refactor(compute-core): rename pdf and whisper modules to clean capability layout

Dockerfile

@@ -58,14 +58,14 @@ COPY --from=app-builder /app/THIRD_PARTY_LICENSES /licenses
 # Include SeaweedFS license text for the copied weed binary.
 COPY --from=seaweedfs-builder /tmp/SeaweedFS-LICENSE.txt /licenses/SeaweedFS-LICENSE.txt
 # Include static model notices for runtime-downloaded assets.
-COPY --from=app-builder /app/compute/core/src/pdf-layout/model/LICENSE.txt /licenses/pp-doclayoutv3-LICENSE.txt
+COPY --from=app-builder /app/compute/core/src/pdf/assets/LICENSE.txt /licenses/pp-doclayoutv3-LICENSE.txt
 
 # Copy seaweedfs weed binary for optional embedded local S3.
 COPY --from=seaweedfs-builder /tmp/weed /usr/local/bin/weed
 RUN chmod +x /usr/local/bin/weed
 
 # Include OpenAI Whisper license text for runtime-downloaded ONNX artifacts.
-COPY --from=app-builder /app/compute/core/src/whisper/model/LICENSE.txt /licenses/openai-whisper-LICENSE.txt
+COPY --from=app-builder /app/compute/core/src/whisper/assets/LICENSE.txt /licenses/openai-whisper-LICENSE.txt
 
 # Expose the port the app runs on
 EXPOSE 3003

compute/core/src/index.ts

@@ -13,7 +13,7 @@ export {
   type ComputeOperationKind,
   type IdleTimeoutAndHardCapInput,
 } from './config/timeout';
-export { renderPage } from './pdf-layout/renderPage';
+export { renderPage } from './pdf/render';
-export { mergeTextWithRegions } from './pdf-layout/mergeTextWithRegions';
+export { mergeTextWithRegions } from './pdf/merge';
-export { stitchCrossPageBlocks } from './pdf-layout/stitchCrossPageBlocks';
+export { stitchCrossPageBlocks } from './pdf/stitch';
-export { normalizeTextItemsForLayout } from './pdf-layout/normalizeTextItemsForLayout';
+export { normalizeTextItemsForLayout } from './pdf/normalize-text';

compute/core/src/local-runtime.ts

@@ -1,7 +1,7 @@
-import { ensureWhisperModel } from './whisper/ensureModel';
+import { ensureWhisperModel } from './whisper/model';
-import { alignAudioWithText } from './whisper/alignment';
+import { alignAudioWithText } from './whisper/align';
-import { ensureModel as ensurePdfLayoutModel } from './pdf-layout/ensureModel';
+import { ensureModel as ensurePdfLayoutModel } from './pdf/model';
-import { parsePdf } from './pdf-layout/parsePdf';
+import { parsePdf } from './pdf/parse';
 
 export async function ensureComputeModels(): Promise<void> {
   await Promise.all([ensureWhisperModel(), ensurePdfLayoutModel()]);

compute/core/src/pdf-layout/index.ts

@@ -1,3 +0,0 @@
-export { renderPage } from './renderPage';
-export { mergeTextWithRegions } from './mergeTextWithRegions';
-export { stitchCrossPageBlocks } from './stitchCrossPageBlocks';

compute/core/src/pdf/index.ts

@@ -0,0 +1,3 @@
+export { renderPage } from './render';
+export { mergeTextWithRegions } from './merge';
+export { stitchCrossPageBlocks } from './stitch';

compute/core/src/pdf/model.ts

@@ -3,7 +3,7 @@ import { fileURLToPath } from 'url';
 import { createHash } from 'crypto';
 import { access, mkdir, rename, writeFile, readFile, unlink, copyFile } from 'fs/promises';
 import { DOCSTORE_DIR } from '../platform/docstore';
-import manifest from './model/manifest.json';
+import manifest from './assets/manifest.json';
 
 const DEFAULT_MODEL_BASE_URL = 'https://huggingface.co/Bei0001/PP-DocLayoutV3-ONNX/resolve/main';
 const PDF_LAYOUT_MODEL_BASE_URL_ENV = 'PDF_LAYOUT_MODEL_BASE_URL';
@@ -14,7 +14,7 @@ export const MODEL_DATA_PATH = path.join(MODEL_DIR, 'PP-DocLayoutV3.onnx.data');
 export const MODEL_CONFIG_PATH = path.join(MODEL_DIR, 'pp-doclayoutv3.config.json');
 export const MODEL_PREPROCESSOR_PATH = path.join(MODEL_DIR, 'pp-doclayoutv3.preprocessor_config.json');
 const LICENSE_PATH = path.join(MODEL_DIR, 'pp-doclayoutv3.LICENSE.txt');
-const STATIC_LICENSE_PATH = path.join(MODULE_DIR, 'model', 'LICENSE.txt');
+const STATIC_LICENSE_PATH = path.join(MODULE_DIR, 'assets', 'LICENSE.txt');
 
 let inflight: Promise<string> | null = null;
 

compute/core/src/pdf/parse.ts

@@ -1,12 +1,12 @@
 import path from 'path';
 import type { TextItem } from 'pdfjs-dist/types/src/display/api';
 import type { ParsedPdfDocument, ParsedPdfPage } from '../types/parsed-pdf';
-import { ensureModel } from './ensureModel';
+import { ensureModel } from './model';
 import { runLayoutModel } from './runLayoutModel';
-import { mergeTextWithRegions } from './mergeTextWithRegions';
+import { mergeTextWithRegions } from './merge';
-import { stitchCrossPageBlocks } from './stitchCrossPageBlocks';
+import { stitchCrossPageBlocks } from './stitch';
-import { renderPage } from './renderPage';
+import { renderPage } from './render';
-import { normalizeTextItemsForLayout } from './normalizeTextItemsForLayout';
+import { normalizeTextItemsForLayout } from './normalize-text';
 
