@clazic/kordoc 2.4.12 → 2.4.13

package/dist/index.d.cts

@@ -433,6 +433,7 @@ interface UnifiedOcrOptions {
     stageWeights?: Partial<Record<UnifiedStage, number>>;
     timeoutMs?: number;
     maxRetriesPerPage?: number;
+    probeConcurrency?: number;
     logger?: Logger;
     runId?: string;
 }

package/dist/index.d.ts

@@ -433,6 +433,7 @@ interface UnifiedOcrOptions {
     stageWeights?: Partial<Record<UnifiedStage, number>>;
     timeoutMs?: number;
     maxRetriesPerPage?: number;
+    probeConcurrency?: number;
     logger?: Logger;
     runId?: string;
 }

package/dist/index.js

@@ -11373,6 +11373,7 @@ async function runUnifiedOcrPipeline(inputPath, options = {}) {
   const maxRetriesPerPage = options.maxRetriesPerPage ?? 5;
   const dpi = options.dpi ?? 300;
   const modelsInput = options.modelCandidates?.length ? options.modelCandidates : DEFAULT_MODELS;
+  const probeConcurrency = Math.max(1, Math.floor(options.probeConcurrency ?? Math.min(3, modelsInput.length)));
   const modelCache = await loadModelCache(modelCachePath);
   const models = sortModelsByCache(modelsInput, modelCache);
   const modelMaxTokens = { ...DEFAULT_MODEL_MAX_TOKENS, ...options.modelMaxTokens ?? {} };
@@ -11423,11 +11424,10 @@ async function runUnifiedOcrPipeline(inputPath, options = {}) {
     const probeStart = Date.now();
     currentStage = "probe";
     markStageStart("probe", "\uBAA8\uB378 \uC18D\uB3C4 \uD504\uB85C\uBE0C \uC218\uD589 \uC911");
-    logStage("info", "probe", "start", "\uBAA8\uB378 \uC18D\uB3C4 \uD504\uB85C\uBE0C \uC2DC\uC791", { models });
+    logStage("info", "probe", "start", "\uBAA8\uB378 \uC18D\uB3C4 \uD504\uB85C\uBE0C \uC2DC\uC791", { models, probeConcurrency });
     const probeImage = await pickRepresentativeImage(images);
-    const probeResults = [];
-    for (let i = 0; i < models.length; i++) {
-      const model = models[i];
+    let probeDone = 0;
+    const probeResultsByIndex = await mapWithConcurrency(models, probeConcurrency, async (model, index) => {
       const t0 = Date.now();
       try {
         await ocrImageViaNim({
@@ -11442,18 +11442,25 @@ async function runUnifiedOcrPipeline(inputPath, options = {}) {
           logger,
           stage: "probe"
         });
-        probeResults.push({ model, durationMs: Date.now() - t0, success: true });
+        const result = { model, durationMs: Date.now() - t0, success: true };
+        probeDone += 1;
+        markStageProgress("probe", Math.round(probeDone / models.length * 100), probeDone, models.length, `\uBAA8\uB378 \uD504\uB85C\uBE0C ${probeDone}/${models.length}`);
+        logStage("debug", "probe", "progress", "\uBAA8\uB378 \uD504\uB85C\uBE0C \uC9C4\uD589", { index: index + 1, total: models.length, model, result });
+        return result;
       } catch (err) {
-        probeResults.push({
+        const result = {
           model,
           durationMs: Date.now() - t0,
           success: false,
           error: err instanceof Error ? err.message : String(err)
-        });
+        };
+        probeDone += 1;
+        markStageProgress("probe", Math.round(probeDone / models.length * 100), probeDone, models.length, `\uBAA8\uB378 \uD504\uB85C\uBE0C ${probeDone}/${models.length}`);
+        logStage("debug", "probe", "progress", "\uBAA8\uB378 \uD504\uB85C\uBE0C \uC9C4\uD589", { index: index + 1, total: models.length, model, result });
+        return result;
       }
-      markStageProgress("probe", Math.round((i + 1) / models.length * 100), i + 1, models.length, `\uBAA8\uB378 \uD504\uB85C\uBE0C ${i + 1}/${models.length}`);
-      logStage("debug", "probe", "progress", "\uBAA8\uB378 \uD504\uB85C\uBE0C \uC9C4\uD589", { index: i + 1, total: models.length, model, result: probeResults.at(-1) });
-    }
+    });
+    const probeResults = probeResultsByIndex;
     const selectedModel = chooseFastestModel(probeResults);
     if (!selectedModel) throw new UnifiedOcrError("PROBE_FAILED", "probe", "\uC18D\uB3C4 \uD504\uB85C\uBE0C \uC2E4\uD328: \uC0AC\uC6A9 \uAC00\uB2A5\uD55C OCR \uBAA8\uB378\uC774 \uC5C6\uC2B5\uB2C8\uB2E4.");
     const fallbackModelOrder = probeResults.filter((r) => r.success).sort((a, b) => a.durationMs - b.durationMs).map((r) => r.model);
@@ -11668,6 +11675,21 @@ async function pickRepresentativeImage(images) {
   use.sort((a, b) => a.size - b.size);
   return use[Math.floor(use.length / 2)].path;
 }
+async function mapWithConcurrency(items, concurrency, mapper) {
+  const results = new Array(items.length);
+  let nextIndex = 0;
+  async function worker() {
+    while (true) {
+      const idx = nextIndex;
+      if (idx >= items.length) return;
+      nextIndex += 1;
+      results[idx] = await mapper(items[idx], idx);
+    }
+  }
+  const workerCount = Math.max(1, Math.min(concurrency, items.length));
+  await Promise.all(Array.from({ length: workerCount }, () => worker()));
+  return results;
+}
 function chooseFastestModel(results) {
   const ok = results.filter((r) => r.success).sort((a, b) => a.durationMs - b.durationMs);
   return ok[0]?.model ?? null;