fix(anki): regenerate sentence furigana from the final sentence (#268)

src/core/services/__tests__/stats-server.test.ts

@@ -4321,3 +4321,101 @@ it('TMDB reassignment returns 404 for a missing library entry before fetching de
   assert.equal(fetches, 1);
   assert.deepEqual(assignments, [1]);
 });
+
+for (const outcome of ['success', 'unavailable', 'throws', 'no-field'] as const) {
+  it(`stats word mining updates full sentence furigana without media (${outcome})`, async () => {
+    await withTempDir(async (dir) => {
+      const sourcePath = path.join(dir, 'episode.mkv');
+      fs.writeFileSync(sourcePath, 'fake media');
+      await withFakeAnkiConnect(
+        async (requests, url) => {
+          let calls = 0;
+          const app = createStatsApp(createMockTracker(), {
+            ankiConnectConfig: {
+              url,
+              deck: 'Mining',
+              media: { generateAudio: false, generateImage: false },
+            },
+            addYomitanNote: async () => 12345,
+            generateSentenceFurigana: async (text, word) => {
+              calls++;
+              assert.equal(text, '猫を見た。');
+              assert.equal(word, '猫');
+              if (outcome === 'throws') throw new Error('parser unavailable');
+              return outcome === 'success' ? '<b> 猫[ねこ]</b>を 見[み]た。' : null;
+            },
+          });
+          const response = await app.request('/api/stats/mine-card?mode=word', {
+            method: 'POST',
+            headers: { 'Content-Type': 'application/json' },
+            body: JSON.stringify({
+              sourcePath,
+              startMs: 1000,
+              endMs: 2000,
+              sentence: '猫を見た。',
+              word: '猫',
+            }),
+          });
+          assert.equal(response.status, 200);
+          const fields = requests.find((request) => request.action === 'updateNoteFields')?.params
+            ?.note?.fields;
+          assert.equal(fields?.Sentence, '<b>猫</b>を見た。');
+          assert.equal(
+            fields?.sentencefurigana,
+            outcome === 'no-field'
+              ? undefined
+              : outcome === 'success'
+                ? '<b> 猫[ねこ]</b>を 見[み]た。'
+                : '',
+          );
+          assert.equal(calls, outcome === 'no-field' ? 0 : 1);
+        },
+        {
+          notesInfoFields: {
+            Sentence: { value: '猫' },
+            ...(outcome === 'no-field' ? {} : { sentencefurigana: { value: ' 猫[ねこ]' } }),
+          },
+        },
+      );
+    });
+  });
+}
+
+it('stats word mining skips furigana highlighting when highlightWord is disabled', async () => {
+  await withTempDir(async (dir) => {
+    const sourcePath = path.join(dir, 'episode.mkv');
+    fs.writeFileSync(sourcePath, 'fake media');
+    await withFakeAnkiConnect(
+      async (_requests, url) => {
+        const highlights: Array<string | undefined> = [];
+        const app = createStatsApp(createMockTracker(), {
+          ankiConnectConfig: {
+            url,
+            deck: 'Mining',
+            media: { generateAudio: false, generateImage: false },
+            behavior: { highlightWord: false },
+          },
+          addYomitanNote: async () => 12345,
+          generateSentenceFurigana: async (_text, highlightedText) => {
+            highlights.push(highlightedText);
+            return ' 猫[ねこ]を 見[み]た。';
+          },
+        });
+        const response = await app.request('/api/stats/mine-card?mode=word', {
+          method: 'POST',
+          headers: { 'Content-Type': 'application/json' },
+          body: JSON.stringify({
+            sourcePath,
+            startMs: 1000,
+            endMs: 2000,
+            sentence: '猫を見た。',
+            word: '猫',
+          }),
+        });
+        assert.equal(response.status, 200);
+        assert.deepEqual(highlights, [undefined]);
+      },
+      { notesInfoFields: { Sentence: { value: '猫' }, SentenceFurigana: { value: ' 猫[ねこ]' } } },
+    );
+  });
+});

