kordoc 2.7.2 → 2.9.0

package/README.md

@@ -7,7 +7,7 @@
 
 > *대한민국에서 둘째가라면 서러울 문서지옥. 거기서 7년 버틴 공무원이 만들었습니다.*
 
-HWP 3.x/5.x, HWPX, PDF, XLS, XLSX, DOCX — 관공서에서 쏟아지는 모든 문서를 파싱하고, 비교하고, 분석하고, 생성합니다.
+HWP 3.x/5.x, HWPX, HWPML, PDF, XLS, XLSX, DOCX — 관공서에서 쏟아지는 모든 문서를 파싱하고, 비교하고, 분석하고, 생성합니다.
 
 [English](./README-EN.md)
 
@@ -312,6 +312,25 @@ const result = await parse(buffer, {
 })
 ```
 
+### PDF 텍스트 품질 신호 (v2.9.0+)
+
+PDF는 텍스트층이 있어도 ToUnicode/CMap이 깨졌거나 NUL 등 제어문자가 섞이는 경우가 많다. `parsePdf` 결과는 페이지별 품질 신호를 함께 반환한다.
+
+```typescript
+const r = await parsePdf(buffer)
+if (r.success && r.qualitySummary?.needsOcr) {
+  // OCR 큐로 라우팅 (kordoc은 OCR을 기본 탑재하지 않음)
+  await routeToOcr(buffer, r.qualitySummary.ocrCandidatePages)
+}
+
+// 페이지 단위 신호
+for (const p of r.pageQuality ?? []) {
+  if (p.needsOcr) console.log(`p${p.page} 검토 필요: ${p.ocrReason}`)
+}
+```
+
+신호 키: `textChars`, `hangulRatio`, `controlCharRatio`, `replacementCharRatio`, `puaRatio` / `needsOcr` (페이지·문서 단위) / `ocrReason` (`low_text` | `high_pua` | `high_control` | `high_replacement`).
+
 ## CLI
 
