kordoc 4.15.7 → 4.16.1
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +346 -1124
- package/dist/{-H6WNRHK6.js → -3DURLXSO.js} +11 -11
- package/dist/{chunk-YFHV4QP3.js → chunk-24N4IJ5I.js} +5 -4
- package/dist/chunk-24N4IJ5I.js.map +1 -0
- package/dist/{chunk-R45AG5K6.cjs → chunk-2BZFKIRB.cjs} +3 -3
- package/dist/chunk-2BZFKIRB.cjs.map +1 -0
- package/dist/{chunk-F5FHYJKJ.js → chunk-3VLVNKJV.js} +4 -2
- package/dist/{chunk-WQWXQVOQ.js.map → chunk-3VLVNKJV.js.map} +1 -1
- package/dist/{chunk-6ULM2YCN.js → chunk-3WIQSM3Z.js} +686 -24
- package/dist/chunk-3WIQSM3Z.js.map +1 -0
- package/dist/{chunk-AWXXLLIG.js → chunk-63YGGAQW.js} +2 -2
- package/dist/{chunk-WMLGS5QL.js → chunk-7FKPPIDZ.js} +245 -119
- package/dist/chunk-7FKPPIDZ.js.map +1 -0
- package/dist/{chunk-4ANWGSUT.cjs → chunk-7IRHSRPU.cjs} +370 -163
- package/dist/chunk-7IRHSRPU.cjs.map +1 -0
- package/dist/{chunk-IHCQ4IE7.js → chunk-7XAGIPOX.js} +2 -2
- package/dist/{chunk-NZMPU24H.js → chunk-ANZTLTKU.js} +5 -4
- package/dist/chunk-ANZTLTKU.js.map +1 -0
- package/dist/{chunk-KJSD2WNF.js → chunk-BIHS4VWR.js} +15 -2
- package/dist/chunk-BIHS4VWR.js.map +1 -0
- package/dist/{chunk-C572DIHX.js → chunk-BTO7ZEE4.js} +13 -6
- package/dist/{chunk-C572DIHX.js.map → chunk-BTO7ZEE4.js.map} +1 -1
- package/dist/{chunk-O5SDUPB2.js → chunk-CXWWX6BQ.js} +2 -2
- package/dist/{chunk-OE6XLTMC.cjs → chunk-D2QCXK2X.cjs} +9 -8
- package/dist/chunk-D2QCXK2X.cjs.map +1 -0
- package/dist/{chunk-ZYJY44BT.js → chunk-DJ2BBQ5P.js} +690 -84
- package/dist/chunk-DJ2BBQ5P.js.map +1 -0
- package/dist/{chunk-YNJLZVEO.cjs → chunk-DO74GYSM.cjs} +1 -1
- package/dist/chunk-DO74GYSM.cjs.map +1 -0
- package/dist/{chunk-UGSRP7E2.js → chunk-EEVAPRPD.js} +3 -3
- package/dist/{chunk-XWA625AN.js → chunk-FSDJLAOW.js} +3 -1
- package/dist/chunk-FSDJLAOW.js.map +1 -0
- package/dist/chunk-GS7T56RP.cjs.map +1 -1
- package/dist/chunk-J55UQC5H.cjs.map +1 -1
- package/dist/{chunk-W2UTB3RG.js → chunk-KBAQLAEL.js} +1 -1
- package/dist/chunk-KBAQLAEL.js.map +1 -0
- package/dist/{chunk-ZWPRA2SE.js → chunk-N65ZXWNQ.js} +2 -2
- package/dist/{chunk-ORMMGTNK.js → chunk-NCXRSQNE.js} +2 -2
- package/dist/{chunk-WQWXQVOQ.js → chunk-NUPXPKIQ.js} +4 -2
- package/dist/{chunk-F5FHYJKJ.js.map → chunk-NUPXPKIQ.js.map} +1 -1
- package/dist/{chunk-ZXAH4AFS.cjs → chunk-NV37CQFN.cjs} +698 -92
- package/dist/chunk-NV37CQFN.cjs.map +1 -0
- package/dist/{chunk-S7PEBTU3.js → chunk-OBRPYCIS.js} +1503 -985
- package/dist/chunk-OBRPYCIS.js.map +1 -0
- package/dist/{chunk-X3PBLSV5.js → chunk-PVMROO3A.js} +319 -112
- package/dist/chunk-PVMROO3A.js.map +1 -0
- package/dist/{chunk-NVG2GQNA.js → chunk-T77M2QP7.js} +1 -1
- package/dist/chunk-T77M2QP7.js.map +1 -0
- package/dist/{chunk-WHZUJH2O.js → chunk-UZHT3LBF.js} +5 -4
- package/dist/{chunk-WHZUJH2O.js.map → chunk-UZHT3LBF.js.map} +1 -1
- package/dist/{chunk-L5MDTA6T.js → chunk-VCRBR72A.js} +2 -2
- package/dist/{chunk-TRB6OYMU.js → chunk-VNOQTDTE.js} +2 -2
- package/dist/{chunk-J6ZU7AIR.cjs → chunk-YRRBA4HW.cjs} +5 -3
- package/dist/chunk-YRRBA4HW.cjs.map +1 -0
- package/dist/chunks-PYYQURNO.js +10 -0
- package/dist/cli.js +40 -36
- package/dist/cli.js.map +1 -1
- package/dist/{formula-UKLCJINT.js → formula-IPLFHEBD.js} +3 -3
- package/dist/{formula-YJIUNSWW.cjs → formula-TRTONZMA.cjs} +9 -9
- package/dist/formula-TRTONZMA.cjs.map +1 -0
- package/dist/{formula-CUSB5R3P.js → formula-U4BTEZ6I.js} +4 -4
- package/dist/{image-ocr-OEEGTGDU.js → image-ocr-3IJ25PS4.js} +14 -9
- package/dist/image-ocr-3IJ25PS4.js.map +1 -0
- package/dist/{image-ocr-S5OLK6WH.js → image-ocr-GTDG4SAJ.js} +9 -11
- package/dist/image-ocr-GTDG4SAJ.js.map +1 -0
- package/dist/{image-ocr-BMCOWLLB.cjs → image-ocr-THFLQ3DC.cjs} +17 -15
- package/dist/image-ocr-THFLQ3DC.cjs.map +1 -0
- package/dist/index.cjs +1905 -1379
- package/dist/index.cjs.map +1 -1
- package/dist/index.d.cts +19 -4
- package/dist/index.d.ts +19 -4
- package/dist/index.js +1352 -826
- package/dist/index.js.map +1 -1
- package/dist/mcp.js +69 -54
- package/dist/mcp.js.map +1 -1
- package/dist/{model-bundle-A6QPDJZ3.js → model-bundle-PVWIAVEX.js} +5 -5
- package/dist/{models-GAP4U6WU.js → models-VCRF3JMT.js} +7 -5
- package/dist/page-range-QHHFTL7X.cjs.map +1 -1
- package/dist/{parser-DZMXH4V5.cjs → parser-2CN2DIPV.cjs} +633 -177
- package/dist/parser-2CN2DIPV.cjs.map +1 -0
- package/dist/{parser-MRTO7R2Y.js → parser-T4RITZDP.js} +512 -53
- package/dist/parser-T4RITZDP.js.map +1 -0
- package/dist/{parser-BCN3SZZC.js → parser-UEROAPYZ.js} +508 -52
- package/dist/parser-UEROAPYZ.js.map +1 -0
- package/dist/pdf-ocr-NARV2AQI.cjs +16 -0
- package/dist/pdf-ocr-NARV2AQI.cjs.map +1 -0
- package/dist/pdf-ocr-SYLMM727.js +18 -0
- package/dist/pdf-ocr-YMQZDJBO.js +15 -0
- package/dist/{profile-io-XAZ2MDXD.js → profile-io-YD54CDXY.js} +3 -3
- package/dist/{rasterize-2YKKVEIR.js → rasterize-3THYOOUN.js} +3 -3
- package/dist/{redact-doc-XPV546P6.js → redact-doc-65L2YJN6.js} +11 -11
- package/dist/{render-ASDQSRPX.js → render-YU7QJVZ7.js} +7 -7
- package/dist/seal-6CQGFTML.js +10 -0
- package/dist/{setup-M3MH2DHU.js → setup-CFHKXDY4.js} +3 -3
- package/dist/visual-ED2H75Q5.js +20 -0
- package/dist/{watch-3DB2LMAP.js → watch-D6FCAYKJ.js} +41 -19
- package/dist/watch-D6FCAYKJ.js.map +1 -0
- package/package.json +1 -1
- package/dist/chunk-4ANWGSUT.cjs.map +0 -1
- package/dist/chunk-6ULM2YCN.js.map +0 -1
- package/dist/chunk-J6ZU7AIR.cjs.map +0 -1
- package/dist/chunk-KJSD2WNF.js.map +0 -1
- package/dist/chunk-NVG2GQNA.js.map +0 -1
- package/dist/chunk-NZMPU24H.js.map +0 -1
- package/dist/chunk-OE6XLTMC.cjs.map +0 -1
- package/dist/chunk-R45AG5K6.cjs.map +0 -1
- package/dist/chunk-S7PEBTU3.js.map +0 -1
- package/dist/chunk-W2UTB3RG.js.map +0 -1
- package/dist/chunk-WMLGS5QL.js.map +0 -1
- package/dist/chunk-X3PBLSV5.js.map +0 -1
- package/dist/chunk-XWA625AN.js.map +0 -1
- package/dist/chunk-YFHV4QP3.js.map +0 -1
- package/dist/chunk-YNJLZVEO.cjs.map +0 -1
- package/dist/chunk-ZXAH4AFS.cjs.map +0 -1
- package/dist/chunk-ZYJY44BT.js.map +0 -1
- package/dist/chunks-PPHEGDUD.js +0 -10
- package/dist/formula-YJIUNSWW.cjs.map +0 -1
- package/dist/image-ocr-BMCOWLLB.cjs.map +0 -1
- package/dist/image-ocr-OEEGTGDU.js.map +0 -1
- package/dist/image-ocr-S5OLK6WH.js.map +0 -1
- package/dist/parser-BCN3SZZC.js.map +0 -1
- package/dist/parser-DZMXH4V5.cjs.map +0 -1
- package/dist/parser-MRTO7R2Y.js.map +0 -1
- package/dist/pdf-ocr-4V7U52JI.js +0 -18
- package/dist/pdf-ocr-A3Q4YMCG.js +0 -15
- package/dist/pdf-ocr-CZDODSMI.cjs +0 -16
- package/dist/pdf-ocr-CZDODSMI.cjs.map +0 -1
- package/dist/seal-3XWSKD6C.js +0 -10
- package/dist/visual-Z6BCNJMU.js +0 -20
- package/dist/watch-3DB2LMAP.js.map +0 -1
- /package/dist/{-H6WNRHK6.js.map → -3DURLXSO.js.map} +0 -0
- /package/dist/{chunk-AWXXLLIG.js.map → chunk-63YGGAQW.js.map} +0 -0
- /package/dist/{chunk-IHCQ4IE7.js.map → chunk-7XAGIPOX.js.map} +0 -0
- /package/dist/{chunk-O5SDUPB2.js.map → chunk-CXWWX6BQ.js.map} +0 -0
- /package/dist/{chunk-UGSRP7E2.js.map → chunk-EEVAPRPD.js.map} +0 -0
- /package/dist/{chunk-ZWPRA2SE.js.map → chunk-N65ZXWNQ.js.map} +0 -0
- /package/dist/{chunk-ORMMGTNK.js.map → chunk-NCXRSQNE.js.map} +0 -0
- /package/dist/{chunk-L5MDTA6T.js.map → chunk-VCRBR72A.js.map} +0 -0
- /package/dist/{chunk-TRB6OYMU.js.map → chunk-VNOQTDTE.js.map} +0 -0
- /package/dist/{chunks-PPHEGDUD.js.map → chunks-PYYQURNO.js.map} +0 -0
- /package/dist/{formula-UKLCJINT.js.map → formula-IPLFHEBD.js.map} +0 -0
- /package/dist/{formula-CUSB5R3P.js.map → formula-U4BTEZ6I.js.map} +0 -0
- /package/dist/{model-bundle-A6QPDJZ3.js.map → model-bundle-PVWIAVEX.js.map} +0 -0
- /package/dist/{models-GAP4U6WU.js.map → models-VCRF3JMT.js.map} +0 -0
- /package/dist/{pdf-ocr-4V7U52JI.js.map → pdf-ocr-SYLMM727.js.map} +0 -0
- /package/dist/{pdf-ocr-A3Q4YMCG.js.map → pdf-ocr-YMQZDJBO.js.map} +0 -0
- /package/dist/{profile-io-XAZ2MDXD.js.map → profile-io-YD54CDXY.js.map} +0 -0
- /package/dist/{rasterize-2YKKVEIR.js.map → rasterize-3THYOOUN.js.map} +0 -0
- /package/dist/{redact-doc-XPV546P6.js.map → redact-doc-65L2YJN6.js.map} +0 -0
- /package/dist/{render-ASDQSRPX.js.map → render-YU7QJVZ7.js.map} +0 -0
- /package/dist/{seal-3XWSKD6C.js.map → seal-6CQGFTML.js.map} +0 -0
- /package/dist/{setup-M3MH2DHU.js.map → setup-CFHKXDY4.js.map} +0 -0
- /package/dist/{visual-Z6BCNJMU.js.map → visual-ED2H75Q5.js.map} +0 -0
|
@@ -50,17 +50,17 @@ import {
|
|
|
50
50
|
unknownFontWarnings,
|
|
51
51
|
usesGaejosikMunche,
|
|
52
52
|
validateHwpx
|
|
53
|
-
} from "./chunk-
|
|
53
|
+
} from "./chunk-OBRPYCIS.js";
|
|
54
54
|
import {
|
|
55
55
|
blocksToPages
|
|
56
|
-
} from "./chunk-
|
|
56
|
+
} from "./chunk-63YGGAQW.js";
|
|
57
57
|
import {
|
|
58
58
|
applySplices,
|
|
59
59
|
buildParagraphSplices,
|
|
60
60
|
buildRangeSplices,
|
|
61
61
|
placeSealHwpx,
|
|
62
62
|
scanSectionXml
|
|
63
|
-
} from "./chunk-
|
|
63
|
+
} from "./chunk-EEVAPRPD.js";
|
|
64
64
|
import {
|
|
65
65
|
DEFAULT_REDACT_RULES,
|
|
66
66
|
redactMarkdown,
|
|
@@ -68,7 +68,7 @@ import {
|
|
|
68
68
|
} from "./chunk-3WSNLH4B.js";
|
|
69
69
|
import {
|
|
70
70
|
blocksToChunks
|
|
71
|
-
} from "./chunk-
|
|
71
|
+
} from "./chunk-VCRBR72A.js";
|
|
72
72
|
import {
|
|
73
73
|
SPACE_EM_FIXED,
|
|
74
74
|
SPACE_EM_FONT,
|
|
@@ -93,20 +93,20 @@ import {
|
|
|