src/core/services/stats-server.ts

@@ -1,3 +1,4 @@
+import type { StatsMiningRouteOptions } from './stats-server/mining-support';
 import { Hono } from 'hono';
 import http, { type IncomingMessage, type ServerResponse } from 'node:http';
 import { Readable } from 'node:stream';
@@ -132,6 +133,7 @@ export interface StatsServerConfig {
   anilistRateLimiter?: AnilistRateLimiter;
   tmdbClient?: TmdbClient;
   addYomitanNote?: (word: string) => Promise<number | null>;
+  generateSentenceFurigana?: StatsMiningRouteOptions['generateSentenceFurigana'];
   resolveAnkiNoteId?: (noteId: number) => number;
   resolveSentenceSearchHeadwords?: (term: string) => Promise<string[]> | string[];
 }
@@ -155,6 +157,7 @@ export function createStatsApp(
     anilistRateLimiter?: AnilistRateLimiter;
     tmdbClient?: TmdbClient;
     addYomitanNote?: (word: string) => Promise<number | null>;
+    generateSentenceFurigana?: StatsMiningRouteOptions['generateSentenceFurigana'];
     resolveAnkiNoteId?: (noteId: number) => number;
     resolveSentenceSearchHeadwords?: (term: string) => Promise<string[]> | string[];
     createMediaGenerator?: () => StatsServerMediaGenerator;
@@ -191,6 +194,7 @@ export async function startStatsServerWithRuntime(
     anilistRateLimiter: config.anilistRateLimiter,
     tmdbClient: config.tmdbClient,
     addYomitanNote: config.addYomitanNote,
+    generateSentenceFurigana: config.generateSentenceFurigana,
     resolveAnkiNoteId: config.resolveAnkiNoteId,
     resolveSentenceSearchHeadwords: config.resolveSentenceSearchHeadwords,
   });

src/core/services/stats-server/mining-routes.ts

@@ -17,7 +17,6 @@ import {
   getStatsDirectMiningAudioFieldNames,
   getStatsWordMiningAudioFieldName,
   resolveStatsNoteFieldName,
-  shouldUseStatsLapisKikuCardFields,
   statsMiningLogger,
   type StatsMiningRouteOptions,
   type StatsServerNoteInfo,
@@ -228,19 +227,11 @@ export function registerStatsMiningRoutes(app: Hono, options?: StatsMiningRouteO
 
       let imageBuffer = imageResult.status === 'fulfilled' ? imageResult.value : null;
       let noteInfo: StatsServerNoteInfo | null = null;
-      if (
-        audioBuffer ||
-        (syncAnimatedImageToWordAudio && generateImage) ||
-        shouldUseStatsLapisKikuCardFields(ankiConfig)
-      ) {
-        try {
-          const noteInfoResult = (await client.notesInfo([noteId])) as StatsServerNoteInfo[];
-          noteInfo = noteInfoResult[0] ?? null;
-        } catch (err) {
-          if (syncAnimatedImageToWordAudio && generateImage) {
-            errors.push(`image: ${(err as Error).message}`);
-          }
-        }
+      try {
+        const noteInfoResult = (await client.notesInfo([noteId])) as StatsServerNoteInfo[];
+        noteInfo = noteInfoResult[0] ?? null;
+      } catch (error) {
+        errors.push(`note fields: ${error instanceof Error ? error.message : String(error)}`);
       }
       if (syncAnimatedImageToWordAudio && generateImage) {
         try {
@@ -272,6 +263,24 @@ export function registerStatsMiningRoutes(app: Hono, options?: StatsMiningRouteO
       const imageFieldName = ankiConfig.fields?.image ?? 'Picture';
 
