kordoc 4.24.1 → 4.24.2

package/dist/{chunk-JSI5T77E.cjs → chunk-PNX5FYHC.cjs}

@@ -10,14 +10,14 @@ var _chunkTK3FDAEEcjs = require('./chunk-TK3FDAEE.cjs');
 
 
 
-var _chunk4VNIEHWMcjs = require('./chunk-4VNIEHWM.cjs');
+var _chunkUNICY546cjs = require('./chunk-UNICY546.cjs');
 
 
-var _chunk5QBTJ4LAcjs = require('./chunk-5QBTJ4LA.cjs');
+var _chunkJFUUFJZWcjs = require('./chunk-JFUUFJZW.cjs');
 
 
 
-var _chunkTYXXTWPHcjs = require('./chunk-TYXXTWPH.cjs');
+var _chunkPNSALBXHcjs = require('./chunk-PNSALBXH.cjs');
 
 // src/ocr/ruling-lines.ts
 var INK_LUMA_MAX = 205;
@@ -1410,7 +1410,7 @@ var OcrEngine = class _OcrEngine {
     this.ort = parts.ort;
     this.sharp = parts.sharp;
   }
-  static async create(models = _chunk4VNIEHWMcjs.DEFAULT_OCR_MODELS) {
+  static async create(models = _chunkUNICY546cjs.DEFAULT_OCR_MODELS) {
     _chunkTK3FDAEEcjs.disableOrtTelemetry.call(void 0, );
     const [ortMod, sharpModRaw] = await Promise.all([
       tryImport("onnxruntime-node", () => Promise.resolve().then(() => _interopRequireWildcard(require("onnxruntime-node")))),
@@ -1430,16 +1430,16 @@ var OcrEngine = class _OcrEngine {
     const local = (s) => "spec" in s ? Promise.resolve(null) : verifiedBytes(s);
     const [detBytes, recBytes, dictBytes] = await Promise.all([local(models.det), local(models.rec), local(models.dict)]);
     const session = (s, bytes) => bytes ? ortMod.InferenceSession.create(bytes, sessionOpts).catch(() => {
-      throw new (0, _chunkTYXXTWPHcjs.KordocError)(`OCR \uBAA8\uB378\uC744 \uC5F4\uC9C0 \uBABB\uD588\uC2B5\uB2C8\uB2E4: ${_path.basename.call(void 0, _chunk4VNIEHWMcjs.ocrSourcePath.call(void 0, s))} \u2014 ONNX \uBAA8\uB378\uC774 \uC544\uB2C8\uAC70\uB098 \uC190\uC0C1\uB410\uC2B5\uB2C8\uB2E4`);
-    }) : ortMod.InferenceSession.create(_chunk4VNIEHWMcjs.ocrSourcePath.call(void 0, s), sessionOpts);
+      throw new (0, _chunkPNSALBXHcjs.KordocError)(`OCR \uBAA8\uB378\uC744 \uC5F4\uC9C0 \uBABB\uD588\uC2B5\uB2C8\uB2E4: ${_path.basename.call(void 0, _chunkUNICY546cjs.ocrSourcePath.call(void 0, s))} \u2014 ONNX \uBAA8\uB378\uC774 \uC544\uB2C8\uAC70\uB098 \uC190\uC0C1\uB410\uC2B5\uB2C8\uB2E4`);
+    }) : ortMod.InferenceSession.create(_chunkUNICY546cjs.ocrSourcePath.call(void 0, s), sessionOpts);
     const [det, rec, dictYml] = await Promise.all([
       session(models.det, detBytes),
       session(models.rec, recBytes),
-      dictBytes ? new TextDecoder().decode(dictBytes) : _promises.readFile.call(void 0, _chunk4VNIEHWMcjs.ocrSourcePath.call(void 0, models.dict), "utf-8")
+      dictBytes ? new TextDecoder().decode(dictBytes) : _promises.readFile.call(void 0, _chunkUNICY546cjs.ocrSourcePath.call(void 0, models.dict), "utf-8")
     ]);
-    const dict = _chunk4VNIEHWMcjs.parseCharacterDict.call(void 0, dictYml);
+    const dict = _chunkUNICY546cjs.parseCharacterDict.call(void 0, dictYml);
     if (dict.length === 0) {