93
93
|
renderSceneToHtml,
|
|
94
94
|
simulateWrap,
|
|
95
95
|
simulateWrapKeepWord
|
|
96
|
-
} from "./chunk-
|
|
96
|
+
} from "./chunk-BTO7ZEE4.js";
|
|
97
97
|
import "./chunk-7FCNOKOR.js";
|
|
98
|
-
import "./chunk-
|
|
98
|
+
import "./chunk-T77M2QP7.js";
|
|
99
99
|
import "./chunk-VQFA55IC.js";
|
|
100
|
-
import "./chunk-
|
|
100
|
+
import "./chunk-CXWWX6BQ.js";
|
|
101
101
|
import {
|
|
102
102
|
blocksToMarkdown,
|
|
103
103
|
flattenLayoutTables,
|
|
104
104
|
hasStructuredCellContent
|
|
105
|
-
} from "./chunk-
|
|
106
|
-
import "./chunk-
|
|
105
|
+
} from "./chunk-UZHT3LBF.js";
|
|
106
|
+
import "./chunk-24N4IJ5I.js";
|
|
107
107
|
import {
|
|
108
108
|
VERSION
|
|
109
|
-
} from "./chunk-
|
|
109
|
+
} from "./chunk-3VLVNKJV.js";
|
|
110
110
|
export {
|
|
111
111
|
BUILTIN_TEMPLATES,
|
|
112
112
|
DEFAULT_REDACT_RULES,
|
|
@@ -196,4 +196,4 @@ export {
|
|
|
196
196
|
usesGaejosikMunche,
|
|
197
197
|
validateHwpx
|
|
198
198
|
};
|
|
199
|
-
//# sourceMappingURL=-
|
|
199
|
+
//# sourceMappingURL=-3DURLXSO.js.map
|
|
@@ -1,7 +1,8 @@
|
|
|
1
1
|
#!/usr/bin/env node
|
|
2
2
|
import {
|
|
3
|
-
KordocError
|
|
4
|
-
|
|
3
|
+
KordocError,
|
|
4
|
+
OPTIONAL_DEP_INSTALL_HINT
|
|
5
|
+
} from "./chunk-3VLVNKJV.js";
|
|
5
6
|
|
|
6
7
|
// src/render/rasterize.ts
|
|
7
8
|
var densityExponent = null;
|
|
@@ -40,7 +41,7 @@ async function loadSharp() {
|
|
|
40
41
|
return mod.default ?? mod;
|
|
41
42
|
} catch {
|
|
42
43
|
throw new KordocError(
|
|
43
|
-
'PNG \uB798\uC2A4\uD130\uC5D0\uB294 sharp\uAC00 \uD544\uC694\uD569\uB2C8\uB2E4 (npm install sharp). sharp \uC5C6\uC774 \uC4F0\uB824\uBA74 format: "svg" + output_path\uB85C SVG \uD30C\uC77C \uC800\uC7A5\uC744 \uC0AC\uC6A9\uD558\uC138\uC694'
|
|
44
|
+
'PNG \uB798\uC2A4\uD130\uC5D0\uB294 sharp\uAC00 \uD544\uC694\uD569\uB2C8\uB2E4 (npm install sharp). sharp \uC5C6\uC774 \uC4F0\uB824\uBA74 format: "svg" + output_path\uB85C SVG \uD30C\uC77C \uC800\uC7A5\uC744 \uC0AC\uC6A9\uD558\uC138\uC694.' + OPTIONAL_DEP_INSTALL_HINT
|
|
44
45
|
);
|
|
45
46
|
}
|
|
46
47
|
}
|
|
@@ -87,4 +88,4 @@ export {
|
|
|
87
88
|
rasterizeSvg,
|
|
88
89
|
rasterizePageSvg
|
|
89
90
|
};
|
|
90
|
-
//# sourceMappingURL=chunk-
|
|
91
|
+
//# sourceMappingURL=chunk-24N4IJ5I.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"sources":["../src/render/rasterize.ts"],"sourcesContent":["/** SVG → PNG/JPEG 래스터 — sharp optional 의존 (미설치 시 KordocError, SVG 경로는 sharp 불필요) */\n\nimport { KordocError, OPTIONAL_DEP_INSTALL_HINT } from \"../utils.js\"\n\nexport interface RasterizeOptions {\n /** 출력 최대 폭 px (기본 1400) */\n maxWidthPx?: number\n /** 출력 최대 높이 px (기본 8000 — 멀티페이지 세로 스택 대비) */\n maxHeightPx?: number\n /** PNG 바이트 상한 (기본 4MB) — 초과 시 절반 스케일로 1회 재시도 */\n maxBytes?: number\n}\n\nexport interface RasterizeResult {\n png: Buffer\n widthPx: number\n heightPx: number\n /** pt → px 배율 */\n scale: number\n}\n\n/** sharp 최소 타입 — optional 의존이라 정적 import 없이 동적 로드 */\nexport type SharpLike = (input: Buffer, opts?: { density?: number; limitInputPixels?: number }) => {\n png(): { toBuffer(): Promise<Buffer> }\n jpeg(opts?: { quality?: number }): { toBuffer(): Promise<Buffer> }\n extract(region: { left: number; top: number; width: number; height: number }): ReturnType<SharpLike>\n toBuffer(): Promise<Buffer>\n metadata(): Promise<{ width?: number; height?: number }>\n}\n\n/**\n * pt 단위 SVG 의 density 배율 지수 — libvips/rsvg 조합에 따라 `density: 72×s` 가 s px/pt(선형)를 내기도,\n * s² px/pt 를 내기도 한다(sharp 0.35·vips 8.18·rsvg 2.62: pt 단위는 제곱, px 단위는 선형 — 실측 2026-09-06).\n * 배율을 가정하면 crop 픽셀 사각형이 실제 래스터와 어긋나 표 영역의 좌상단 1/4 만 잘려 나온다(v4.13.0 실사고).\n * 프로세스당 1회 프로브로 지수를 재고, 산출 픽셀 크기는 항상 실제 이미지에서 다시 읽는다.\n */\nlet densityExponent: number | null = null\nconst PROBE_SVG = `<svg xmlns=\"http://www.w3.org/2000/svg\" viewBox=\"0 0 100 100\" width=\"100pt\" height=\"100pt\"><rect width=\"100\" height=\"100\" fill=\"#fff\"/></svg>`\n\n/** PNG IHDR 폭 (바이트 16~19, BE) — 프로브는 metadata 없이 읽는다 */\nfunction pngWidth(buf: Buffer): number {\n return buf.length >= 24 && buf[0] === 0x89 && buf[1] === 0x50 ? buf.readUInt32BE(16) : 0\n}\n\nasync function probeDensityExponent(sharp: SharpLike): Promise<number> {\n if (densityExponent !== null) return densityExponent\n try {\n const png = await sharp(Buffer.from(PROBE_SVG), { density: 144 }).png().toBuffer()\n const k = pngWidth(png) / 100\n const e = k > 0 ? Math.log(k) / Math.log(2) : 1\n densityExponent = Number.isFinite(e) && e > 0.5 ? Math.round(e * 4) / 4 : 1\n } catch {\n densityExponent = 1\n }\n return densityExponent\n}\n\n/** 요청 배율(px/pt) → sharp density */\nfunction densityFor(scale: number, exponent: number): number {\n return 72 * Math.pow(scale, 1 / exponent)\n}\n\n/** 실제 픽셀 크기 — 보고값은 가정이 아니라 산출 이미지에서 읽는다 */\nasync function actualSize(sharp: SharpLike, data: Buffer, fallbackW: number, fallbackH: number): Promise<{ w: number; h: number }> {\n const w = pngWidth(data)\n if (w > 0 && data.length >= 24) return { w, h: data.readUInt32BE(20) }\n try {\n const m = await sharp(data).metadata()\n if (m.width && m.height) return { w: m.width, h: m.height }\n } catch { /* 폴백 */ }\n return { w: fallbackW, h: fallbackH }\n}\n\nexport async function loadSharp(): Promise<SharpLike> {\n try {\n const mod: any = await import(\"sharp\")\n return mod.default ?? mod\n } catch {\n throw new KordocError(\n 'PNG 래스터에는 sharp가 필요합니다 (npm install sharp). sharp 없이 쓰려면 format: \"svg\" + output_path로 SVG 파일 저장을 사용하세요.' +\n OPTIONAL_DEP_INSTALL_HINT,\n )\n }\n}\n\n/** SVG 문자열을 PNG로 래스터. widthPt/heightPt는 SVG 캔버스 크기(pt). */\nexport async function rasterizeSvg(\n svg: string,\n widthPt: number,\n heightPt: number,\n options?: RasterizeOptions,\n): Promise<RasterizeResult> {\n const sharp = await loadSharp()\n if (!(widthPt > 0) || !(heightPt > 0)) throw new KordocError(`잘못된 SVG 크기: ${widthPt}x${heightPt}pt`)\n const maxW = options?.maxWidthPx ?? 1400\n const maxH = options?.maxHeightPx ?? 8000\n const maxBytes = options?.maxBytes ?? 4 * 1024 * 1024\n\n const exponent = await probeDensityExponent(sharp)\n const render = async (scale: number): Promise<RasterizeResult> => {\n // sharp 의 SVG 기본 밀도 72DPI 기준. 배율 지수는 프로브 값, 픽셀 크기·실배율은 산출물에서 읽는다\n const png = await sharp(Buffer.from(svg), { density: densityFor(scale, exponent), limitInputPixels: 268402689 })\n .png()\n .toBuffer()\n const { w, h } = await actualSize(sharp, png, Math.round(widthPt * scale), Math.round(heightPt * scale))\n return { png, widthPx: w, heightPx: h, scale: w / widthPt }\n }\n\n const scale = Math.min(maxW / widthPt, maxH / heightPt, 2)\n let result = await render(scale)\n if (result.png.length > maxBytes && scale > 0.25) {\n result = await render(scale / 2)\n }\n if (result.png.length > maxBytes) {\n throw new KordocError(\n `PNG가 상한(${(maxBytes / 1024 / 1024).toFixed(0)}MB)을 초과합니다 (${(result.png.length / 1024 / 1024).toFixed(1)}MB) — format: \"svg\" + output_path로 파일 저장을 사용하세요`,\n )\n }\n return result\n}\n\n// ─── 페이지 단위 래스터 (#75 Task 5) ────────────────────\n\nexport type RasterFormat = \"png\" | \"jpeg\"\n\nexport interface PageRasterOptions {\n format?: RasterFormat\n /** 출력 최대 폭 px (기본 1400) */\n maxWidthPx?: number\n /** 출력 최대 높이 px (기본 4000 — 페이지 1장) */\n maxHeightPx?: number\n /** JPEG 품질 (기본 85) */\n quality?: number\n}\n\nexport interface PageRasterResult {\n data: Buffer\n mimeType: \"image/png\" | \"image/jpeg\"\n widthPx: number\n heightPx: number\n /** pt → px 실배율 — crop 픽셀 좌표는 이 값으로만 환산한다 (DPI 재추정 금지) */\n scale: number\n}\n\n/**\n * 페이지 standalone SVG 1장 → PNG/JPEG. 두 포맷은 같은 SVG·같은 배율을 쓴다.\n * 픽셀 크기·scale 은 산출 이미지에서 읽은 실제값(crop 은 이 scale 로만 환산).\n */\nexport async function rasterizePageSvg(svg: string, widthPt: number, heightPt: number, options: PageRasterOptions = {}): Promise<PageRasterResult> {\n const sharp = await loadSharp()\n if (!(widthPt > 0) || !(heightPt > 0)) throw new KordocError(`잘못된 페이지 크기: ${widthPt}x${heightPt}pt`)\n const maxW = options.maxWidthPx ?? 1400\n const maxH = options.maxHeightPx ?? 4000\n const scale = Math.min(maxW / widthPt, maxH / heightPt, 2)\n const format = options.format ?? \"png\"\n const exponent = await probeDensityExponent(sharp)\n const pipeline = sharp(Buffer.from(svg), { density: densityFor(scale, exponent), limitInputPixels: 268402689 })\n const data = format === \"jpeg\" ? await pipeline.jpeg({ quality: options.quality ?? 