 ```bash

package/dist/{chunk-4NWDJGAU.js → chunk-M24KMDAR.js}

@@ -22,7 +22,7 @@ import {
   sanitizeHref,
   stripDtd,
   toArrayBuffer
-} from "./chunk-4SK2PDMQ.js";
+} from "./chunk-SJ5TPMBT.js";
 import {
   parsePageRange
 } from "./chunk-MOL7MDBG.js";
@@ -3928,6 +3928,8 @@ var CHAR_H1 = 5;
 var CHAR_H2 = 6;
 var CHAR_H3 = 7;
 var CHAR_H4 = 8;
+var CHAR_TABLE_HEADER = 9;
+var CHAR_QUOTE = 10;
 var PARA_NORMAL = 0;
 var PARA_H1 = 1;
 var PARA_H2 = 2;
@@ -3936,14 +3938,30 @@ var PARA_H4 = 4;
 var PARA_CODE = 5;
 var PARA_QUOTE = 6;
 var PARA_LIST = 7;
-async function markdownToHwpx(markdown) {
+var DEFAULT_TEXT_COLOR = "#000000";
+function resolveTheme(theme) {
+  return {
+    h1: theme?.headingColors?.[1] ?? DEFAULT_TEXT_COLOR,
+    h2: theme?.headingColors?.[2] ?? DEFAULT_TEXT_COLOR,
+    h3: theme?.headingColors?.[3] ?? DEFAULT_TEXT_COLOR,
+    h4: theme?.headingColors?.[4] ?? theme?.headingColors?.[3] ?? DEFAULT_TEXT_COLOR,
+    body: theme?.bodyColor ?? DEFAULT_TEXT_COLOR,
+    quote: theme?.quoteColor ?? DEFAULT_TEXT_COLOR,
+    /** quoteColor가 명시되었는지 — blockquote charPr 분기에 사용 (baseline 호환) */
+    hasQuoteOption: theme?.quoteColor !== void 0,
+    tableHeader: theme?.tableHeaderColor ?? theme?.bodyColor ?? DEFAULT_TEXT_COLOR,
+    tableHeaderBold: !!theme?.tableHeaderBold
+  };
+}
+async function markdownToHwpx(markdown, options) {
+  const theme = resolveTheme(options?.theme);
   const blocks = parseMarkdownToBlocks(markdown);
-  const sectionXml = blocksToSectionXml(blocks);
+  const sectionXml = blocksToSectionXml(blocks, theme);
   const zip = new JSZip3();
   zip.file("mimetype", "application/hwp+zip", { compression: "STORE" });
   zip.file("META-INF/container.xml", generateContainerXml());
   zip.file("Contents/content.hpf", generateManifest());
-  zip.file("Contents/header.xml", generateHeaderXml());
+  zip.file("Contents/header.xml", generateHeaderXml(theme));
   zip.file("Contents/section0.xml", sectionXml);
   zip.file("Preview/PrvText.txt", buildPrvText(blocks));
   return await zip.generateAsync({ type: "arraybuffer" });
@@ -4123,11 +4141,11 @@ function generateManifest() {
   </opf:spine>
 </opf:package>`;
 }
-function charPr(id, height, bold, italic, fontId = 0) {
+function charPr(id, height, bold, italic, fontId = 0, textColor = DEFAULT_TEXT_COLOR) {
   const boldAttr = bold ? ` bold="1"` : "";
   const italicAttr = italic ? ` italic="1"` : "";
   const effFont = bold ? 2 : fontId;
-  return `      <hh:charPr id="${id}" height="${height}" textColor="#000000" shadeColor="none" useFontSpace="0" useKerning="0" symMark="NONE" borderFillIDRef="0"${boldAttr}${italicAttr}>
+  return `      <hh:charPr id="${id}" height="${height}" textColor="${textColor}" shadeColor="none" useFontSpace="0" useKerning="0" symMark="NONE" borderFillIDRef="0"${boldAttr}${italicAttr}>
         <hh:fontRef hangul="${effFont}" latin="${effFont}" hanja="${effFont}" japanese="${effFont}" other="${effFont}" symbol="${effFont}" user="${effFont}"/>
         <hh:ratio hangul="100" latin="100" hanja="100" japanese="100" other="100" symbol="100" user="100"/>
         <hh:spacing hangul="0" latin="0" hanja="0" japanese="0" other="0" symbol="0" user="0"/>
@@ -4147,7 +4165,7 @@ function paraPr(id, opts = {}) {
         <hh:border borderFillIDRef="0" offsetLeft="0" offsetRight="0" offsetTop="0" offsetBottom="0" connect="0" ignoreMargin="0"/>
       </hh:paraPr>`;
 }
-function generateHeaderXml() {
+function generateHeaderXml(theme) {
   return `<?xml version="1.0" encoding="UTF-8" standalone="yes" ?>
 <hh:head xmlns:hh="${NS_HEAD}" xmlns:hp="${NS_PARA}" version="1.4" secCnt="1">
   <hh:beginNum page="1" footnote="1" endnote="1" pic="1" tbl="1" equation="1"/>
@@ -4223,16 +4241,18 @@ function generateHeaderXml() {
         <hh:fillInfo/>
       </hh:borderFill>
     </hh:borderFills>
-    <hh:charProperties itemCnt="9">
-${charPr(0, 1e3, false, false)}
-${charPr(1, 1e3, true, false)}
-${charPr(2, 1e3, false, true)}
-${charPr(3, 1e3, true, true)}
+    <hh:charProperties itemCnt="11">
+${charPr(0, 1e3, false, false, 0, theme.body)}
+${charPr(1, 1e3, true, false, 0, theme.body)}
+${charPr(2, 1e3, false, true, 0, theme.body)}
+${charPr(3, 1e3, true, true, 0, theme.body)}
 ${charPr(4, 900, false, false, 1)}
-${charPr(5, 1800, true, false, 1)}
-${charPr(6, 1400, true, false, 1)}
-${charPr(7, 1200, true, false, 1)}
-${charPr(8, 1100, true, false, 1)}
+${charPr(5, 1800, true, false, 1, theme.h1)}
+${charPr(6, 1400, true, false, 1, theme.h2)}
+${charPr(7, 1200, true, false, 1, theme.h3)}
+${charPr(8, 1100, true, false, 1, theme.h4)}
+${charPr(CHAR_TABLE_HEADER, 1e3, theme.tableHeaderBold, false, 0, theme.tableHeader)}
+${charPr(CHAR_QUOTE, 1e3, false, true, 0, theme.quote)}
     </hh:charProperties>
     <hh:tabProperties itemCnt="0"/>
     <hh:numberings itemCnt="0"/>