-      throw "spec" in models.dict ? new Error("OCR \uC0AC\uC804 \uD30C\uC2F1 \uC2E4\uD328 \u2014 \uBAA8\uB378 \uCE90\uC2DC\uB97C \uC0AD\uC81C \uD6C4 \uC7AC\uB2E4\uC6B4\uB85C\uB4DC\uD558\uC138\uC694") : new (0, _chunkTYXXTWPHcjs.KordocError)(`OCR \uC0AC\uC804 \uD30C\uC2F1 \uC2E4\uD328 \u2014 ${_path.basename.call(void 0, models.dict.file.path)} \uC5D0 PostProcess.character_dict \uBAA9\uB85D\uC774 \uC5C6\uC2B5\uB2C8\uB2E4`);
+      throw "spec" in models.dict ? new Error("OCR \uC0AC\uC804 \uD30C\uC2F1 \uC2E4\uD328 \u2014 \uBAA8\uB378 \uCE90\uC2DC\uB97C \uC0AD\uC81C \uD6C4 \uC7AC\uB2E4\uC6B4\uB85C\uB4DC\uD558\uC138\uC694") : new (0, _chunkPNSALBXHcjs.KordocError)(`OCR \uC0AC\uC804 \uD30C\uC2F1 \uC2E4\uD328 \u2014 ${_path.basename.call(void 0, models.dict.file.path)} \uC5D0 PostProcess.character_dict \uBAA9\uB85D\uC774 \uC5C6\uC2B5\uB2C8\uB2E4`);
     }
     return new _OcrEngine({ det, rec, dict, ort: ortMod, sharp: sharpMod });
   }
@@ -1648,7 +1648,7 @@ var OcrEngine = class _OcrEngine {
       const out = await this.rec.run({ [this.rec.inputNames[0]]: tensor });
       const logits = out[this.rec.outputNames[0]];
       const [, T, C] = logits.dims;
-      if (C !== this.dict.length + 2) throw new (0, _chunkTYXXTWPHcjs.KordocError)(`OCR rec \uCD9C\uB825 \uD074\uB798\uC2A4 ${C}\uAC1C\uAC00 \uC0AC\uC804 ${this.dict.length}\uC790 + 2 \uC640 \uB2E4\uB985\uB2C8\uB2E4 \u2014 \uBAA8\uB378\uC5D0 \uB9DE\uB294 \uC0AC\uC804(ocrRecDict)\uC744 \uC9C0\uC815\uD558\uC138\uC694`);
+      if (C !== this.dict.length + 2) throw new (0, _chunkPNSALBXHcjs.KordocError)(`OCR rec \uCD9C\uB825 \uD074\uB798\uC2A4 ${C}\uAC1C\uAC00 \uC0AC\uC804 ${this.dict.length}\uC790 + 2 \uC640 \uB2E4\uB985\uB2C8\uB2E4 \u2014 \uBAA8\uB378\uC5D0 \uB9DE\uB294 \uC0AC\uC804(ocrRecDict)\uC744 \uC9C0\uC815\uD558\uC138\uC694`);
       const data = logits.data;
       idx.forEach((ci, k) => {
         const r = ctcDecode(data.subarray(k * T * C, (k + 1) * T * C), T, C, this.dict);
@@ -1771,13 +1771,13 @@ async function tryImport(name, loader) {
     return await loader();
   } catch (e) {
     throw new Error(
-      `\uB0B4\uC7A5 OCR \uC744 \uC0AC\uC6A9\uD558\uB824\uBA74 optional dependency '${name}' \uC774 \uD544\uC694\uD569\uB2C8\uB2E4. \`npm install ${name}\` \uD6C4 \uB2E4\uC2DC \uC2E4\uD589\uD558\uC138\uC694.${_chunkTYXXTWPHcjs.OPTIONAL_DEP_INSTALL_HINT} \uC6D0\uC778: ${e.message}`
+      `\uB0B4\uC7A5 OCR \uC744 \uC0AC\uC6A9\uD558\uB824\uBA74 optional dependency '${name}' \uC774 \uD544\uC694\uD569\uB2C8\uB2E4. \`npm install ${name}\` \uD6C4 \uB2E4\uC2DC \uC2E4\uD589\uD558\uC138\uC694.${_chunkPNSALBXHcjs.OPTIONAL_DEP_INSTALL_HINT} \uC6D0\uC778: ${e.message}`
     );
   }
 }
 async function verifiedBytes(s) {
   const name = _path.basename.call(void 0, s.file.path);
-  const fail = (why2) => new (0, _chunkTYXXTWPHcjs.KordocError)(`OCR \uBAA8\uB378 \uD30C\uC77C\uC744 \uC77D\uC9C0 \uBABB\uD588\uC2B5\uB2C8\uB2E4: ${name} (${why2})`);