85 }).toBuffer() : await pipeline.png().toBuffer()\n const { w, h } = await actualSize(sharp, data, Math.round(widthPt * scale), Math.round(heightPt * scale))\n return { data, mimeType: format === \"jpeg\" ? \"image/jpeg\" : \"image/png\", widthPx: w, heightPx: h, scale: w / widthPt }\n}\n"],"mappings":";;;;;;;AAoCA,IAAI,kBAAiC;AACrC,IAAM,YAAY;AAGlB,SAAS,SAAS,KAAqB;AACrC,SAAO,IAAI,UAAU,MAAM,IAAI,CAAC,MAAM,OAAQ,IAAI,CAAC,MAAM,KAAO,IAAI,aAAa,EAAE,IAAI;AACzF;AAEA,eAAe,qBAAqB,OAAmC;AACrE,MAAI,oBAAoB,KAAM,QAAO;AACrC,MAAI;AACF,UAAM,MAAM,MAAM,MAAM,OAAO,KAAK,SAAS,GAAG,EAAE,SAAS,IAAI,CAAC,EAAE,IAAI,EAAE,SAAS;AACjF,UAAM,IAAI,SAAS,GAAG,IAAI;AAC1B,UAAM,IAAI,IAAI,IAAI,KAAK,IAAI,CAAC,IAAI,KAAK,IAAI,CAAC,IAAI;AAC9C,sBAAkB,OAAO,SAAS,CAAC,KAAK,IAAI,MAAM,KAAK,MAAM,IAAI,CAAC,IAAI,IAAI;AAAA,EAC5E,QAAQ;AACN,sBAAkB;AAAA,EACpB;AACA,SAAO;AACT;AAGA,SAAS,WAAW,OAAe,UAA0B;AAC3D,SAAO,KAAK,KAAK,IAAI,OAAO,IAAI,QAAQ;AAC1C;AAGA,eAAe,WAAW,OAAkB,MAAc,WAAmB,WAAsD;AACjI,QAAM,IAAI,SAAS,IAAI;AACvB,MAAI,IAAI,KAAK,KAAK,UAAU,GAAI,QAAO,EAAE,GAAG,GAAG,KAAK,aAAa,EAAE,EAAE;AACrE,MAAI;AACF,UAAM,IAAI,MAAM,MAAM,IAAI,EAAE,SAAS;AACrC,QAAI,EAAE,SAAS,EAAE,OAAQ,QAAO,EAAE,GAAG,EAAE,OAAO,GAAG,EAAE,OAAO;AAAA,EAC5D,QAAQ;AAAA,EAAW;AACnB,SAAO,EAAE,GAAG,WAAW,GAAG,UAAU;AACtC;AAEA,eAAsB,YAAgC;AACpD,MAAI;AACF,UAAM,MAAW,MAAM,OAAO,OAAO;AACrC,WAAO,IAAI,WAAW;AAAA,EACxB,QAAQ;AACN,UAAM,IAAI;AAAA,MACR,mPACE;AAAA,IACJ;AAAA,EACF;AACF;AAGA,eAAsB,aACpB,KACA,SACA,UACA,SAC0B;AAC1B,QAAM,QAAQ,MAAM,UAAU;AAC9B,MAAI,EAAE,UAAU,MAAM,EAAE,WAAW,GAAI,OAAM,IAAI,YAAY,wCAAe,OAAO,IAAI,QAAQ,IAAI;AACnG,QAAM,OAAO,SAAS,cAAc;AACpC,QAAM,OAAO,SAAS,eAAe;AACrC,QAAM,WAAW,SAAS,YAAY,IAAI,OAAO;AAEjD,QAAM,WAAW,MAAM,qBAAqB,KAAK;AACjD,QAAM,SAAS,OAAOA,WAA4C;AAEhE,UAAM,MAAM,MAAM,MAAM,OAAO,KAAK,GAAG,GAAG,EAAE,SAAS,WAAWA,QAAO,QAAQ,GAAG,kBAAkB,UAAU,CAAC,EAC5G,IAAI,EACJ,SAAS;AACZ,UAAM,EAAE,GAAG,EAAE,IAAI,MAAM,WAAW,OAAO,KAAK,KAAK,MAAM,UAAUA,MAAK,GAAG,KAAK,MAAM,WAAWA,MAAK,CAAC;AACvG,WAAO,EAAE,KAAK,SAAS,GAAG,UAAU,GAAG,OAAO,IAAI,QAAQ;AAAA,EAC5D;AAEA,QAAM,QAAQ,KAAK,IAAI,OAAO,SAAS,OAAO,UAAU,CAAC;AACzD,MAAI,SAAS,MAAM,OAAO,KAAK;AAC/B,MAAI,OAAO,IAAI,SAAS,YAAY,QAAQ,MAAM;AAChD,aAAS,MAAM,OAAO,QAAQ,CAAC;AAAA,EACjC;AACA,MAAI,OAAO,IAAI,SAAS,UAAU;AAChC,UAAM,IAAI;AAAA,MACR,2BAAY,WAAW,OAAO,MAAM,QAAQ,CAAC,CAAC,8CAAgB,OAAO,IAAI,SAAS,OAAO,MAAM,QAAQ,CAAC,CAAC;AAAA,IAC3G;AAAA,EACF;AACA,SAAO;AACT;AA6BA,eAAsB,iBAAiB,KAAa,SAAiB,UAAkB,UAA6B,CAAC,GAA8B;AACjJ,QAAM,QAAQ,MAAM,UAAU;AAC9B,MAAI,EAAE,UAAU,MAAM,EAAE,WAAW,GAAI,OAAM,IAAI,YAAY,uDAAe,OAAO,IAAI,QAAQ,IAAI;AACnG,QAAM,OAAO,QAAQ,cAAc;AACnC,QAAM,OAAO,QAAQ,eAAe;AACpC,QAAM,QAAQ,KAAK,IAAI,OAAO,SAAS,OAAO,UAAU,CAAC;AACzD,QAAM,SAAS,QAAQ,UAAU;AACjC,QAAM,WAAW,MAAM,qBAAqB,KAAK;AACjD,QAAM,WAAW,MAAM,OAAO,KAAK,GAAG,GAAG,EAAE,SAAS,WAAW,OAAO,QAAQ,GAAG,kBAAkB,UAAU,CAAC;AAC9G,QAAM,OAAO,WAAW,SAAS,MAAM,SAAS,KAAK,EAAE,SAAS,QAAQ,WAAW,GAAG,CAAC,EAAE,SAAS,IAAI,MAAM,SAAS,IAAI,EAAE,SAAS;AACpI,QAAM,EAAE,GAAG,EAAE,IAAI,MAAM,WAAW,OAAO,MAAM,KAAK,MAAM,UAAU,KAAK,GAAG,KAAK,MAAM,WAAW,KAAK,CAAC;AACxG,SAAO,EAAE,MAAM,UAAU,WAAW,SAAS,eAAe,aAAa,SAAS,GAAG,UAAU,GAAG,OAAO,IAAI,QAAQ;AACvH;","names":["scale"]}
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
"use strict";Object.defineProperty(exports, "__esModule", {value: true}); function _optionalChain(ops) { let lastAccessLHS = undefined; let value = ops[0]; let i = 1; while (i < ops.length) { const op = ops[i]; const fn = ops[i + 1]; i += 2; if ((op === 'optionalAccess' || op === 'optionalCall') && value == null) { return undefined; } if (op === 'access' || op === 'optionalAccess') { lastAccessLHS = value; value = fn(value); } else if (op === 'call' || op === 'optionalCall') { value = fn((...args) => value.call(lastAccessLHS, ...args)); lastAccessLHS = undefined; } } return value; }
|
|
2
2
|
|
|
3
|
-
var
|
|
3
|
+
var _chunkYRRBA4HWcjs = require('./chunk-YRRBA4HW.cjs');
|
|
4
4
|
|
|
5
5
|
// src/pdf/formula/models.ts
|
|
6
6
|
var _crypto = require('crypto');
|
|
@@ -23,7 +23,7 @@ function isOfflineMode() {
|
|
|
23
23
|
}
|
|
24
24
|
function assertNetworkAllowed(what, hint) {
|
|
25
25
|
if (!isOfflineMode()) return;
|
|
26
|
-
throw new (0,
|
|
26
|
+
throw new (0, _chunkYRRBA4HWcjs.KordocError)(
|
|
27
27
|
`\uD3D0\uC1C4\uB9DD \uBAA8\uB4DC(KORDOC_OFFLINE)\uC5D0\uC11C \uCC28\uB2E8\uB428: ${what}` + (hint ? ` \u2014 ${hint}` : "")
|
|
28
28
|
);
|
|
29
29
|
}
|
|
@@ -266,4 +266,4 @@ async function sha256OfFile(p) {
|
|
|
266
266
|
|
|
267
267
|
|
|
268
268
|
exports.MFD_MODEL = MFD_MODEL; exports.MFR_ENCODER_MODEL = MFR_ENCODER_MODEL; exports.MFR_DECODER_MODEL = MFR_DECODER_MODEL; exports.MFR_TOKENIZER = MFR_TOKENIZER; exports.ALL_FORMULA_MODELS = ALL_FORMULA_MODELS; exports.getModelsDir = getModelsDir; exports.getFormulaModelsDir = getFormulaModelsDir; exports.getFormulaModelStatus = getFormulaModelStatus; exports.ensureModelsIn = ensureModelsIn; exports.ensureFormulaModels = ensureFormulaModels; exports.ensureSingleModel = ensureSingleModel;
|
|
269
|
-
//# sourceMappingURL=chunk-
|
|
269
|
+
//# sourceMappingURL=chunk-2BZFKIRB.cjs.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"sources":["/Users/mong-e/workspace/kordoc/dist/chunk-2BZFKIRB.cjs","../src/pdf/formula/models.ts","../src/shared/offline.ts"],"names":[],"mappings":"AAAA;AACE;AACF,wDAA6B;AAC7B;AACA;ACMA,gCAA2B;AAC3B,wBAAiC;AACjC,uCAA4C;AAC5C;AACA,wBAAwB;AACxB,4BAA8B;AAC9B,4CAAyB;AACzB,gCAAyB;ADJzB;AACA;AEKA,IAAM,OAAA,kBAAS,IAAI,GAAA,CAAI,CAAC,GAAA,EAAK,MAAA,EAAQ,KAAA,EAAO,IAAI,CAAC,CAAA;AAEjD,SAAS,OAAA,CAAQ,IAAA,EAAuB;AACtC,EAAA,MAAM,EAAA,EAAI,OAAA,CAAQ,GAAA,CAAI,IAAI,CAAA;AAC1B,EAAA,OAAO,CAAC,CAAC,EAAA,GAAK,MAAA,CAAO,GAAA,CAAI,CAAA,CAAE,IAAA,CAAK,CAAA,CAAE,WAAA,CAAY,CAAC,CAAA;AACjD;AAGO,SAAS,aAAA,CAAA,EAAyB;AACvC,EAAA,OAAO,OAAA,CAAQ,gBAAgB,CAAA;AACjC;AAOO,SAAS,oBAAA,CAAqB,IAAA,EAAc,IAAA,EAAqB;AACtE,EAAA,GAAA,CAAI,CAAC,aAAA,CAAc,CAAA,EAAG,MAAA;AACtB,EAAA,MAAM,IAAI,kCAAA;AAAA,IACR,CAAA,gFAAA,EAAiC,IAAI,CAAA,EAAA;AACvC,EAAA;AACF;AFZ4F;AACA;ACDxD;AAC5B,EAAA;AACI,EAAA;AACL,EAAA;AACG,EAAA;AACA,EAAA;AACV;AAE4C;AACpC,EAAA;AACI,EAAA;AACL,EAAA;AACG,EAAA;AACA,EAAA;AACV;AAE4C;AACpC,EAAA;AACI,EAAA;AACL,EAAA;AACG,EAAA;AACA,EAAA;AACV;AAEwC;AAChC,EAAA;AACI,EAAA;AACL,EAAA;AACG,EAAA;AACA,EAAA;AACV;AAE4D;AAC1D,EAAA;AACA,EAAA;AACA,EAAA;AACA,EAAA;AACF;AAOqD;AACtB,EAAA;AACI,EAAA;AACH,IAAA;AAC9B,EAAA;AAC2D,EAAA;AAC7D;AAE8C;AACd,EAAA;AAChC;AAa6G;AAC5E,EAAA;AACL,EAAA;AACiB,IAAA;AAC5B,IAAA;AACT,IAAA;AAC4B,MAAA;AACE,MAAA;AAC1B,IAAA;AACG,MAAA;AACX,IAAA;AACa,IAAA;AACoD,MAAA;AAC/D,MAAA;AACF,IAAA;AACI,IAAA;AACyC,MAAA;AACf,MAAA;AACmC,QAAA;AACxD,MAAA;AACO,QAAA;AACV,UAAA;AACA,UAAA;AACQ,UAAA;AACE,UAAA;AAC4D,UAAA;AACvE,QAAA;AACH,MAAA;AACU,IAAA;AACE,MAAA;AACV,QAAA;AACA,QAAA;AACQ,QAAA;AACE,QAAA;AACgD,QAAA;AAC3D,MAAA;AACH,IAAA;AACF,EAAA;AACO,EAAA;AACT;AAGsE;AACH,EAAA;AACnE;AAoBgI;AAC1F,EAAA;AAEV,EAAA;AACiB,IAAA;AAEU,IAAA;AACpC,sBAAA;AACX,QAAA;AACY,QAAA;AACL,QAAA;AACA,QAAA;AACE,QAAA;AACV,MAAA;AACD,MAAA;AACF,IAAA;AAGI,IAAA;AACoB,MAAA;AAChB,IAAA;AAER,IAAA;AAEgD,IAAA;AAClD,EAAA;AACF;AAGuF;AACV,EAAA;AAC7E;AAGsG;AACpE,EAAA;AACI,EAAA;AACK,EAAA;AACU,EAAA;AACe,oBAAA;AAChE,IAAA;AACF,EAAA;AACI,EAAA;AACoB,IAAA;AAChB,EAAA;AAER,EAAA;AACgD,EAAA;AAClD;AAE4F;AACtF,EAAA;AAC4B,IAAA;AACU,IAAA;AAClC,EAAA;AACC,IAAA;AACT,EAAA;AACI,EAAA;AACyC,IAAA;AACzB,IAAA;AACZ,EAAA;AACC,IAAA;AACT,EAAA;AACF;AAMiB;AAEf,EAAA;AACc,IAAA;AACZ,IAAA;AACF,EAAA;AAG6B,EAAA;AACsB,EAAA;AAEhB,EAAA;AACxB,IAAA;AAAA;AAEO,MAAA;AAChB,IAAA;AACD,EAAA;AAC2B,EAAA;AAChB,IAAA;AAC0C,MAAA;AACpD,IAAA;AACF,EAAA;AAEmD,EAAA;AACQ,EAAA;AAC1C,EAAA;AAEoB,EAAA;AACjC,EAAA;AACyF,IAAA;AACzC,IAAA;AAC5B,MAAA;AACP,sBAAA;AACX,QAAA;AACA,QAAA;AACA,QAAA;AACO,QAAA;AACR,MAAA;AACF,IAAA;AACwB,IAAA;AACf,EAAA;AACN,IAAA;AACmB,MAAA;AACf,IAAA;AAER,IAAA;AAC+D,IAAA;AACjE,EAAA;AAEa,kBAAA;AACX,IAAA;AACA,IAAA;AACA,IAAA;AACO,IAAA;AACR,EAAA;AAGG,EAAA;AACA,EAAA;AACkC,IAAA;AAC1B,EAAA;AACN,IAAA;AACmB,MAAA;AACf,IAAA;AAER,IAAA;AACiE,IAAA;AACnE,EAAA;AAE4B,EAAA;AACtB,IAAA;AACmB,MAAA;AACf,IAAA;AAER,IAAA;AACU,IAAA;AAC4D,MAAA;AACtE,IAAA;AACF,EAAA;AAEgC,EAAA;AACnB,kBAAA;AACX,IAAA;AACA,IAAA;AACA,IAAA;AACO,IAAA;AACR,EAAA;AACH;AAEwD;AACzB,EAAA;AACI,EAAA;AACY,EAAA;AACZ,IAAA;AACf,MAAA;AAEhB,IAAA;AACD,EAAA;AACoB,EAAA;AACvB;AD1E4F;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA;AACA","file":"/Users/mong-e/workspace/kordoc/dist/chunk-2BZFKIRB.cjs","sourcesContent":[null,"/**\n * 수식 OCR 모델 자동 다운로드 + SHA-256 검증\n *\n * 캐시 위치: `~/.cache/kordoc/models/pix2text/`\n * 총 ~155MB (MFD 44MB + MFR encoder 87MB + MFR decoder 30MB + tokenizer 40KB)\n *\n * HF URL + SHA 는 Docufinder archive/rust-formula-ocr-wip 에서 검증 완료됨.\n * 모델 버전 갱신 시 URL + SHA 를 함께 갱신해야 한다.\n */\n\nimport { createHash } from \"crypto\"\nimport { createReadStream } from \"fs\"\nimport { mkdir, stat, unlink, rename } from \"fs/promises\"\nimport { createWriteStream } from \"fs\"\nimport { homedir } from \"os\"\nimport { join, dirname } from \"path\"\nimport { pipeline } from \"stream/promises\"\nimport { Readable } from \"stream\"\nimport { assertNetworkAllowed } from \"../../shared/offline.js\"\n\nexport interface ModelSpec {\n name: string\n filename: string\n url: string\n sha256: string\n sizeMb: number\n}\n\n/** 모델 스펙 (SHA-256 은 실제 다운로드로 검증됨 — 변경 금지) */\nexport const MFD_MODEL: ModelSpec = {\n name: \"Pix2Text MFD\",\n filename: \"mfd.onnx\",\n url: \"https://huggingface.co/breezedeus/pix2text-mfd/resolve/main/mfd-v20240618.onnx\",\n sha256: \"51a8854743b17ae654729af8db82a630c1ccfa06debf4856c8b28055f87d02c1\",\n sizeMb: 42,\n}\n\nexport const MFR_ENCODER_MODEL: ModelSpec = {\n name: \"Pix2Text MFR encoder\",\n filename: \"encoder_model.onnx\",\n url: \"https://huggingface.co/breezedeus/pix2text-mfr/resolve/main/encoder_model.onnx\",\n sha256: \"bd8d5c322792e9ec45793af5569e9748f82a3d728a9e00213dbfc56c1486f37d\",\n sizeMb: 87,\n}\n\nexport const MFR_DECODER_MODEL: ModelSpec = {\n name: \"Pix2Text MFR decoder\",\n filename: \"decoder_model.onnx\",\n url: \"https://huggingface.co/breezedeus/pix2text-mfr/resolve/main/decoder_model.onnx\",\n sha256: \"fd0f92d7a012f3dae41e1ac79421aea0ea888b5a66cb3f9a004e424f82f3daed\",\n sizeMb: 30,\n}\n\nexport const MFR_TOKENIZER: ModelSpec = {\n name: \"Pix2Text MFR tokenizer\",\n filename: \"tokenizer.json\",\n url: \"https://huggingface.co/breezedeus/pix2text-mfr/resolve/main/tokenizer.json\",\n sha256: \"3e2ab757277d22639bec28c9d7972e352d3d1dba223051fa674002dc5ab64df3\",\n sizeMb: 1,\n}\n\nexport const ALL_FORMULA_MODELS: ReadonlyArray<ModelSpec> = [\n MFD_MODEL,\n MFR_ENCODER_MODEL,\n MFR_DECODER_MODEL,\n MFR_TOKENIZER,\n]\n\n/**\n * 모델 캐시 하위 디렉토리 경로. 