@@ -4262,7 +4282,7 @@ var tableIdCounter = TABLE_ID_BASE;
 function nextTableId() {
   return ++tableIdCounter;
 }
-function generateTable(rows) {
+function generateTable(rows, theme) {
   const rowCnt = rows.length;
   const colCnt = Math.max(...rows.map((r) => r.length), 1);
   const cellW = Math.floor(44e3 / colCnt);
@@ -4270,12 +4290,15 @@ function generateTable(rows) {
   const tblW = cellW * colCnt;
   const tblH = cellH * rowCnt;
   const tblId = nextTableId();
+  const useHeaderStyle = theme.tableHeader !== theme.body || theme.tableHeaderBold;
   const trElements = rows.map((row, rowIdx) => {
     const cells = row.length < colCnt ? [...row, ...Array(colCnt - row.length).fill("")] : row;
+    const isHeaderRow = rowIdx === 0;
+    const headerCharPr = isHeaderRow && useHeaderStyle ? CHAR_TABLE_HEADER : CHAR_NORMAL;
     const tdElements = cells.map((cell, colIdx) => {
-      const runs = generateRuns(cell);
+      const runs = generateRuns(cell, headerCharPr);
       const p = `<hp:p paraPrIDRef="0" styleIDRef="0">${runs}</hp:p>`;
-      return `<hp:tc name="" header="${rowIdx === 0 ? 1 : 0}" hasMargin="0" protect="0" editable="1" dirty="0" borderFillIDRef="1"><hp:subList id="" textDirection="HORIZONTAL" lineWrap="BREAK" vertAlign="TOP" linkListIDRef="0" linkListNextIDRef="0" textWidth="0" textHeight="0" hasTextRef="0" hasNumRef="0">${p}</hp:subList><hp:cellAddr colAddr="${colIdx}" rowAddr="${rowIdx}"/><hp:cellSpan colSpan="1" rowSpan="1"/><hp:cellSz width="${cellW}" height="${cellH}"/><hp:cellMargin left="141" right="141" top="141" bottom="141"/></hp:tc>`;
+      return `<hp:tc name="" header="${isHeaderRow ? 1 : 0}" hasMargin="0" protect="0" editable="1" dirty="0" borderFillIDRef="1"><hp:subList id="" textDirection="HORIZONTAL" lineWrap="BREAK" vertAlign="TOP" linkListIDRef="0" linkListNextIDRef="0" textWidth="0" textHeight="0" hasTextRef="0" hasNumRef="0">${p}</hp:subList><hp:cellAddr colAddr="${colIdx}" rowAddr="${rowIdx}"/><hp:cellSpan colSpan="1" rowSpan="1"/><hp:cellSz width="${cellW}" height="${cellH}"/><hp:cellMargin left="141" right="141" top="141" bottom="141"/></hp:tc>`;
     }).join("");
     return `<hp:tr>${tdElements}</hp:tr>`;
   }).join("");
@@ -4283,7 +4306,7 @@ function generateTable(rows) {
   const tbl = `<hp:tbl id="${tblId}" zOrder="0" numberingType="TABLE" pageBreak="CELL" repeatHeader="0" rowCnt="${rowCnt}" colCnt="${colCnt}" cellSpacing="0" borderFillIDRef="1" noShading="0">${tblInner}</hp:tbl>`;
   return `<hp:p paraPrIDRef="0" styleIDRef="0"><hp:run charPrIDRef="0">${tbl}</hp:run></hp:p>`;
 }
-function blocksToSectionXml(blocks) {
+function blocksToSectionXml(blocks, theme) {
   const paraXmls = [];
   let isFirst = true;
   const orderedCounters = {};
@@ -4312,7 +4335,11 @@ function blocksToSectionXml(blocks) {
         break;
       }
       case "blockquote":
-        xml = generateParagraph(block.text || "", PARA_QUOTE);
+        xml = generateParagraph(
+          block.text || "",
+          PARA_QUOTE,
+          theme.hasQuoteOption ? CHAR_QUOTE : CHAR_NORMAL
+        );
         break;
       case "list_item": {
         const indent = block.indent || 0;
@@ -4345,7 +4372,7 @@ function blocksToSectionXml(blocks) {
             paraXmls.push(`<hp:p paraPrIDRef="0" styleIDRef="0">${secRun}</hp:p>`);
             isFirst = false;
           }
-          xml = generateTable(block.rows);
+          xml = generateTable(block.rows, theme);
         }
         break;
     }
@@ -18714,7 +18741,7 @@ async function parseHwp(buffer, options) {
 async function parsePdf(buffer, options) {
   let parsePdfDocument;
   try {
-    const mod = await import("./parser-QMMQ7Y7R.js");
+    const mod = await import("./parser-XBYGROQB.js");
     parsePdfDocument = mod.parsePdfDocument;
   } catch {
     return {
@@ -18725,8 +18752,8 @@ async function parsePdf(buffer, options) {
     };
   }
   try {
-    const { markdown, blocks, metadata, outline, warnings, isImageBased } = await parsePdfDocument(buffer, options);
-    return { success: true, fileType: "pdf", markdown, blocks, metadata, outline, warnings, isImageBased };
+    const { markdown, blocks, metadata, outline, warnings, isImageBased, pageQuality, qualitySummary } = await parsePdfDocument(buffer, options);
+    return { success: true, fileType: "pdf", markdown, blocks, metadata, outline, warnings, isImageBased, pageQuality, qualitySummary };
   } catch (err) {
     const isImageBased = err instanceof Error && "isImageBased" in err ? true : void 0;
     return { success: false, fileType: "pdf", error: err instanceof Error ? err.message : "PDF \uD30C\uC2F1 \uC2E4\uD328", code: classifyError(err), isImageBased };
@@ -18952,4 +18979,4 @@ export {
   compare,
   parse
 };
-//# sourceMappingURL=chunk-4NWDJGAU.js.map
+//# sourceMappingURL=chunk-M24KMDAR.js.map