC6D0\uACE0\uC5D0 \uC5C6\uB294 \uB2F4\uB2F9 \uC904 \uCE78\uC740 \uACAC\uBCF8 \uADF8\uB300\uB85C \uB460 \u2014 ${kept.join(", ")}`);
+  if (slots.length === 1) return [{ cell: slots[0], lines: [texts[0]], post: (p) => undistribute(ctx, p) }];
+  return slots.flatMap((cell2, i) => {
+    const next = texts[i];
+    if (next === cell2.text) return [];
+    const before = LABELLED_RE.exec(cell2.text.trim()), after = LABELLED_RE.exec(next.trim());
+    const swap = before && after && before[1].trim() === after[1].trim() ? [before[3], after[3]] : !before && !after && DATE_IN_RE.test(cell2.text) && next ? [cell2.text.trim(), next.trim()] : null;
+    return [{ cell: cell2, lines: [next], ...swap ? { raw: (first) => replaceCellValue(first, swap[0], swap[1]) } : {} }];
+  });
+}
+var LABELLED_RE = /^(.+?)([\s　]*[:：][\s　]*)(\S.*)$/u;
+function parseContactPiece(piece) {
+  const m = LABELLED_RE.exec(piece.trim());
+  if (m) {
+    const kw = ROLE_KW_RE.exec(m[1])?.[1];
+    return kw ? { kw, label: m[1].trim(), value: m[3].trim() } : null;
+  }
+  const k = ROLE_KW_RE.exec(piece);
+  if (!k) return null;
+  const end = (k.index ?? 0) + k[1].length;
+  const value = piece.slice(end).trim();
+  return value ? { kw: k[1], label: piece.slice(0, end).trim(), value } : null;
 }
+var STAFF_KW = /* @__PURE__ */ new Set(["\uB2F4\uB2F9", "\uC8FC\uBB34\uAD00"]);
 function matchContactPieces(slotTexts, contact) {
   const out = slotTexts.map(() => "");
   if (!contact || !slotTexts.length) return out;
+  if (slotTexts.length > 1) {
+    const done = matchRoleSlots(slotTexts, splitContact(contact));
+    if (done) return done;
+  }
   const pieces = slotTexts.length === 1 ? [splitContact(contact).join("  ")] : splitContact(contact);
   const free = slotTexts.map((_, i) => i);
   const take = (pred) => {
@@ -27424,6 +27459,28 @@ function matchContactPieces(slotTexts, contact) {
   }
   return out;
 }
+function matchRoleSlots(slotTexts, pieces) {
+  const slots = slotTexts.map((t) => {
+    const date = DATE_IN_RE.test(t) && !ROLE_KW_RE.test(t);
+    const m = LABELLED_RE.exec(t.trim());
+    return { date, kw: m ? ROLE_KW_RE.exec(m[1])?.[1] ?? null : null, label: m?.[1].trim() ?? "", sep: m?.[2] ?? " : " };
+  });
+  if (slots.some((s) => !s.date && !s.kw)) return null;
+  const out = slotTexts.slice();
+  for (const piece of pieces) {
+    if (DATE_RE.test(piece)) {
+      const i2 = slots.findIndex((s) => s.date);
+      if (i2 >= 0) out[i2] = piece;
+      continue;
+    }
+    const entry = parseContactPiece(piece);
+    if (!entry) return null;
+    const i = slots.findIndex((s) => s.kw === entry.kw);
+    if (i < 0) return null;
+    out[i] = `${STAFF_KW.has(entry.kw) ? slots[i].label : entry.label}${slots[i].sep}${entry.value}`;
+  }
+  return out;
+}
 function contactXml(ctx, contact) {
   const proto = ctx.scan.protos.get("contact");
   if (!proto) return bodyPara(ctx, contact);
@@ -28548,10 +28605,13 @@ var RULES = [
   // 숫자 런 첫머리에서만 시작 — 아니면 긴 숫자 줄에서 시작 자리마다 런을 다시 훑어 이차 폭주(숫자 5만 자 2.9초)