환경 변수 `KORDOC_MODEL_CACHE` 가 설정되면\n * 해당 경로를 우선 사용, 없으면 `~/.cache/kordoc/models/<subdir>/`.\n * (수식 OCR 외 텍스트 OCR 등 다른 모델군도 공용 — subdir 로 격리)\n */\nexport function getModelsDir(subdir: string): string {\n const override = process.env.KORDOC_MODEL_CACHE\n if (override && override.trim()) {\n return join(override, subdir)\n }\n return join(homedir(), \".cache\", \"kordoc\", \"models\", subdir)\n}\n\nexport function getFormulaModelsDir(): string {\n return getModelsDir(\"pix2text\")\n}\n\nexport interface ModelStatus {\n spec: ModelSpec\n localPath: string\n exists: boolean\n /** SHA 검증까지 끝났는가 (exists && valid) */\n verified: boolean\n /** 검증 실패 시 이유 */\n invalidReason?: string\n}\n\n/** 지정 모델군의 존재/유효성 상태 반환 (다운로드 없이 확인만) */\nexport async function getModelStatusIn(dir: string, specs: ReadonlyArray<ModelSpec>): Promise<ModelStatus[]> {\n const result: ModelStatus[] = []\n for (const spec of specs) {\n const localPath = join(dir, spec.filename)\n let exists = false\n try {\n const s = await stat(localPath)\n exists = s.isFile() && s.size > 0\n } catch {\n exists = false\n }\n if (!exists) {\n result.push({ spec, localPath, exists: false, verified: false })\n continue\n }\n try {\n const actual = await sha256OfFile(localPath)\n if (actual === spec.sha256) {\n result.push({ spec, localPath, exists: true, verified: true })\n } else {\n result.push({\n spec,\n localPath,\n exists: true,\n verified: false,\n invalidReason: `SHA256 mismatch: expected ${spec.sha256}, got ${actual}`,\n })\n }\n } catch (e) {\n result.push({\n spec,\n localPath,\n exists: true,\n verified: false,\n invalidReason: `SHA compute failed: ${(e as Error).message}`,\n })\n }\n }\n return result\n}\n\n/** 모든 수식 모델의 존재/유효성 상태 반환 (다운로드 없이 확인만) */\nexport async function getFormulaModelStatus(): Promise<ModelStatus[]> {\n return getModelStatusIn(getFormulaModelsDir(), ALL_FORMULA_MODELS)\n}\n\nexport interface DownloadProgress {\n spec: ModelSpec\n /** 현재까지 다운로드한 바이트 */\n downloaded: number\n /** 전체 바이트 (알 수 있으면) */\n total: number | null\n /** \"download\" | \"verify\" | \"done\" | \"skip\" */\n phase: \"download\" | \"verify\" | \"done\" | \"skip\" | \"error\"\n message?: string\n}\n\nexport type ProgressHandler = (p: DownloadProgress) => void\n\n/**\n * 지정 모델군을 다운로드/검증한다.\n * - 이미 있고 SHA 일치 → skip\n * - 없음 → 다운로드 후 SHA 검증. 실패 시 파일 삭제.\n */\nexport async function ensureModelsIn(dir: string, specs: ReadonlyArray<ModelSpec>, onProgress?: ProgressHandler): Promise<void> {\n await mkdir(dir, { recursive: true })\n\n for (const spec of specs) {\n const localPath = join(dir, spec.filename)\n\n if (await isExistingValid(localPath, spec.sha256)) {\n onProgress?.({\n spec,\n downloaded: 0,\n total: null,\n phase: \"skip\",\n message: \"이미 존재 + SHA 일치\",\n })\n continue\n }\n\n // 기존 파일 있지만 SHA 불일치 → 삭제\n try {\n await unlink(localPath)\n } catch {\n // 없을 수 있음\n }\n\n await downloadToFile(spec, localPath, onProgress)\n }\n}\n\n/** 필요한 모든 수식 모델을 다운로드/검증한다. */\nexport async function ensureFormulaModels(onProgress?: ProgressHandler): Promise<void> {\n return ensureModelsIn(getFormulaModelsDir(), ALL_FORMULA_MODELS, onProgress)\n}\n\n/** 단일 모델만 확인/다운로드 (진단 UI 에서 개별 상태 새로고침용) */\nexport async function ensureSingleModel(spec: ModelSpec, onProgress?: ProgressHandler): Promise<void> {\n const dir = getFormulaModelsDir()\n await mkdir(dir, { recursive: true })\n const localPath = join(dir, spec.filename)\n if (await isExistingValid(localPath, spec.sha256)) {\n onProgress?.({ spec, downloaded: 0, total: null, phase: \"skip\" })\n return\n }\n try {\n await unlink(localPath)\n } catch {\n // 재다운로드 전 손상/잔존 파일 정리 best-effort — 애초에 없으면 실패해도 무방\n }\n await downloadToFile(spec, localPath, onProgress)\n}\n\nasync function isExistingValid(localPath: string, sha256Expected: string): Promise<boolean> {\n try {\n const s = await stat(localPath)\n if (!s.isFile() || s.size === 0) return false\n } catch {\n return false\n }\n try {\n const actual = await sha256OfFile(localPath)\n return actual === sha256Expected\n } catch {\n return false\n }\n}\n\nasync function downloadToFile(\n spec: ModelSpec,\n localPath: string,\n onProgress?: ProgressHandler,\n): Promise<void> {\n // 폐쇄망 모드에서는 요청 자체를 보내지 않는다 (사이드로드 안내로 대체)\n assertNetworkAllowed(\n `${spec.name} 모델 다운로드`,\n \"온라인 PC에서 `kordoc models --export <디렉토리>` 로 내보낸 뒤 이 PC에서 `kordoc models --import <디렉토리>` 하거나, KORDOC_MODEL_CACHE 로 모델 캐시 경로를 지정하세요\",\n )\n\n // 먼저 .part 로 받고 검증 후 rename — 중단된 다운로드가 \"정상 파일\"로 오인되는 걸 방지\n const partPath = `${localPath}.part`\n await mkdir(dirname(localPath), { recursive: true })\n\n const resp = await fetch(spec.url, {\n headers: {\n // HF CDN 은 UA 없으면 가끔 403 을 뱉는다\n \"User-Agent\": \"kordoc-formula-ocr/1.0 (+https://github.com/chrisryugj/kordoc)\",\n },\n })\n if (!resp.ok || !resp.body) {\n throw new Error(\n `${spec.name} 다운로드 실패: HTTP ${resp.status} ${resp.statusText} (${spec.url})`,\n )\n }\n\n const lenHeader = resp.headers.get(\"content-length\")\n const total = lenHeader ? Number.parseInt(lenHeader, 10) : null\n let downloaded = 0\n\n const ws = createWriteStream(partPath)\n try {\n const reader = Readable.fromWeb(resp.body as unknown as import(\"stream/web\").ReadableStream)\n reader.on(\"data\", (chunk: Buffer | Uint8Array) => {\n downloaded += chunk.length\n onProgress?.({\n spec,\n downloaded,\n total,\n phase: \"download\",\n })\n })\n await pipeline(reader, ws)\n } catch (e) {\n try {\n await unlink(partPath)\n } catch {\n // .part 임시파일 정리 best-effort — 본 에러(스트리밍 실패)를 가리지 않도록 무시\n }\n throw new Error(`${spec.name} 스트리밍 실패: ${(e as Error).message}`)\n }\n\n onProgress?.({\n spec,\n downloaded,\n total,\n phase: \"verify\",\n })\n\n // SHA 검증\n let actual: string\n try {\n actual = await sha256OfFile(partPath)\n } catch (e) {\n try {\n await unlink(partPath)\n } catch {\n // .part 임시파일 정리 best-effort — 본 에러(SHA 계산 실패)를 가리지 않도록 무시\n }\n throw new Error(`${spec.name} SHA 계산 실패: ${(e as Error).message}`)\n }\n\n if (actual !== spec.sha256) {\n try {\n await unlink(partPath)\n } catch {\n // 오염된 .part 정리 best-effort — 본 에러(SHA mismatch)를 가리지 않도록 무시\n }\n throw new Error(\n `${spec.name} SHA256 mismatch: expected ${spec.sha256}, got ${actual} — 모델 URL 이 오염되었거나 전송 중 손상되었습니다.`,\n )\n }\n\n await rename(partPath, localPath)\n onProgress?.({\n spec,\n downloaded,\n total,\n phase: \"done\",\n })\n}\n\nasync function sha256OfFile(p: string): Promise<string> {\n const h = createHash(\"sha256\")\n const stream = createReadStream(p)\n await pipeline(stream, async function* (src) {\n for await (const chunk of src) {\n h.update(chunk)\n // yield 안 함 (소비만)\n }\n })\n return h.digest(\"hex\")\n}\n","/**\n * 폐쇄망(내부망) 배포용 아웃바운드 통신 킬스위치 + 파일 접근 루트 제한.\n *\n * kordoc 자체는 문서 파서라 상시 외부 통신이 없지만, 두 경로만 나간다:\n * 1. OCR/수식 모델 최초 1회 다운로드 (huggingface.co)\n * 2. `kordoc watch --webhook` 사용자가 명시한 URL\n *\n * `KORDOC_OFFLINE=1` 이면 위 경로를 **시도 전에** 차단한다. 보안성 검토에서\n * \"외부 통신 없음\"을 런타임으로 증명하기 위한 단일 관문 — 새 아웃바운드 경로를\n * 추가하면 반드시 `assertNetworkAllowed()` 를 먼저 통과시킬 것.\n *\n * `KORDOC_ROOT` 는 파일 접근을 특정 디렉토리 하위로 제한한다(미설정 시 제한 없음 —\n * 기존 동작 보존). 내부망 배포에서는 두 변수를 함께 고정하는 것을 권장.\n */\n\nimport { realpathSync } from \"fs\"\nimport { isAbsolute, relative, resolve, sep } from \"path\"\nimport { KordocError } from \"../utils.js\"\n\nconst TRUTHY = new Set([\"1\", \"true\", \"yes\", \"on\"])\n\nfunction envFlag(name: string): boolean {\n const v = process.env[name]\n return !!v && TRUTHY.has(v.trim().toLowerCase())\n}\n\n/** 폐쇄망 모드 여부 (`KORDOC_OFFLINE=1|true|yes|on`) */\nexport function isOfflineMode(): boolean {\n return envFlag(\"KORDOC_OFFLINE\")\n}\n\n/**\n * 아웃바운드 통신 관문. 폐쇄망 모드면 요청을 보내기 전에 예외로 끊는다.\n * @param what 차단 대상 설명 (사용자에게 노출됨 — 호스트/URL 은 넣지 말 것)\n * @param hint 대안 안내 (오프라인 사이드로드 방법 등)\n */\nexport function assertNetworkAllowed(what: string, hint?: string): void {\n if (!isOfflineMode()) return\n throw new KordocError(\n `폐쇄망 모드(KORDOC_OFFLINE)에서 차단됨: ${what}` + (hint ? ` — ${hint}` : \"\"),\n )\n}\n\n/** 파일 접근 루트 (`KORDOC_ROOT`). 미설정이면 null = 제한 없음 */\nexport function getAccessRoot(): string | null {\n const raw = process.env.KORDOC_ROOT\n if (!raw || !raw.trim()) return null\n const resolved = resolve(raw.trim())\n try {\n return realpathSync(resolved)\n } catch {\n // 루트가 실재하지 않으면 모든 접근이 막히는 게 안전한 기본값\n return resolved\n }\n}\n\n/** 경로가 루트 하위인지 (루트 자신 포함). 경계 문자열 매칭이 아닌 세그먼트 단위 비교 */\nexport function isWithinRoot(realPath: string, root: string): boolean {\n if (realPath === root) return true\n const rel = relative(root, realPath)\n return rel !== \"\" && !rel.startsWith(\"..\") && !isAbsolute(rel) && !rel.startsWith(`..${sep}`)\n}\n\n/**\n * `KORDOC_ROOT` 가 설정된 경우 경로가 루트 하위인지 검증한다.\n * 인자는 이미 realpath 로 해석된 절대경로여야 한다 (심볼릭 링크 탈출 방지).\n */\nexport function assertWithinRoot(realPath: string): void {\n const root = getAccessRoot()\n if (!root) return\n if (!isWithinRoot(realPath, root)) {\n throw new KordocError(`허용된 작업 디렉토리(KORDOC_ROOT) 밖의 경로입니다: ${realPath}`)\n }\n}\n"]}
|
|
@@ -2,7 +2,7 @@
|
|
|
2
2
|
|
|
3
3
|
// src/utils.ts
|
|
4
4
|
import { format } from "util";
|
|
5
|
-
var VERSION = true ? "4.