 interface ParsePdfInput {
   documentId: string;

compute/core/src/pdf/runLayoutModel.ts

@@ -1,7 +1,7 @@
 import * as ort from 'onnxruntime-node';
 import { readFile } from 'fs/promises';
 import type { LayoutRegion, PdfTextItem } from './types';
-import { ensureModel, MODEL_CONFIG_PATH, MODEL_PREPROCESSOR_PATH } from './ensureModel';
+import { ensureModel, MODEL_CONFIG_PATH, MODEL_PREPROCESSOR_PATH } from './model';
 import { getOnnxThreadsPerJob } from '../config/cpu-budget';
 
 interface RunLayoutInput {

compute/core/src/whisper/align.ts

@@ -14,7 +14,7 @@ import { getComputeTimeoutConfig } from '../config/timeout';
 import {
   mapWordsToSentenceOffsets,
   type WhisperWord,
-} from './alignment-mapping';
+} from './alignment-map';
 import { buildGoertzelCoefficients, goertzelPower } from './spectral';
 import {
   buildWordsFromTimestampedTokens,
@@ -29,7 +29,7 @@ import {
   WHISPER_ENCODER_MODEL_PATH,
   WHISPER_DECODER_MERGED_MODEL_PATH,
   WHISPER_DECODER_WITH_PAST_MODEL_PATH,
-} from './ensureModel';
+} from './model';
 
 interface WhisperAlignmentOptions {
   lang?: string;
@@ -114,7 +114,7 @@ const hannWindow = buildHannWindow(N_FFT);
 const goertzelCoefficients = buildGoertzelCoefficients(MEL_FILTER_BINS, N_FFT);
 
 const MODULE_DIR = dirname(fileURLToPath(import.meta.url));
-const MEL_FILTERS_NPZ_PATH = join(MODULE_DIR, 'model', 'mel_filters.npz');
+const MEL_FILTERS_NPZ_PATH = join(MODULE_DIR, 'assets', 'mel_filters.npz');
 
 function buildHannWindow(length: number): Float32Array {
   const window = new Float32Array(length);

compute/core/src/whisper/index.ts

@@ -2,7 +2,7 @@ export {
   alignAudioWithText,
   makeWhisperCacheKey,
   type WhisperRequestBody,
-} from './alignment';
+} from './align';
 
 export {
   ensureWhisperModel,
@@ -11,8 +11,8 @@ export {
   type WhisperArtifactSpec,
   type WhisperStaticArtifactSpec,
   type WhisperFetch,
-} from './ensureModel';
+} from './model';
 
-export { mapWordsToSentenceOffsets, type WhisperWord } from './alignment-mapping';
+export { mapWordsToSentenceOffsets, type WhisperWord } from './alignment-map';
 export { buildGoertzelCoefficients, goertzelPower } from './spectral';
 export { buildWordsFromTimestampedTokens, extractTokenStartTimestamps } from './token-timestamps';

compute/core/src/whisper/model.ts

@@ -3,11 +3,11 @@ import { fileURLToPath } from 'url';
 import { createHash } from 'crypto';
 import { access, copyFile, mkdir, readFile, rename, unlink, writeFile } from 'fs/promises';
 import { DOCSTORE_DIR } from '../platform/docstore';
-import manifest from './model/manifest.json';
+import manifest from './assets/manifest.json';
 
 const MODULE_DIR = path.dirname(fileURLToPath(import.meta.url));
 const MODEL_DIR = path.join(DOCSTORE_DIR, 'model', 'whisper-base_timestamped');
-const STATIC_LICENSE_PATH = path.join(MODULE_DIR, 'model', 'LICENSE.txt');
+const STATIC_LICENSE_PATH = path.join(MODULE_DIR, 'assets', 'LICENSE.txt');
 
 export const WHISPER_CONFIG_PATH = path.join(MODEL_DIR, 'config.json');
 export const WHISPER_GENERATION_CONFIG_PATH = path.join(MODEL_DIR, 'generation_config.json');

tests/unit/whisper-alignment-mapping.spec.ts

@@ -1,7 +1,7 @@
 import { test, expect } from '@playwright/test';
 import {
   mapWordsToSentenceOffsets,
-} from '../../compute/core/src/whisper/alignment-mapping';
+} from '../../compute/core/src/whisper/alignment-map';
 
 test.describe('whisper alignment mapping', () => {
   test('maps words to sentence offsets with punctuation and repeated spaces', () => {

tests/unit/whisper-alignment-smoke.spec.ts

@@ -1,7 +1,7 @@
 import { test, expect } from '@playwright/test';
 import { readFile } from 'fs/promises';
 import path from 'path';
-import { alignAudioWithText } from '../../compute/core/src/whisper/alignment';
+import { alignAudioWithText } from '../../compute/core/src/whisper/align';
 
 test.describe('whisper alignment smoke', () => {
   test('runs ONNX alignment end-to-end without decoder reshape errors', async () => {

tests/unit/whisper-ensure-model.spec.ts

@@ -6,7 +6,7 @@ import path from 'path';
 import {
   createSingleflightRunner,
   ensureWhisperArtifacts,
-} from '../../compute/core/src/whisper/ensureModel';
+} from '../../compute/core/src/whisper/model';
 
 function sha256(bytes: Uint8Array): string {
   return createHash('sha256').update(bytes).digest('hex');