   // "359,600천원"·"1,200천원" 처럼 쉼표를 둔 수 뒤 천원은 예산서 단위 표기(단위: 천원)라 줄임말이 아니다 —
   // 쉼표 묶음 뒤(앞이 숫자·쉼표)는 건너뛴다. 걸리는 것은 "3천원"·"1500천원" 같은 쉼표 없는 수
+  // 표 칸("금액(천원)" 열)·산출근거 곱셈 줄("× … = 900천원")의 천원도 예산서 단위 표기다 — 건너뛴다
   {
     code: "MONEY_CHEONWON",
     severity: "error",
     pattern: /(?<![\d,])\d+\s*천\s*원/g,
+    skipTable: true,
+    check: (m) => /[×✕]|\s=\s/.test(m.input ?? "") ? null : "",
     message: "\uAE08\uC561\uC740 '\uCC9C\uC6D0'\uC73C\uB85C \uC904\uC774\uC9C0 \uC54A\uACE0 \uC544\uB77C\uBE44\uC544 \uC22B\uC790\uB85C",
     suggest: "\uC608) 345,000\uC6D0"
   },
@@ -29526,7 +29586,8 @@ function checkCalc(line, no, out) {
       rule: "BUDGET_CALC_MISMATCH",
       severity: "error",
       message: "\uC0B0\uCD9C\uADFC\uAC70 \uACF1\uC148 \uACB0\uACFC\uAC00 \uC801\uC740 \uAE08\uC561\uACFC \uB2E4\uB984",
-      suggest: `${m[1].trim()} = ${fmt(expected)}${m[3]}`
+      // "원래 식 → 바른 식" — 그대로 바꿔 넣을 수 있는 꼴(편집 도구가 바로 반영)
+      suggest: `${m[0]} \u2192 ${m[1].trim()} = ${fmt(expected)}${m[3]}`
     });
   }
 }
@@ -29836,6 +29897,104 @@ function runLintSuite(text, opts = {}) {
   };
 }
 
+// src/hwpx/lint-blocks.ts
+var MARKDOWN_STRUCTURE = /^\s*(?:```|~~~|#|\||>)/;
+var safeLine = (text) => text.replace(/[\r\n]/g, " ");
+var safeCell = (text) => safeLine(text).replace(/\|/g, "\uFF5C");
+function blocksToLines(blocks) {
+  const lines = [];
+  const paragraph = (block) => {
+    const text = safeLine(block.text);
+    const prefix = MARKDOWN_STRUCTURE.test(text) ? "\\" : "";
+    lines.push({ text: prefix + text, segments: [{ start: prefix.length, length: text.length, block: block.id }] });
+  };
+  const tables = /* @__PURE__ */ new Map();
+  for (const b of blocks) if (b.table) (tables.get(b.table.id) ?? tables.set(b.table.id, []).get(b.table.id)).push(b);
+  const emitted = /* @__PURE__ */ new Set();
+  for (const b of blocks) {
+    if (!b.table) {
+      paragraph(b);
+      continue;
+    }
+    if (emitted.has(b.table.id)) continue;
+    emitted.add(b.table.id);
+    const cells = tables.get(b.table.id);
+    if (cells.length === 1) {
+      paragraph(cells[0]);
+      continue;
+    }
+    const rows = cells.reduce((n, c) => Math.max(n, c.table.row + 1), 0);
+    const cols = cells.reduce((n, c) => Math.max(n, c.table.col + 1), 0);
+    const at = new Map(cells.map((c) => [`${c.table.row},${c.table.col}`, c]));
+    lines.push({ text: "", segments: [] });
+    for (let r = 0; r < rows; r++) {
+      let text = "|";
+      const segments = [];
+      for (let c = 0; c < cols; c++) {
+        text += " ";
+        const cell2 = at.get(`${r},${c}`);
+        if (cell2) {
+          const t = safeCell(cell2.text);
+          segments.push({ start: text.length, length: t.length, block: cell2.id });
+          text += t;
+        }
+        text += " |";
+      }
+      lines.push({ text, segments });
+      if (r === 0) lines.push({ text: "|" + " --- |".repeat(cols), segments: [] });
+    }
+    lines.push({ text: "", segments: [] });
+  }
+  return lines;
+}
+function locate(lines, f) {