|
|
5
|
+
var VERSION = true ? "4.16.1" : "0.0.0-dev";
|
|
6
6
|
function toArrayBuffer(buf) {
|
|
7
7
|
if (buf.byteOffset === 0 && buf.byteLength === buf.buffer.byteLength) {
|
|
8
8
|
return buf.buffer;
|
|
@@ -122,6 +122,7 @@ function safeMax(arr) {
|
|
|
122
122
|
for (let i = 0; i < arr.length; i++) if (arr[i] > max) max = arr[i];
|
|
123
123
|
return max;
|
|
124
124
|
}
|
|
125
|
+
var OPTIONAL_DEP_INSTALL_HINT = " \uB124\uD2B8\uC6CC\uD06C\uAC00 \uC81C\uD55C\uB41C linux/x64 \uC5D0\uC11C npx \uB85C \uC124\uCE58\uD588\uB2E4\uBA74 onnxruntime-node \uC758 CUDA \uB2E4\uC6B4\uB85C\uB4DC \uC2E4\uD328\uB85C \uD568\uAED8 \uBE60\uC84C\uC744 \uC218 \uC788\uC2B5\uB2C8\uB2E4 \u2014 `ONNXRUNTIME_NODE_INSTALL=skip npx -y kordoc@^4 \u2026` \uB85C \uB2E4\uC2DC \uC124\uCE58\uD558\uC138\uC694.";
|
|
125
126
|
function classifyError(err) {
|
|
126
127
|
if (!(err instanceof Error)) return "PARSE_ERROR";
|
|
127
128
|
const msg = err.message;
|
|
@@ -155,6 +156,7 @@ export {
|
|
|
155
156
|
routeConsoleToStderr,
|
|
156
157
|
safeMin,
|
|
157
158
|
safeMax,
|
|
159
|
+
OPTIONAL_DEP_INSTALL_HINT,
|
|
158
160
|
classifyError
|
|
159
161
|
};
|
|
160
|
-
//# sourceMappingURL=chunk-
|
|
162
|
+
//# sourceMappingURL=chunk-3VLVNKJV.js.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"sources":["../src/utils.ts"],"sourcesContent":["/** kordoc 공용 유틸리티 */\n\nimport { format } from \"util\"\n\n/** 빌드 타임에 tsup define으로 주입되는 버전 */\ndeclare const __KORDOC_VERSION__: string\nexport const VERSION: string = typeof __KORDOC_VERSION__ !== \"undefined\" ? __KORDOC_VERSION__ : \"0.0.0-dev\"\n\n/**\n * Node.js Buffer → ArrayBuffer 변환\n * pool Buffer의 공유 ArrayBuffer 문제를 안전하게 처리.\n * offset=0이고 전체 ArrayBuffer를 차지하면 복사 없이 직접 반환.\n */\nexport function toArrayBuffer(buf: Buffer): ArrayBuffer {\n if (buf.byteOffset === 0 && buf.byteLength === buf.buffer.byteLength) {\n return buf.buffer as ArrayBuffer\n }\n return buf.buffer.slice(buf.byteOffset, buf.byteOffset + buf.byteLength) as ArrayBuffer\n}\n\n/**\n * kordoc 내부 에러 클래스 — 사용자에게 노출해도 안전한 메시지만 포함.\n * MCP 에러 정제에서 instanceof로 판별하여 allowlist 패턴 매칭 없이 안전하게 통과.\n */\nexport class KordocError extends Error {\n constructor(message: string) {\n super(message)\n this.name = \"KordocError\"\n }\n}\n\n/**\n * ENOENT 판별 — 입력 경로가 존재하지 않음.\n * 래핑된 KordocError 를 뚫고 올라온 경우도 있으므로 메시지까지 함께 본다.\n */\nfunction isFileNotFound(err: unknown): boolean {\n if (typeof err !== \"object\" || err === null) return false\n if ((err as NodeJS.ErrnoException).code === \"ENOENT\") return true\n const msg = (err as Error).message\n return typeof msg === \"string\" && /ENOENT|no such file or directory/i.test(msg)\n}\n\n/**\n * 에러 메시지 정제 — KordocError는 그대로, ENOENT 만 구분 문구, 나머지는 일반 메시지로 대체.\n * 파일시스템 경로, 스택 트레이스 등 내부 정보 노출 방지.\n */\nexport function sanitizeError(err: unknown): string {\n if (err instanceof KordocError) return err.message\n // 없는 파일은 문서 문제가 아니라 고칠 수 있는 입력 문제다 — 사유를 숨기지 않는다.\n // 경로는 노출하지 않는다(어느 파일인지는 stderr·실패 JSON 의 basename 이 알린다).\n // 문구는 MCP describeError 의 ENOENT 힌트와 같은 표현을 쓴다.\n if (isFileNotFound(err)) return \"파일 또는 디렉토리를 찾을 수 없습니다\"\n return \"문서 처리 중 오류가 발생했습니다\"\n}\n\n/**\n * ZIP 엔트리 경로의 경로 순회 여부 판별.\n * 백슬래시 정규화, .., 절대경로, Windows 드라이브 문자 모두 차단.\n */\nexport function isPathTraversal(name: string): boolean {\n if (name.includes(\"\\x00\")) return true\n const normalized = name.replace(/\\\\/g, \"/\")\n const segments = normalized.split(\"/\")\n return segments.some(s => s === \"..\") || normalized.startsWith(\"/\") || /^[A-Za-z]:/.test(normalized)\n}\n\n// ─── HWPX 섹션 href 해석 (parser/roundtrip 공용) ────────────────────\n\n/**\n * manifest href를 본문 섹션 경로(`Contents/sectionN.xml`)로 정규화.\n * 비본문 XML(header/script/settings)·path traversal·드라이브 경로는 null로 거른다.\n * 백슬래시 구분자를 슬래시로 통일해 parser와 roundtrip이 동일한 목록을 만든다.\n */\nexport function normalizeSectionHref(href: string): string | null {\n if (!href) return null\n let normalized = href.replace(/\\\\/g, \"/\").replace(/^\\/+/, \"\")\n if (isPathTraversal(normalized)) return null\n if (/^[Ss]ection\\d+\\.xml$/.test(normalized)) normalized = \"Contents/\" + normalized\n return /(?:^|\\/)[Ss]ection\\d+\\.xml$/.test(normalized) ? normalized : null\n}\n\n/** sectionN.xml을 N 숫자 순서로 정렬 (section10 > section2). */\nexport function compareSectionPaths(a: string, b: string): number {\n const ai = Number(a.match(/[Ss]ection(\\d+)\\.xml$/)?.[1] ?? Number.MAX_SAFE_INTEGER)\n const bi = Number(b.match(/[Ss]ection(\\d+)\\.xml$/)?.[1] ?? Number.MAX_SAFE_INTEGER)\n return ai === bi ? a.localeCompare(b) : ai - bi\n}\n\n// ─── ZIP 안전 로딩 (ZIP bomb 방지) ────────────────────\n\n/** ZIP 비압축 상한 env 의 최대값 (MB) — 오타로 사실상 무제한이 되지 않게 */\nconst MAX_UNZIP_ENV_MB = 8192\n\n/**\n * ZIP 비압축 크기 상한 — `KORDOC_MAX_UNZIP_MB` 가 양수면 그 값(MB, 최대 8192), 아니면 포맷 기본값.\n * 빈 행이 수백만 개인 XLSX 처럼 내용은 작고 XML 만 큰 실문서를 폐쇄망 대량 적재에서 받으려는 설정 (#91)\n */\nexport function unzipLimitBytes(defaultBytes: number): number {\n const mb = Number(process.env.KORDOC_MAX_UNZIP_MB)\n return Number.isFinite(mb) && mb > 0 ? Math.min(mb, MAX_UNZIP_ENV_MB) * 1024 * 1024 : defaultBytes\n}\n\n/**\n * ZIP bomb 사전 검사 — Central Directory에서 비압축 합계와 엔트리 수 확인.\n * HWPX/XLSX/DOCX 등 모든 ZIP 기반 포맷에서 공통 사용.\n */\nexport function precheckZipSize(\n buffer: ArrayBuffer,\n maxUncompressedSize = 256 * 1024 * 1024, // parser-shared MAX_DECOMPRESS_SIZE 와 동기\n maxEntries = 500,\n): { totalUncompressed: number; entryCount: number } {\n try {\n const data = new DataView(buffer)\n const len = buffer.byteLength\n // EOCD 시그니처 역방향 스캔\n let eocdOffset = -1\n for (let i = len - 22; i >= Math.max(0, len - 65557); i--) {\n if (data.getUint32(i, true) === 0x06054b50) { eocdOffset = i; break }\n }\n if (eocdOffset < 0) return { totalUncompressed: 0, entryCount: 0 }\n\n const entryCount = data.getUint16(eocdOffset + 10, true)\n if (entryCount > maxEntries) {\n throw new KordocError(`ZIP 엔트리 수 초과: ${entryCount} (최대 ${maxEntries})`)\n }\n\n const cdSize = data.getUint32(eocdOffset + 12, true)\n const cdOffset = data.getUint32(eocdOffset + 16, true)\n if (cdOffset + cdSize > len) return { totalUncompressed: 0, entryCount }\n\n let totalUncompressed = 0\n let pos = cdOffset\n for (let i = 0; i < entryCount && pos + 46 <= cdOffset + cdSize; i++) {\n if (data.getUint32(pos, true) !== 0x02014b50) break\n totalUncompressed += data.getUint32(pos + 24, true)\n const nameLen = data.getUint16(pos + 28, true)\n const extraLen = data.getUint16(pos + 30, true)\n const commentLen = data.getUint16(pos + 32, true)\n pos += 46 + nameLen + extraLen + commentLen\n }\n\n if (totalUncompressed > maxUncompressedSize) {\n throw new KordocError(`ZIP 비압축 크기 초과: ${(totalUncompressed / 1024 / 1024).toFixed(1)}MB (최대 ${maxUncompressedSize / 1024 / 1024}MB)`)\n }\n\n return { totalUncompressed, entryCount }\n } catch (err) {\n if (err instanceof KordocError) throw err\n return { totalUncompressed: 0, entryCount: 0 }\n }\n}\n\n/**\n * XML 파싱 전 프롤로그 정리 — 선두 BOM 제거 + DOCTYPE 제거(XXE/Billion Laughs 방지).\n *\n * BOM: 일부 OpenXML 라이터가 `[Content_Types].xml`·`.rels` 앞에 EF BB BF를 붙인다.\n * XML 스펙상 적법(엔티티는 BOM으로 시작할 수 있음)이고 엑셀·리브레오피스는 여는데,\n * xmldom은 \"선언이 문서 시작이 아님\" fatalError로 파일 전체를 거부했다 (#63).\n */\nexport function stripDtd(xml: string): string {\n return xml.replace(/^\\uFEFF/, \"\").replace(/<!DOCTYPE\\s[^[>]*(\\[[\\s\\S]*?\\])?