       mediaFields[sentenceFieldName] = highlightedSentence;
+      const furiganaFieldName = noteInfo
+        ? resolveStatsNoteFieldName(noteInfo, 'SentenceFurigana')
+        : null;
+      if (furiganaFieldName) {
+        let furigana: string | null = null;
+        try {
+          furigana =
+            (await options?.generateSentenceFurigana?.(
+              sentence,
+              ankiConfig.behavior?.highlightWord === false ? undefined : word,
+            )) ?? null;
+        } catch (error) {
+          statsMiningLogger.warn('Failed to generate sentence furigana:', error);
+        }
+        mediaFields[furiganaFieldName] = furigana ?? '';
+        if (furigana === null)
+          errors.push('furigana: unavailable; using the full sentence without readings');
+      }
       applyStatsWordCardFields(mediaFields, noteInfo, ankiConfig);
 
       if (audioBuffer) {

src/core/services/stats-server/mining-support.ts

@@ -39,6 +39,7 @@ export type StatsMiningRouteOptions = {
     input: RetimedSecondarySubtitleInput,
   ) => Promise<string> | string;
   addYomitanNote?: (word: string) => Promise<number | null>;
+  generateSentenceFurigana?: (text: string, highlightedText?: string) => Promise<string | null>;
   createMediaGenerator?: () => StatsServerMediaGenerator;
   onMiningTiming?: (event: StatsMiningTimingEvent) => void;
   nowMs?: () => number;

src/core/services/tokenizer/sentence-furigana.test.ts

@@ -0,0 +1,46 @@
+import assert from 'node:assert/strict';
+import test from 'node:test';
+import { generateSentenceFurigana } from './sentence-furigana';
+import { createDeps, runInjectedYomitanScript } from './yomitan-scan-test-harness';
+
+test('generates sentence readings through the parser runtime bridge', async () => {
+  const requests: string[] = [];
+  const deps = createDeps((script) =>
+    runInjectedYomitanScript(script, (action, params) => {
+      requests.push(action);
+      if (action === 'optionsGetFull')
+        return { profileCurrent: 0, profiles: [{ options: { scanning: { length: 20 } } }] };
+      assert.equal(action, 'parseText');
+      assert.ok(typeof params === 'object' && params !== null && 'text' in params);
+      assert.equal(params.text, '猫がいる。');
+      return [
+        {
+          source: 'scanning-parser',
+          content: [[{ text: '猫', reading: 'ねこ' }], [{ text: 'がいる。', reading: '' }]],
+        },
+      ];
+    }),
+  );
+  assert.equal(
+    await generateSentenceFurigana('猫がいる。', '猫', deps, { error: assert.fail }),
+    '<b> 猫[ねこ]</b>がいる。',
+  );
+  assert.ok(requests.includes('parseText'));
+});
+
+test(
+  'a stalled parser cannot indefinitely block sentence and media updates',
+  { timeout: 15_000 },
+  async () => {
+    const warnings: string[] = [];
+    const deps = createDeps(() => new Promise<never>(() => {}));
+    assert.equal(
+      await generateSentenceFurigana('猫', undefined, deps, {
+        error: () => undefined,
+        warn: (message) => warnings.push(message),
+      }),
+      null,
+    );
+    assert.equal(warnings.length, 1);
+  },
+);

src/core/services/tokenizer/sentence-furigana.ts

@@ -0,0 +1,28 @@
+import { formatSentenceFurigana } from '../../../anki-integration/sentence-furigana';
+import { requestYomitanParseResults } from './yomitan-parser-runtime';
+
+export async function generateSentenceFurigana(
+  text: string,
+  highlightedText: string | undefined,
+  deps: Parameters<typeof requestYomitanParseResults>[1],
+  logger: Parameters<typeof requestYomitanParseResults>[2],
+): Promise<string | null> {
+  let timer: ReturnType<typeof setTimeout> | undefined;
+  try {
+    const results = await Promise.race([
+      requestYomitanParseResults(text, deps, logger),
+      new Promise<never>((_, reject) => {
+        timer = setTimeout(
+          () => reject(new Error('Sentence furigana generation timed out')),
+          10_000,
+        );
+      }),
+    ]);
+    return formatSentenceFurigana(text, results, highlightedText);
+  } catch (error) {
+    logger.warn?.('Failed to generate sentence furigana:', error);
+    return null;
+  } finally {
+    clearTimeout(timer);
+  }
+}