+  const fail = (why2) => new (0, _chunkPNSALBXHcjs.KordocError)(`OCR \uBAA8\uB378 \uD30C\uC77C\uC744 \uC77D\uC9C0 \uBABB\uD588\uC2B5\uB2C8\uB2E4: ${name} (${why2})`);
   const why = (e) => {
     const code = e.code;
     return code === "ENOENT" ? "\uD30C\uC77C\uC774 \uC5C6\uC74C" : code === "EACCES" || code === "EPERM" ? "\uAD8C\uD55C \uC5C6\uC74C" : "\uC77D\uAE30 \uC2E4\uD328";
@@ -1798,13 +1798,13 @@ async function verifiedBytes(s) {
   }
   const actual = _crypto.createHash.call(void 0, "sha256").update(bytes).digest("hex");
   if (actual !== s.file.sha256) {
-    throw new (0, _chunkTYXXTWPHcjs.KordocError)(`OCR \uBAA8\uB378 SHA256 mismatch: ${_path.basename.call(void 0, s.file.path)} \u2014 expected ${s.file.sha256}, got ${actual} (\uB85C\uB4DC \uAC70\uBD80)`);
+    throw new (0, _chunkPNSALBXHcjs.KordocError)(`OCR \uBAA8\uB378 SHA256 mismatch: ${_path.basename.call(void 0, s.file.path)} \u2014 expected ${s.file.sha256}, got ${actual} (\uB85C\uB4DC \uAC70\uBD80)`);
   }
   return bytes;
 }
 var engines = /* @__PURE__ */ new Map();
-function getOcrEngine(models = _chunk4VNIEHWMcjs.DEFAULT_OCR_MODELS) {
-  const key = _chunk4VNIEHWMcjs.ocrSelectionKey.call(void 0, models);
+function getOcrEngine(models = _chunkUNICY546cjs.DEFAULT_OCR_MODELS) {
+  const key = _chunkUNICY546cjs.ocrSelectionKey.call(void 0, models);
   let p = engines.get(key);
   if (!p) {
     p = OcrEngine.create(models).catch((err) => {
@@ -1907,13 +1907,13 @@ function rotateRgba(rgba, width, height, deg) {
 var OCR_RENDER_SCALE = 3;
 var MAX_OCR_PIXELS = 24e6;
 var PAGE_TIMEOUT_MS = 12e4;
-var DEFAULT_BUILTIN_OCR = Object.freeze({ models: _chunk4VNIEHWMcjs.DEFAULT_OCR_MODELS, tuning: DEFAULT_OCR_TUNING });
+var DEFAULT_BUILTIN_OCR = Object.freeze({ models: _chunkUNICY546cjs.DEFAULT_OCR_MODELS, tuning: DEFAULT_OCR_TUNING });
 function builtinOcrConfig(o) {
   if (!o || o.ocrDet === void 0 && o.ocrDetLongSide === void 0 && o.ocrRec === void 0 && o.ocrRecDict === void 0 && !o.ocrLines) return DEFAULT_BUILTIN_OCR;
-  const bad = _chunk5QBTJ4LAcjs.validateOcrOptions.call(void 0, o);
-  if (bad) throw new (0, _chunkTYXXTWPHcjs.KordocError)(bad);
+  const bad = _chunkJFUUFJZWcjs.validateOcrOptions.call(void 0, o);
+  if (bad) throw new (0, _chunkPNSALBXHcjs.KordocError)(bad);
   return {
-    models: _chunk4VNIEHWMcjs.resolveOcrModels.call(void 0, o),
+    models: _chunkUNICY546cjs.resolveOcrModels.call(void 0, o),
     tuning: o.ocrDetLongSide === void 0 ? DEFAULT_OCR_TUNING : { ...DEFAULT_OCR_TUNING, detLongSide: o.ocrDetLongSide },
     lines: o.ocrLines === "all" ? "all" : o.ocrLines ? "kept" : void 0
   };
@@ -1939,7 +1939,7 @@ async function runPdfOcr(buffer, targets, mode, warnings, onProgress, detectTabl
     "@hyzyla/pdfium",
     () => Promise.resolve().then(() => _interopRequireWildcard(require("@hyzyla/pdfium")))
   );
-  if (mode === "builtin") await _chunk4VNIEHWMcjs.ensureOcrModelsFor.call(void 0, config.models, reportModelDownload);