\\s*>/gi, \"\")\n}\n\n/** 하이퍼링크 URL 살균 — javascript: 등 XSS 위험 스킴 차단 */\nconst SAFE_HREF_RE = /^(?:https?:|mailto:|tel:|#)/i\nconst HREF_ESCAPE: Record<string, string> = { \"(\": \"%28\", \")\": \"%29\", \"<\": \"%3C\", \">\": \"%3E\", \"\\\"\": \"%22\", \"'\": \"%27\", \"`\": \"%60\" }\nexport function sanitizeHref(href: string): string | null {\n const trimmed = href.trim()\n if (!trimmed || !SAFE_HREF_RE.test(trimmed)) return null\n // 괄호 percent-encoding — 마크다운 링크 목적지에서 불균형 ')'는 링크를 조기 종료시켜\n // 문법이 깨진다 (원문 하이퍼링크에 ')'가 박힌 실문서 존재). %28/%29는 의미 동일.\n // 공백·<>\"'` 도 — 스킴만 보고 통과시키면 \"mailto:a <img src=x onerror=…>\" 가 링크 목적지를 깨고 날 태그로\n // 렌더됐다(pdfjs 는 mailto·tel URI 를 정규화하지 않는다, v4.14.4 리뷰 재현). 인코딩해도 같은 주소다\n return trimmed.replace(/[()<>\"'`]/g, c => HREF_ESCAPE[c]).replace(/\\s/g, c => encodeURIComponent(c))\n}\n\n/**\n * HTML 엔티티 이스케이프 — 글은 `& < >`, 속성값(attr)은 `\"` 까지. HTML 표 경로(builder tableToHtml)와\n * 인쇄 렌더러가 같이 쓰는 한 벌. 읽는 쪽은 unescapeHtml 로 푼다\n */\nexport function escapeHtml(text: string, attr = false): string {\n const s = text.replace(/&/g, \"&\").replace(/</g, \"<\").replace(/>/g, \">\")\n return attr ? s.replace(/\"/g, \""\") : s\n}\n\n/** escapeHtml 의 역 (+ 손으로 쓴 HTML 표의 ') — 한 번에 풀어 \"&lt;\" 는 \"<\" 로 남는다 */\nexport function unescapeHtml(text: string): string {\n return text.replace(/&(lt|gt|quot|#39|amp);/g, (_, e: string) => HTML_ENTITY[e])\n}\nconst HTML_ENTITY: Record<string, string> = { lt: \"<\", gt: \">\", quot: \"\\\"\", \"#39\": \"'\", amp: \"&\" }\n\n/**\n * console.log·info·warn·debug 를 stderr 로 — stdout 이 기계 출력 채널(MCP JSON-RPC·CLI 마크다운/JSON·parse-worker\n * NDJSON)인 진입점용. 라이브러리(pdfjs 등)가 console 로 찍는 경고가 그 채널을 깨지 않게 한다\n */\nexport function routeConsoleToStderr(): void {\n const toStderr = (...args: unknown[]): void => void process.stderr.write(format(...args) + \"\\n\")\n console.log = toStderr\n console.info = toStderr\n console.warn = toStderr\n console.debug = toStderr\n}\n\n// ─── 안전한 min/max (스택 오버플로 방지) ─────────────\n\n/** Math.min(...arr) 대체 — 대형 배열에서 스택 오버플로 방지 */\nexport function safeMin(arr: number[]): number {\n let min = Infinity\n for (let i = 0; i < arr.length; i++) if (arr[i] < min) min = arr[i]\n return min\n}\n\n/** Math.max(...arr) 대체 — 대형 배열에서 스택 오버플로 방지 */\nexport function safeMax(arr: number[]): number {\n let max = -Infinity\n for (let i = 0; i < arr.length; i++) if (arr[i] > max) max = arr[i]\n return max\n}\n\n// ─── 에러 분류 ──────────────────────────────────────\n\nimport type { ErrorCode } from \"./types.js\"\n\n/** 에러를 구조화된 ErrorCode로 분류 — KordocError 메시지 패턴 매칭 */\nexport function classifyError(err: unknown): ErrorCode {\n if (!(err instanceof Error)) return \"PARSE_ERROR\"\n const msg = err.message\n // 입력 파일 없음 — 문서 내용 문제가 아니라 호출자의 경로 문제라 PARSE_ERROR 로 덮지 않는다.\n if (isFileNotFound(err)) return \"FILE_NOT_FOUND\"\n // DRM 을 암호화보다 먼저 — \"DRM 암호화된 HWPX…\" 류 메시지가 ENCRYPTED 로 떨어지지 않도록\n if (msg.includes(\"DRM\")) return \"DRM_PROTECTED\"\n if (msg.includes(\"암호화\") || msg.includes(\"암호로 보호\")) return \"ENCRYPTED\"\n if (msg.includes(\"optional dependency\")) return \"MISSING_DEPENDENCY\"\n // V8 문자열 한계 초과 — 이미지 base64 총량이 큰 문서의 JSON 직렬화 (#65)\n if (msg.includes(\"Invalid string length\") || msg.includes(\"Cannot create a string longer\")) return \"OUTPUT_TOO_LARGE\"\n if (msg.includes(\"ZIP bomb\") || msg.includes(\"ZIP 비압축 크기 초과\") || msg.includes(\"ZIP 엔트리 수 초과\")) return \"ZIP_BOMB\"\n if (msg.includes(\"bomb\") || msg.includes(\"크기 초과\") || msg.includes(\"압축 해제\")) return \"DECOMPRESSION_BOMB\"\n if (msg.includes(\"이미지 기반\")) return \"IMAGE_BASED_PDF\"\n if (msg.includes(\"섹션\") && (msg.includes(\"찾을 수 없\") || msg.includes(\"없음\"))) return \"NO_SECTIONS\"\n if (msg.includes(\"시그니처\") || msg.includes(\"복구할 수 없\")) return \"CORRUPTED\"\n return \"PARSE_ERROR\"\n}\n"],"mappings":";AAMO,IAAM,UAAkB,OAA4C,WAAqB;AAOzF,SAAS,cAAc,KAA0B;AACtD,MAAI,IAAI,eAAe,KAAK,IAAI,eAAe,IAAI,OAAO,YAAY;AACpE,WAAO,IAAI;AAAA,EACb;AACA,SAAO,IAAI,OAAO,MAAM,IAAI,YAAY,IAAI,aAAa,IAAI,UAAU;AACzE;AAMO,IAAM,cAAN,cAA0B,MAAM;AAAA,EACrC,YAAY,SAAiB;AAC3B,UAAM,OAAO;AACb,SAAK,OAAO;AAAA,EACd;AACF;AAMA,SAAS,eAAe,KAAuB;AAC7C,MAAI,OAAO,QAAQ,YAAY,QAAQ,KAAM,QAAO;AACpD,MAAK,IAA8B,SAAS,SAAU,QAAO;AAC7D,QAAM,MAAO,IAAc;AAC3B,SAAO,OAAO,QAAQ,YAAY,oCAAoC,KAAK,GAAG;AAChF;AAMO,SAAS,cAAc,KAAsB;AAClD,MAAI,eAAe,YAAa,QAAO,IAAI;AAI3C,MAAI,eAAe,GAAG,EAAG,QAAO;AAChC,SAAO;AACT;AAMO,SAAS,gBAAgB,MAAuB;AACrD,MAAI,KAAK,SAAS,IAAM,EAAG,QAAO;AAClC,QAAM,aAAa,KAAK,QAAQ,OAAO,GAAG;AAC1C,QAAM,WAAW,WAAW,MAAM,GAAG;AACrC,SAAO,SAAS,KAAK,OAAK,MAAM,IAAI,KAAK,WAAW,WAAW,GAAG,KAAK,aAAa,KAAK,UAAU;AACrG;AASO,SAAS,qBAAqB,MAA6B;AAChE,MAAI,CAAC,KAAM,QAAO;AAClB,MAAI,aAAa,KAAK,QAAQ,OAAO,GAAG,EAAE,QAAQ,QAAQ,EAAE;AAC5D,MAAI,gBAAgB,UAAU,EAAG,QAAO;AACxC,MAAI,uBAAuB,KAAK,UAAU,EAAG,cAAa,cAAc;AACxE,SAAO,8BAA8B,KAAK,UAAU,IAAI,aAAa;AACvE;AAGO,SAAS,oBAAoB,GAAW,GAAmB;AAChE,QAAM,KAAK,OAAO,EAAE,MAAM,uBAAuB,IAAI,CAAC,KAAK,OAAO,gBAAgB;AAClF,QAAM,KAAK,OAAO,EAAE,MAAM,uBAAuB,IAAI,CAAC,KAAK,OAAO,gBAAgB;AAClF,SAAO,OAAO,KAAK,EAAE,cAAc,CAAC,IAAI,KAAK;AAC/C;AAKA,IAAM,mBAAmB;AAMlB,SAAS,gBAAgB,cAA8B;AAC5D,QAAM,KAAK,OAAO,QAAQ,IAAI,mBAAmB;AACjD,SAAO,OAAO,SAAS,EAAE,KAAK,KAAK,IAAI,KAAK,IAAI,IAAI,gBAAgB,IAAI,OAAO,OAAO;AACxF;AAMO,SAAS,gBACd,QACA,sBAAsB,MAAM,OAAO,MACnC,aAAa,KACsC;AACnD,MAAI;AACF,UAAM,OAAO,IAAI,SAAS,MAAM;AAChC,UAAM,MAAM,OAAO;AAEnB,QAAI,aAAa;AACjB,aAAS,IAAI,MAAM,IAAI,KAAK,KAAK,IAAI,GAAG,MAAM,KAAK,GAAG,KAAK;AACzD,UAAI,KAAK,UAAU,GAAG,IAAI,MAAM,WAAY;AAAE,qBAAa;AAAG;AAAA,MAAM;AAAA,IACtE;AACA,QAAI,aAAa,EAAG,QAAO,EAAE,mBAAmB,GAAG,YAAY,EAAE;AAEjE,UAAM,aAAa,KAAK,UAAU,aAAa,IAAI,IAAI;AACvD,QAAI,aAAa,YAAY;AAC3B,YAAM,IAAI,YAAY,+CAAiB,UAAU,kBAAQ,UAAU,GAAG;AAAA,IACxE;AAEA,UAAM,SAAS,KAAK,UAAU,aAAa,IAAI,IAAI;AACnD,UAAM,WAAW,KAAK,UAAU,aAAa,IAAI,IAAI;AACrD,QAAI,WAAW,SAAS,IAAK,QAAO,EAAE,mBAAmB,GAAG,WAAW;AAEvE,QAAI,oBAAoB;AACxB,QAAI,MAAM;AACV,aAAS,IAAI,GAAG,IAAI,cAAc,MAAM,MAAM,WAAW,QAAQ,KAAK;AACpE,UAAI,KAAK,UAAU,KAAK,IAAI,MAAM,SAAY;AAC9C,2BAAqB,KAAK,UAAU,MAAM,IAAI,IAAI;AAClD,YAAM,UAAU,KAAK,UAAU,MAAM,IAAI,IAAI;AAC7C,YAAM,WAAW,KAAK,UAAU,MAAM,IAAI,IAAI;AAC9C,YAAM,aAAa,KAAK,UAAU,MAAM,IAAI,IAAI;AAChD,aAAO,KAAK,UAAU,WAAW;AAAA,IACnC;AAEA,QAAI,oBAAoB,qBAAqB;AAC3C,YAAM,IAAI,YAAY,sDAAmB,oBAAoB,OAAO,MAAM,QAAQ,CAAC,CAAC,oBAAU,sBAAsB,OAAO,IAAI,KAAK;AAAA,IACtI;AAEA,WAAO,EAAE,mBAAmB,WAAW;AAAA,EACzC,SAAS,KAAK;AACZ,QAAI,eAAe,YAAa,OAAM;AACtC,WAAO,EAAE,mBAAmB,GAAG,YAAY,EAAE;AAAA,EAC/C;AACF;AASO,SAAS,SAAS,KAAqB;AAC5C,SAAO,IAAI,QAAQ,WAAW,EAAE,EAAE,QAAQ,0CAA0C,EAAE;AACxF;AAGA,IAAM,eAAe;AACrB,IAAM,cAAsC,EAAE,KAAK,OAAO,KAAK,OAAO,KAAK,OAAO,KAAK,OAAO,KAAM,OAAO,KAAK,OAAO,KAAK,MAAM;AAC3H,SAAS,aAAa,MAA6B;AACxD,QAAM,UAAU,KAAK,KAAK;AAC1B,MAAI,CAAC,WAAW,CAAC,aAAa,KAAK,OAAO,EAAG,QAAO;AAKpD,SAAO,QAAQ,QAAQ,cAAc,OAAK,YAAY,CAAC,CAAC,EAAE,QAAQ,OAAO,OAAK,mBAAmB,CAAC,CAAC;AACrG;AAMO,SAAS,WAAW,MAAc,OAAO,OAAe;AAC7D,QAAM,IAAI,KAAK,QAAQ,MAAM,OAAO,EAAE,QAAQ,MAAM,MAAM,EAAE,QAAQ,MAAM,MAAM;AAChF,SAAO,OAAO,EAAE,QAAQ,MAAM,QAAQ,IAAI;AAC5C;AAGO,SAAS,aAAa,MAAsB;AACjD,SAAO,KAAK,QAAQ,2BAA2B,CAAC,GAAG,MAAc,YAAY,CAAC,CAAC;AACjF;AACA,IAAM,cAAsC,EAAE,IAAI,KAAK,IAAI,KAAK,MAAM,KAAM,OAAO,KAAK,KAAK,IAAI;AAiB1F,SAAS,QAAQ,KAAuB;AAC7C,MAAI,MAAM;AACV,WAAS,IAAI,GAAG,IAAI,IAAI,QAAQ,IAAK,KAAI,IAAI,CAAC,IAAI,IAAK,OAAM,IAAI,CAAC;AAClE,SAAO;AACT;AAGO,SAAS,QAAQ,KAAuB;AAC7C,MAAI,MAAM;AACV,WAAS,IAAI,GAAG,IAAI,IAAI,QAAQ,IAAK,KAAI,IAAI,CAAC,IAAI,IAAK,OAAM,IAAI,CAAC;AAClE,SAAO;AACT;AAOO,SAAS,cAAc,KAAyB;AACrD,MAAI,EAAE,eAAe,OAAQ,QAAO;AACpC,QAAM,MAAM,IAAI;AAEhB,MAAI,eAAe,GAAG,EAAG,QAAO;AAEhC,MAAI,IAAI,SAAS,KAAK,EAAG,QAAO;AAChC,MAAI,IAAI,SAAS,oBAAK,KAAK,IAAI,SAAS,iCAAQ,EAAG,QAAO;AAC1D,MAAI,IAAI,SAAS,qBAAqB,EAAG,QAAO;AAEhD,MAAI,IAAI,SAAS,uBAAuB,KAAK,IAAI,SAAS,+BAA+B,EAAG,QAAO;AACnG,MAAI,IAAI,SAAS,UAAU,KAAK,IAAI,SAAS,kDAAe,KAAK,IAAI,SAAS,4CAAc,EAAG,QAAO;AACtG,MAAI,IAAI,SAAS,MAAM,KAAK,IAAI,SAAS,2BAAO,KAAK,IAAI,SAAS,2BAAO,EAAG,QAAO;AACnF,MAAI,IAAI,SAAS,iCAAQ,EAAG,QAAO;AACnC,MAAI,IAAI,SAAS,cAAI,MAAM,IAAI,SAAS,4BAAQ,KAAK,IAAI,SAAS,cAAI,GAAI,QAAO;AACjF,MAAI,IAAI,SAAS,0BAAM,KAAK,IAAI,SAAS,kCAAS,EAAG,QAAO;AAC5D,SAAO;AACT;","names":[]}