+  const line = lines[f.line - 1];
+  if (!line) return { block: null, offset: null };
+  if (f.column !== void 0) {
+    const at = f.column - 1;
+    const seg2 = line.segments.find((s) => at >= s.start && at < s.start + Math.max(1, s.length)) ?? line.segments.find((s) => line.text.indexOf(f.match, s.start) >= s.start && line.text.indexOf(f.match, s.start) < s.start + s.length);
+    if (seg2) {
+      const found = line.text.indexOf(f.match, Math.max(seg2.start, at));
+      const start = found >= seg2.start && found < seg2.start + seg2.length ? found : at;
+      return { block: seg2.block, offset: Math.min(Math.max(0, start - seg2.start), seg2.length) };
+    }
+    return { block: null, offset: null };
+  }
+  if (line.segments.length !== 1) return { block: null, offset: null };
+  const seg = line.segments[0];
+  const body = line.text.slice(seg.start, seg.start + seg.length);
+  const first = f.match ? body.indexOf(f.match) : -1;
+  const unique = first >= 0 && body.indexOf(f.match, first + 1) < 0;
+  return { block: seg.block, offset: unique ? first : null };
+}
+function lintBlocks(blocks, opts = {}) {
+  const lines = blocksToLines(blocks);
+  const suite = runLintSuite(lines.map((l) => l.text).join("\n") + "\n", opts);
+  const findings = [];
+  const take = (category, list) => {
+    for (const f of list ?? []) {
+      const { block, offset } = locate(lines, f);
+      findings.push({
+        block,
+        offset,
+        length: f.match.length,
+        category,
+        rule: f.rule,
+        severity: f.severity,
+        match: f.match,
+        message: f.message,
+        ...f.suggest ? { suggest: f.suggest } : {}
+      });
+    }
+  };
+  take("notation", suite.findings);
+  take("munche", suite.munche);
+  take("slop", suite.slop);
+  take("logic", suite.logic);
+  take("budget", suite.budget);
+  return { findings, ...suite.numbers ? { numbers: suite.numbers } : {}, summary: suite.summary };
+}
+
 // src/roundtrip/patcher.ts
 import JSZip11 from "jszip";
 
@@ -32880,7 +33039,7 @@ function scriptsOff(r, off) {
 }
 async function parseImage(buffer, options) {
   try {
-    const { parseImageDocument } = await import("./image-ocr-IYFAHFQH.js");
+    const { parseImageDocument } = await import("./image-ocr-CGJJCD4E.js");
     const { blocks, warnings, lines } = await parseImageDocument(buffer, options);
     stripScriptTags({ blocks });
     return {
@@ -32944,7 +33103,7 @@ async function parseHwp(buffer, options) {
 async function parsePdf(buffer, options) {
   let parsePdfDocument;
   try {
-    const mod = await import("./parser-OQ4FLNBE.js");
+    const mod = await import("./parser-IHCFHLHN.js");
     parsePdfDocument = mod.parsePdfDocument;
   } catch {
     return {
@@ -33153,6 +33312,7 @@ export {
   formatGateFinding,
   lintBudgetText,
   runLintSuite,
+  lintBlocks,
   replaceOleStream,
   wipeFreeSectors,
   patchHwp,
@@ -33176,4 +33336,4 @@ export {
   parseHwpml,
   fillForm
 };
-//# sourceMappingURL=chunk-MSUBZT53.js.map
+//# sourceMappingURL=chunk-GEY6ZMXH.js.map