@clazic/kordoc 2.3.2 → 2.3.3

package/dist/index.js

@@ -2425,7 +2425,7 @@ async function resolveOcrProvider(mode, warnings, concurrency, batchSize) {
     return createCliOcrProvider(mode);
   }
   const detected = detectAvailableOcr();
-  if (detected !== "gemini") {
+  if (detected !== "codex") {
     if (detected === "tesseract") {
       warnings?.push({
         message: getTesseractFallbackMessage(),
@@ -2433,7 +2433,7 @@ async function resolveOcrProvider(mode, warnings, concurrency, batchSize) {
       });
     } else {
       warnings?.push({
-        message: `OCR: '${detected}' \uC0AC\uC6A9 \uC911 (gemini CLI\uAC00 \uC5C6\uC5B4 fallback). \uB354 \uB098\uC740 \uD488\uC9C8\uC744 \uC704\uD574 gemini CLI \uC124\uCE58\uB97C \uAD8C\uC7A5\uD569\uB2C8\uB2E4.`,
+        message: `OCR: '${detected}' \uC0AC\uC6A9 \uC911 (codex CLI\uAC00 \uC5C6\uC5B4 fallback). \uB354 \uB098\uC740 \uD488\uC9C8\uC744 \uC704\uD574 codex CLI \uC124\uCE58\uB97C \uAD8C\uC7A5\uD569\uB2C8\uB2E4.`,
         code: "OCR_CLI_FALLBACK"
       });
     }
@@ -2790,7 +2790,7 @@ import JSZip2 from "jszip";
 import { DOMParser } from "@xmldom/xmldom";
 
 // src/utils.ts
-var VERSION = true ? "2.3.2" : "0.0.0-dev";
+var VERSION = true ? "2.3.3" : "0.0.0-dev";
 function toArrayBuffer(buf) {
   if (buf.byteOffset === 0 && buf.byteLength === buf.buffer.byteLength) {
     return buf.buffer;
@@ -6318,15 +6318,26 @@ async function parsePdfDocument(buffer, options) {
         warnings.push({ page: i, message: `\uD398\uC774\uC9C0 ${i} \uD30C\uC2F1 \uC2E4\uD328: ${pageErr instanceof Error ? pageErr.message : "\uC54C \uC218 \uC5C6\uB294 \uC624\uB958"}`, code: "PARTIAL_PARSE" });
       }
     };
-    const sampleCount = Math.min(5, targetPageNums.length);
-    for (let si = 0; si < sampleCount; si++) {
+    const SAMPLE_SIZE = Math.min(10, targetPageNums.length);
+    const sampledIndices = /* @__PURE__ */ new Set();
+    if (targetPageNums.length <= SAMPLE_SIZE) {
+      for (let i = 0; i < targetPageNums.length; i++) sampledIndices.add(i);
+    } else {
+      for (let i = 0; i < SAMPLE_SIZE; i++) {
+        const idx = Math.round(i * (targetPageNums.length - 1) / (SAMPLE_SIZE - 1));
+        sampledIndices.add(idx);
+      }
+    }
+    for (const si of sampledIndices) {
       await parseSinglePage(targetPageNums[si]);
     }
-    const sampleParsed = parsedPages || sampleCount;
+    const sampleParsed = parsedPages || sampledIndices.size;
     const isImageBased = totalChars / Math.max(sampleParsed, 1) < 10;
     if (!isImageBased) {
-      for (let si = sampleCount; si < targetPageNums.length; si++) {
-        await parseSinglePage(targetPageNums[si]);
+      for (let si = 0; si < targetPageNums.length; si++) {
+        if (!sampledIndices.has(si)) {
+          await parseSinglePage(targetPageNums[si]);
+        }
       }
     }
     const parsedPageCount = parsedPages || (pageFilter ? pageFilter.size : effectivePageCount);