|
|
1
|
+
{"version":3,"sources":["../src/utils.ts"],"sourcesContent":["/** kordoc 공용 유틸리티 */\n\nimport { format } from \"util\"\n\n/** 빌드 타임에 tsup define으로 주입되는 버전 */\ndeclare const __KORDOC_VERSION__: string\nexport const VERSION: string = typeof __KORDOC_VERSION__ !== \"undefined\" ? __KORDOC_VERSION__ : \"0.0.0-dev\"\n\n/**\n * Node.js Buffer → ArrayBuffer 변환\n * pool Buffer의 공유 ArrayBuffer 문제를 안전하게 처리.\n * offset=0이고 전체 ArrayBuffer를 차지하면 복사 없이 직접 반환.\n */\nexport function toArrayBuffer(buf: Buffer): ArrayBuffer {\n if (buf.byteOffset === 0 && buf.byteLength === buf.buffer.byteLength) {\n return buf.buffer as ArrayBuffer\n }\n return buf.buffer.slice(buf.byteOffset, buf.byteOffset + buf.byteLength) as ArrayBuffer\n}\n\n/**\n * kordoc 내부 에러 클래스 — 사용자에게 노출해도 안전한 메시지만 포함.\n * MCP 에러 정제에서 instanceof로 판별하여 allowlist 패턴 매칭 없이 안전하게 통과.\n */\nexport class KordocError extends Error {\n constructor(message: string) {\n super(message)\n this.name = \"KordocError\"\n }\n}\n\n/**\n * ENOENT 판별 — 입력 경로가 존재하지 않음.\n * 래핑된 KordocError 를 뚫고 올라온 경우도 있으므로 메시지까지 함께 본다.\n */\nfunction isFileNotFound(err: unknown): boolean {\n if (typeof err !== \"object\" || err === null) return false\n if ((err as NodeJS.ErrnoException).code === \"ENOENT\") return true\n const msg = (err as Error).message\n return typeof msg === \"string\" && /ENOENT|no such file or directory/i.test(msg)\n}\n\n/**\n * 에러 메시지 정제 — KordocError는 그대로, ENOENT 만 구분 문구, 나머지는 일반 메시지로 대체.\n * 파일시스템 경로, 스택 트레이스 등 내부 정보 노출 방지.\n */\nexport function sanitizeError(err: unknown): string {\n if (err instanceof KordocError) return err.message\n // 없는 파일은 문서 문제가 아니라 고칠 수 있는 입력 문제다 — 사유를 숨기지 않는다.\n // 경로는 노출하지 않는다(어느 파일인지는 stderr·실패 JSON 의 basename 이 알린다).\n // 문구는 MCP describeError 의 ENOENT 힌트와 같은 표현을 쓴다.\n if (isFileNotFound(err)) return \"파일 또는 디렉토리를 찾을 수 없습니다\"\n return \"문서 처리 중 오류가 발생했습니다\"\n}\n\n/**\n * ZIP 엔트리 경로의 경로 순회 여부 판별.\n * 백슬래시 정규화, .., 절대경로, Windows 드라이브 문자 모두 차단.\n */\nexport function isPathTraversal(name: string): boolean {\n if (name.includes(\"\\x00\")) return true\n const normalized = name.replace(/\\\\/g, \"/\")\n const segments = normalized.split(\"/\")\n return segments.some(s => s === \"..\") || normalized.startsWith(\"/\") || /^[A-Za-z]:/.test(normalized)\n}\n\n// ─── HWPX 섹션 href 해석 (parser/roundtrip 공용) ────────────────────\n\n/**\n * manifest href를 본문 섹션 경로(`Contents/sectionN.xml`)로 정규화.\n * 비본문 XML(header/script/settings)·path traversal·드라이브 경로는 null로 거른다.\n * 백슬래시 구분자를 슬래시로 통일해 parser와 roundtrip이 동일한 목록을 만든다.\n */\nexport function normalizeSectionHref(href: string): string | null {\n if (!href) return null\n let normalized = href.replace(/\\\\/g, \"/\").replace(/^\\/+/, \"\")\n if (isPathTraversal(normalized)) return null\n if (/^[Ss]ection\\d+\\.xml$/.test(normalized)) normalized = \"Contents/\" + normalized\n return /(?:^|\\/)[Ss]ection\\d+\\.xml$/.test(normalized) ? normalized : null\n}\n\n/** sectionN.xml을 N 숫자 순서로 정렬 (section10 > section2). */\nexport function compareSectionPaths(a: string, b: string): number {\n const ai = Number(a.match(/[Ss]ection(\\d+)\\.xml$/)?.[1] ?? Number.MAX_SAFE_INTEGER)\n const bi = Number(b.match(/[Ss]ection(\\d+)\\.xml$/)?.[1] ?? Number.MAX_SAFE_INTEGER)\n return ai === bi ? a.localeCompare(b) : ai - bi\n}\n\n// ─── ZIP 안전 로딩 (ZIP bomb 방지) ────────────────────\n\n/** ZIP 비압축 상한 env 의 최대값 (MB) — 오타로 사실상 무제한이 되지 않게 */\nconst MAX_UNZIP_ENV_MB = 8192\n\n/**\n * ZIP 비압축 크기 상한 — `KORDOC_MAX_UNZIP_MB` 가 양수면 그 값(MB, 최대 8192), 아니면 포맷 기본값.\n * 빈 행이 수백만 개인 XLSX 처럼 내용은 작고 XML 만 큰 실문서를 폐쇄망 대량 적재에서 받으려는 설정 (#91)\n */\nexport function unzipLimitBytes(defaultBytes: number): number {\n const mb = Number(process.env.KORDOC_MAX_UNZIP_MB)\n return Number.isFinite(mb) && mb > 0 ? Math.min(mb, MAX_UNZIP_ENV_MB) * 1024 * 1024 : defaultBytes\n}\n\n/**\n * ZIP bomb 사전 검사 — Central Directory에서 비압축 합계와 엔트리 수 확인.\n * HWPX/XLSX/DOCX 등 모든 ZIP 기반 포맷에서 공통 사용.\n */\nexport function precheckZipSize(\n buffer: ArrayBuffer,\n maxUncompressedSize = 256 * 1024 * 1024, // parser-shared MAX_DECOMPRESS_SIZE 와 동기\n maxEntries = 500,\n): { totalUncompressed: number; entryCount: number } {\n try {\n const data = new DataView(buffer)\n const len = buffer.byteLength\n // EOCD 시그니처 역방향 스캔\n let eocdOffset = -1\n for (let i = len - 22; i >= Math.max(0, len - 65557); i--) {\n if (data.getUint32(i, true) === 0x06054b50) { eocdOffset = i; break }\n }\n if (eocdOffset < 0) return { totalUncompressed: 0, entryCount: 0 }\n\n const entryCount = data.getUint16(eocdOffset + 10, true)\n if (entryCount > maxEntries) {\n throw new KordocError(`ZIP 엔트리 수 초과: ${entryCount} (최대 ${maxEntries})`)\n }\n\n const cdSize = data.getUint32(eocdOffset + 12, true)\n const cdOffset = data.getUint32(eocdOffset + 16, true)\n if (cdOffset + cdSize > len) return { totalUncompressed: 0, entryCount }\n\n let totalUncompressed = 0\n let pos = cdOffset\n for (let i = 0; i < entryCount && pos + 46 <= cdOffset + cdSize; i++) {\n if (data.getUint32(pos, true) !== 0x02014b50) break\n totalUncompressed += data.getUint32(pos + 24, true)\n const nameLen = data.getUint16(pos + 28, true)\n const extraLen = data.getUint16(pos + 30, true)\n const commentLen = data.getUint16(pos + 32, true)\n pos += 46 + nameLen + extraLen + commentLen\n }\n\n if (totalUncompressed > maxUncompressedSize) {\n throw new KordocError(`ZIP 비압축 크기 초과: ${(totalUncompressed / 1024 / 1024).toFixed(1)}MB (최대 ${maxUncompressedSize / 1024 / 1024}MB)`)\n }\n\n return { totalUncompressed, entryCount }\n } catch (err) {\n if (err instanceof KordocError) throw err\n return { totalUncompressed: 0, entryCount: 0 }\n }\n}\n\n/**\n * XML 파싱 전 프롤로그 정리 — 선두 BOM 제거 + DOCTYPE 제거(XXE/Billion Laughs 방지).\n *\n * BOM: 일부 OpenXML 라이터가 `[Content_Types].xml`·`.rels` 앞에 EF BB BF를 붙인다.\n * XML 스펙상 적법(엔티티는 BOM으로 시작할 수 있음)이고 엑셀·리브레오피스는 여는데,\n * xmldom은 \"선언이 문서 시작이 아님\" fatalError로 파일 전체를 거부했다 (#63).\n */\nexport function stripDtd(xml: string): string {\n return xml.replace(/^\\uFEFF/, \"\").replace(/<!DOCTYPE\\s[^[>]*(\\[[\\s\\S]*?\\])?\\s*>/gi, \"\")\n}\n\n/** 하이퍼링크 URL 살균 — javascript: 등 XSS 위험 스킴 차단 */\nconst SAFE_HREF_RE = /^(?:https?:|mailto:|tel:|#)/i\nconst HREF_ESCAPE: Record<string, string> = { \"(\": \"%28\", \")\": \"%29\", \"<\": \"%3C\", \">\": \"%3E\", \"\\\"\": \"%22\", \"'\": \"%27\", \"`\": \"%60\" }\nexport function sanitizeHref(href: string): string | null {\n const trimmed = href.trim()\n if (!trimmed || !SAFE_HREF_RE.test(trimmed)) return null\n // 괄호 percent-encoding — 마크다운 링크 목적지에서 불균형 ')'는 링크를 조기 종료시켜\n // 문법이 깨진다 (원문 하이퍼링크에 ')'가 박힌 실문서 존재). %28/%29는 의미 동일.\n // 공백·<>\"'` 도 — 스킴만 보고 통과시키면 \"mailto:a <img src=x onerror=…>\" 가 링크 목적지를 깨고 날 태그로\n // 렌더됐다(pdfjs 는 mailto·tel URI 를 정규화하지 않는다, v4.14.4 리뷰 재현). 인코딩해도 같은 주소다\n return trimmed.replace(/[()<>\"'`]/g, c => HREF_ESCAPE[c]).replace(/\\s/g, c => encodeURIComponent(c))\n}\n\n/**\n * HTML 엔티티 이스케이프 — 글은 `& < >`, 속성값(attr)은 `\"` 까지. HTML 표 경로(builder tableToHtml)와\n * 인쇄 렌더러가 같이 쓰는 한 벌. 