+  if (mode === "builtin") await _chunkUNICY546cjs.ensureOcrModelsFor.call(void 0, config.models, reportModelDownload);
   const engine = mode === "builtin" ? await getOcrEngine(config.models) : null;
   const pdfium = await pdfiumMod.PDFiumLibrary.init();
   const doc = await pdfium.loadDocument(new Uint8Array(buffer));
@@ -2023,7 +2023,7 @@ async function ocrOnePage(page, pageNo, mode, engine, warnings, detectTables, ve
         const x1 = r.x1 * scale - 2, x2 = r.x2 * scale + 2, y1 = (pdfH - r.y2) * scale - 2, y2 = (pdfH - r.y1) * scale + 2;
         const columnRules = _nullishCoalesce(_optionalChain([ruling, 'optionalAccess', _26 => _26.verticals, 'access', _27 => _27.filter, 'call', _28 => _28((s) => Math.min(s.x1, s.x2) >= x1 && Math.max(s.x1, s.x2) <= x2 && Math.min(s.y1, s.y2) >= y1 && Math.max(s.y1, s.y2) <= y2), 'access', _29 => _29.length]), () => ( 0));
         if (columnRules < 2) {
-          for (const b of blocks) if (b.type === "table") _chunk4VNIEHWMcjs.UNRULED_REGION_TABLES.add(b);
+          for (const b of blocks) if (b.type === "table") _chunkUNICY546cjs.UNRULED_REGION_TABLES.add(b);
         }
         return blocks;
       });
@@ -2159,14 +2159,14 @@ function ocrItemsToBlocks(items, pageNumber, pdfW, pdfH, scale, extraLines, dete
       ...xs[i].space ? { hasSpaceBefore: true } : {}
     };
   });
-  return _chunk4VNIEHWMcjs.extractPageBlocksWithLines.call(void 0, norm, pageNumber, opList, pdfW, pdfH, extraLines, detectTables);
+  return _chunkUNICY546cjs.extractPageBlocksWithLines.call(void 0, norm, pageNumber, opList, pdfW, pdfH, extraLines, detectTables);
 }
 async function tryImport2(name, loader) {
   try {
     return await loader();
   } catch (e) {
     throw new Error(
-      `OCR \uC744 \uC0AC\uC6A9\uD558\uB824\uBA74 optional dependency '${name}' \uC774 \uD544\uC694\uD569\uB2C8\uB2E4. \`npm install ${name}\` \uD6C4 \uB2E4\uC2DC \uC2E4\uD589\uD558\uC138\uC694.${_chunkTYXXTWPHcjs.OPTIONAL_DEP_INSTALL_HINT} \uC6D0\uC778: ${e.message}`
+      `OCR \uC744 \uC0AC\uC6A9\uD558\uB824\uBA74 optional dependency '${name}' \uC774 \uD544\uC694\uD569\uB2C8\uB2E4. \`npm install ${name}\` \uD6C4 \uB2E4\uC2DC \uC2E4\uD589\uD558\uC138\uC694.${_chunkPNSALBXHcjs.OPTIONAL_DEP_INSTALL_HINT} \uC6D0\uC778: ${e.message}`
     );
   }
 }
@@ -2212,4 +2212,4 @@ function bgraToRgba(bgra) {
 
 
 exports.getOcrEngine = getOcrEngine; exports.detectRulingLines = detectRulingLines; exports.rulingToPdfLines = rulingToPdfLines; exports.deskewPage = deskewPage; exports.MAX_OCR_PIXELS = MAX_OCR_PIXELS; exports.DEFAULT_BUILTIN_OCR = DEFAULT_BUILTIN_OCR; exports.builtinOcrConfig = builtinOcrConfig; exports.reportModelDownload = reportModelDownload; exports.toOcrLine = toOcrLine; exports.pageLines = pageLines; exports.runPdfOcr = runPdfOcr; exports.ocrOnePage = ocrOnePage; exports.groupOcrLines = groupOcrLines; exports.ocrItemsToBlocks = ocrItemsToBlocks;
-//# sourceMappingURL=chunk-JSI5T77E.cjs.map
+//# sourceMappingURL=chunk-PNX5FYHC.cjs.map