읽는 쪽은 unescapeHtml 로 푼다\n */\nexport function escapeHtml(text: string, attr = false): string {\n const s = text.replace(/&/g, \"&\").replace(/</g, \"<\").replace(/>/g, \">\")\n return attr ? s.replace(/\"/g, \""\") : s\n}\n\n/** escapeHtml 의 역 (+ 손으로 쓴 HTML 표의 ') — 한 번에 풀어 \"&lt;\" 는 \"<\" 로 남는다 */\nexport function unescapeHtml(text: string): string {\n return text.replace(/&(lt|gt|quot|#39|amp);/g, (_, e: string) => HTML_ENTITY[e])\n}\nconst HTML_ENTITY: Record<string, string> = { lt: \"<\", gt: \">\", quot: \"\\\"\", \"#39\": \"'\", amp: \"&\" }\n\n/**\n * console.log·info·warn·debug 를 stderr 로 — stdout 이 기계 출력 채널(MCP JSON-RPC·CLI 마크다운/JSON·parse-worker\n * NDJSON)인 진입점용. 라이브러리(pdfjs 등)가 console 로 찍는 경고가 그 채널을 깨지 않게 한다\n */\nexport function routeConsoleToStderr(): void {\n const toStderr = (...args: unknown[]): void => void process.stderr.write(format(...args) + \"\\n\")\n console.log = toStderr\n console.info = toStderr\n console.warn = toStderr\n console.debug = toStderr\n}\n\n// ─── 안전한 min/max (스택 오버플로 방지) ─────────────\n\n/** Math.min(...arr) 대체 — 대형 배열에서 스택 오버플로 방지 */\nexport function safeMin(arr: number[]): number {\n let min = Infinity\n for (let i = 0; i < arr.length; i++) if (arr[i] < min) min = arr[i]\n return min\n}\n\n/** Math.max(...arr) 대체 — 대형 배열에서 스택 오버플로 방지 */\nexport function safeMax(arr: number[]): number {\n let max = -Infinity\n for (let i = 0; i < arr.length; i++) if (arr[i] > max) max = arr[i]\n return max\n}\n\n// ─── 에러 분류 ──────────────────────────────────────\n\nimport type { ErrorCode } from \"./types.js\"\n\n/** 에러를 구조화된 ErrorCode로 분류 — KordocError 메시지 패턴 매칭 */\n/** 선택 의존 설치 실패 안내 꼬리 — 네트워크가 막힌 linux/x64 에서 onnxruntime-node 설치 스크립트가 CUDA 바이너리를 받다 실패하면\n * npm 이 sharp·@huggingface/transformers 까지 선택 의존 설치에서 뺀다(#99). npx 캐시 실행은 npm install 로 못 고친다 */\nexport const OPTIONAL_DEP_INSTALL_HINT =\n \" 네트워크가 제한된 linux/x64 에서 npx 로 설치했다면 onnxruntime-node 의 CUDA 다운로드 실패로 함께 빠졌을 수 있습니다 — \" +\n \"`ONNXRUNTIME_NODE_INSTALL=skip npx -y kordoc@^4 …` 로 다시 설치하세요.\"\n\nexport function classifyError(err: unknown): ErrorCode {\n if (!(err instanceof Error)) return \"PARSE_ERROR\"\n const msg = err.message\n // 입력 파일 없음 — 문서 내용 문제가 아니라 호출자의 경로 문제라 PARSE_ERROR 로 덮지 않는다.\n if (isFileNotFound(err)) return \"FILE_NOT_FOUND\"\n // DRM 을 암호화보다 먼저 — \"DRM 암호화된 HWPX…\" 류 메시지가 ENCRYPTED 로 떨어지지 않도록\n if (msg.includes(\"DRM\")) return \"DRM_PROTECTED\"\n if (msg.includes(\"암호화\") || msg.includes(\"암호로 보호\")) return \"ENCRYPTED\"\n if (msg.includes(\"optional dependency\")) return \"MISSING_DEPENDENCY\"\n // V8 문자열 한계 초과 — 이미지 base64 총량이 큰 문서의 JSON 직렬화 (#65)\n if (msg.includes(\"Invalid string length\") || msg.includes(\"Cannot create a string longer\")) return \"OUTPUT_TOO_LARGE\"\n if (msg.includes(\"ZIP bomb\") || msg.includes(\"ZIP 비압축 크기 초과\") || msg.includes(\"ZIP 엔트리 수 초과\")) return \"ZIP_BOMB\"\n if (msg.includes(\"bomb\") || msg.includes(\"크기 초과\") || msg.includes(\"압축 해제\")) return \"DECOMPRESSION_BOMB\"\n if (msg.includes(\"이미지 기반\")) return \"IMAGE_BASED_PDF\"\n if (msg.includes(\"섹션\") && (msg.includes(\"찾을 수 없\") || msg.includes(\"없음\"))) return \"NO_SECTIONS\"\n if (msg.includes(\"시그니처\") || msg.includes(\"복구할 수 없\")) return \"CORRUPTED\"\n return \"PARSE_ERROR\"\n}\n"],"mappings":";;;AAEA,SAAS,cAAc;AAIhB,IAAM,UAAkB,OAA4C,WAAqB;AAOzF,SAAS,cAAc,KAA0B;AACtD,MAAI,IAAI,eAAe,KAAK,IAAI,eAAe,IAAI,OAAO,YAAY;AACpE,WAAO,IAAI;AAAA,EACb;AACA,SAAO,IAAI,OAAO,MAAM,IAAI,YAAY,IAAI,aAAa,IAAI,UAAU;AACzE;AAMO,IAAM,cAAN,cAA0B,MAAM;AAAA,EACrC,YAAY,SAAiB;AAC3B,UAAM,OAAO;AACb,SAAK,OAAO;AAAA,EACd;AACF;AAMA,SAAS,eAAe,KAAuB;AAC7C,MAAI,OAAO,QAAQ,YAAY,QAAQ,KAAM,QAAO;AACpD,MAAK,IAA8B,SAAS,SAAU,QAAO;AAC7D,QAAM,MAAO,IAAc;AAC3B,SAAO,OAAO,QAAQ,YAAY,oCAAoC,KAAK,GAAG;AAChF;AAMO,SAAS,cAAc,KAAsB;AAClD,MAAI,eAAe,YAAa,QAAO,IAAI;AAI3C,MAAI,eAAe,GAAG,EAAG,QAAO;AAChC,SAAO;AACT;AAMO,SAAS,gBAAgB,MAAuB;AACrD,MAAI,KAAK,SAAS,IAAM,EAAG,QAAO;AAClC,QAAM,aAAa,KAAK,QAAQ,OAAO,GAAG;AAC1C,QAAM,WAAW,WAAW,MAAM,GAAG;AACrC,SAAO,SAAS,KAAK,OAAK,MAAM,IAAI,KAAK,WAAW,WAAW,GAAG,KAAK,aAAa,KAAK,UAAU;AACrG;AASO,SAAS,qBAAqB,MAA6B;AAChE,MAAI,CAAC,KAAM,QAAO;AAClB,MAAI,aAAa,KAAK,QAAQ,OAAO,GAAG,EAAE,QAAQ,QAAQ,EAAE;AAC5D,MAAI,gBAAgB,UAAU,EAAG,QAAO;AACxC,MAAI,uBAAuB,KAAK,UAAU,EAAG,cAAa,cAAc;AACxE,SAAO,8BAA8B,KAAK,UAAU,IAAI,aAAa;AACvE;AAGO,SAAS,oBAAoB,GAAW,GAAmB;AAChE,QAAM,KAAK,OAAO,EAAE,MAAM,uBAAuB,IAAI,CAAC,KAAK,OAAO,gBAAgB;AAClF,QAAM,KAAK,OAAO,EAAE,MAAM,uBAAuB,IAAI,CAAC,KAAK,OAAO,gBAAgB;AAClF,SAAO,OAAO,KAAK,EAAE,cAAc,CAAC,IAAI,KAAK;AAC/C;AAKA,IAAM,mBAAmB;AAMlB,SAAS,gBAAgB,cAA8B;AAC5D,QAAM,KAAK,OAAO,QAAQ,IAAI,mBAAmB;AACjD,SAAO,OAAO,SAAS,EAAE,KAAK,KAAK,IAAI,KAAK,IAAI,IAAI,gBAAgB,IAAI,OAAO,OAAO;AACxF;AAMO,SAAS,gBACd,QACA,sBAAsB,MAAM,OAAO,MACnC,aAAa,KACsC;AACnD,MAAI;AACF,UAAM,OAAO,IAAI,SAAS,MAAM;AAChC,UAAM,MAAM,OAAO;AAEnB,QAAI,aAAa;AACjB,aAAS,IAAI,MAAM,IAAI,KAAK,KAAK,IAAI,GAAG,MAAM,KAAK,GAAG,KAAK;AACzD,UAAI,KAAK,UAAU,GAAG,IAAI,MAAM,WAAY;AAAE,qBAAa;AAAG;AAAA,MAAM;AAAA,IACtE;AACA,QAAI,aAAa,EAAG,QAAO,EAAE,mBAAmB,GAAG,YAAY,EAAE;AAEjE,UAAM,aAAa,KAAK,UAAU,aAAa,IAAI,IAAI;AACvD,QAAI,aAAa,YAAY;AAC3B,YAAM,IAAI,YAAY,+CAAiB,UAAU,kBAAQ,UAAU,GAAG;AAAA,IACxE;AAEA,UAAM,SAAS,KAAK,UAAU,aAAa,IAAI,IAAI;AACnD,UAAM,WAAW,KAAK,UAAU,aAAa,IAAI,IAAI;AACrD,QAAI,WAAW,SAAS,IAAK,QAAO,EAAE,mBAAmB,GAAG,WAAW;AAEvE,QAAI,oBAAoB;AACxB,QAAI,MAAM;AACV,aAAS,IAAI,GAAG,IAAI,cAAc,MAAM,MAAM,WAAW,QAAQ,KAAK;AACpE,UAAI,KAAK,UAAU,KAAK,IAAI,MAAM,SAAY;AAC9C,2BAAqB,KAAK,UAAU,MAAM,IAAI,IAAI;AAClD,YAAM,UAAU,KAAK,UAAU,MAAM,IAAI,IAAI;AAC7C,YAAM,WAAW,KAAK,UAAU,MAAM,IAAI,IAAI;AAC9C,YAAM,aAAa,KAAK,UAAU,MAAM,IAAI,IAAI;AAChD,aAAO,KAAK,UAAU,WAAW;AAAA,IACnC;AAEA,QAAI,oBAAoB,qBAAqB;AAC3C,YAAM,IAAI,YAAY,sDAAmB,oBAAoB,OAAO,MAAM,QAAQ,CAAC,CAAC,oBAAU,sBAAsB,OAAO,IAAI,KAAK;AAAA,IACtI;AAEA,WAAO,EAAE,mBAAmB,WAAW;AAAA,EACzC,SAAS,KAAK;AACZ,QAAI,eAAe,YAAa,OAAM;AACtC,WAAO,EAAE,mBAAmB,GAAG,YAAY,EAAE;AAAA,EAC/C;AACF;AASO,SAAS,SAAS,KAAqB;AAC5C,SAAO,IAAI,QAAQ,WAAW,EAAE,EAAE,QAAQ,0CAA0C,EAAE;AACxF;AAGA,IAAM,eAAe;AACrB,IAAM,cAAsC,EAAE,KAAK,OAAO,KAAK,OAAO,KAAK,OAAO,KAAK,OAAO,KAAM,OAAO,KAAK,OAAO,KAAK,MAAM;AAC3H,SAAS,aAAa,MAA6B;AACxD,QAAM,UAAU,KAAK,KAAK;AAC1B,MAAI,CAAC,WAAW,CAAC,aAAa,KAAK,OAAO,EAAG,QAAO;AAKpD,SAAO,QAAQ,QAAQ,cAAc,OAAK,YAAY,CAAC,CAAC,EAAE,QAAQ,OAAO,OAAK,mBAAmB,CAAC,CAAC;AACrG;AAMO,SAAS,WAAW,MAAc,OAAO,OAAe;AAC7D,QAAM,IAAI,KAAK,QAAQ,MAAM,OAAO,EAAE,QAAQ,MAAM,MAAM,EAAE,QAAQ,MAAM,MAAM;AAChF,SAAO,OAAO,EAAE,QAAQ,MAAM,QAAQ,IAAI;AAC5C;AAGO,SAAS,aAAa,MAAsB;AACjD,SAAO,KAAK,QAAQ,2BAA2B,CAAC,GAAG,MAAc,YAAY,CAAC,CAAC;AACjF;AACA,IAAM,cAAsC,EAAE,IAAI,KAAK,IAAI,KAAK,MAAM,KAAM,OAAO,KAAK,KAAK,IAAI;AAM1F,SAAS,uBAA6B;AAC3C,QAAM,WAAW,IAAI,SAA0B,KAAK,QAAQ,OAAO,MAAM,OAAO,GAAG,IAAI,IAAI,IAAI;AAC/F,UAAQ,MAAM;AACd,UAAQ,OAAO;AACf,UAAQ,OAAO;AACf,UAAQ,QAAQ;AAClB;AAKO,SAAS,QAAQ,KAAuB;AAC7C,MAAI,MAAM;AACV,WAAS,IAAI,GAAG,IAAI,IAAI,QAAQ,IAAK,KAAI,IAAI,CAAC,IAAI,IAAK,OAAM,IAAI,CAAC;AAClE,SAAO;AACT;AAGO,SAAS,QAAQ,KAAuB;AAC7C,MAAI,MAAM;AACV,WAAS,IAAI,GAAG,IAAI,IAAI,QAAQ,IAAK,KAAI,IAAI,CAAC,IAAI,IAAK,OAAM,IAAI,CAAC;AAClE,SAAO;AACT;AASO,IAAM,4BACX;AAGK,SAAS,cAAc,KAAyB;AACrD,MAAI,EAAE,eAAe,OAAQ,QAAO;AACpC,QAAM,MAAM,IAAI;AAEhB,MAAI,eAAe,GAAG,EAAG,QAAO;AAEhC,MAAI,IAAI,SAAS,KAAK,EAAG,QAAO;AAChC,MAAI,IAAI,SAAS,oBAAK,KAAK,IAAI,SAAS,iCAAQ,EAAG,QAAO;AAC1D,MAAI,IAAI,SAAS,qBAAqB,EAAG,QAAO;AAEhD,MAAI,IAAI,SAAS,uBAAuB,KAAK,IAAI,SAAS,+BAA+B,EAAG,QAAO;AACnG,MAAI,IAAI,SAAS,UAAU,KAAK,IAAI,SAAS,kDAAe,KAAK,IAAI,SAAS,4CAAc,EAAG,QAAO;AACtG,MAAI,IAAI,SAAS,MAAM,KAAK,IAAI,SAAS,2BAAO,KAAK,IAAI,SAAS,2BAAO,EAAG,QAAO;AACnF,MAAI,IAAI,SAAS,iCAAQ,EAAG,QAAO;AACnC,MAAI,IAAI,SAAS,cAAI,MAAM,IAAI,SAAS,4BAAQ,KAAK,IAAI,SAAS,cAAI,GAAI,QAAO;AACjF,MAAI,IAAI,SAAS,0BAAM,KAAK,IAAI,SAAS,kCAAS,EAAG,QAAO;AAC5D,SAAO;AACT;","names":[]}
|