feat(i18n): persist language preference and thread through chat, memoir, TTS

- Add users.language_preference (Alembic 0018, default zh); capture at signup/SMS
  only; expose on auth and profile APIs
- Lite English prompts for chat and memoir; localized stage labels and agent
  names (Life Echo / 岁月知己)
- Tencent TTS: language-aware synthesis, ModelType=1 for 501004, English chunking
- WebSocket pipeline: emit all AGENT_RESPONSE segments when TTS cancels; INFO logs
  for tts_this_turn and TTS decisions; on-demand TTS logging
- Expo: device language on auth, i18n tiers/agent name, [SPLIT] streaming UX fixes
- Tests for migration, prompts, pipeline, router tts_this_turn, reply segments

Co-authored-by: Cursor <cursoragent@cursor.com>

api/app/features/memoir/story_pipeline_sync.py

@@ -254,7 +254,11 @@ def _slot_snippets_for_narrative(
     return slot_snippets
 
 
-def _placeholder_title(chapter_category: str) -> str:
+def _placeholder_title(chapter_category: str, language: str = "zh") -> str:
+    if language == "en":
+        from app.agents.stage_constants import chapter_category_display
+
+        return chapter_category_display(chapter_category, language="en")
     return CHAPTER_CATEGORIES.get(chapter_category, chapter_category)
 
 
@@ -301,16 +305,19 @@ def _strip_ungrounded_title_segments(
     hay: str,
     *,
     chapter_category: str,
+    language: str = "zh",
 ) -> str:
     """
     按 · / • 分节丢弃含未落地履历短语的小节；全部丢弃则占位。
     """
     if not settings.memoir_title_hay_grounding_strict_phrases_enabled:
-        return (title or "").strip() or _placeholder_title(chapter_category)
+        return (title or "").strip() or _placeholder_title(
+            chapter_category, language=language
+        )
     t = (title or "").strip()
     h = (hay or "").strip()
     if not t:
-        return _placeholder_title(chapter_category)
+        return _placeholder_title(chapter_category, language=language)
     segments = [s.strip() for s in re.split(r"\s*[·•]\s*", t) if s.strip()]
     if not segments:
         segments = [t]
@@ -329,7 +336,7 @@ def _strip_ungrounded_title_segments(
             continue
         kept.append(seg)
     if not kept:
-        return _placeholder_title(chapter_category)
+        return _placeholder_title(chapter_category, language=language)
     if len(kept) == 1:
         return kept[0]
     return " · ".join(kept)
@@ -346,11 +353,12 @@ def _maybe_generate_title(
     llm: Any,
     oral_scope: str = "",
     narrow_profile_for_title: bool = True,
+    language: str = "zh",
 ) -> str:
     """Generate a title only when body is long enough; otherwise return placeholder."""
     body_len = len((md or "").strip())
     if body_len < settings.story_title_min_body_chars:
-        return _placeholder_title(chapter_category)
+        return _placeholder_title(chapter_category, language=language)
     content_excerpt = (md or "").strip()[:300]
     merged_slots = _title_slots_filtered_for_generation(
         slot_snippets, md=md, oral_scope=oral_scope
@@ -366,10 +374,14 @@ def _maybe_generate_title(
         user_profile=profile_for_title,
         birth_year=user_birth_year,
         llm=llm,
+        language=language,
     )
     hay = _title_hay_for_grounding(merged_slots, md, oral_scope)
     return _strip_ungrounded_title_segments(
-        raw_title, hay, chapter_category=chapter_category
+        raw_title,
+        hay,
+        chapter_category=chapter_category,
+        language=language,
     )
 
 
@@ -733,6 +745,7 @@ def _execute_narrative_unit(
     occupation: str = "",
     memoir_correlation_id: str | None = None,
     fidelity_llm: Any | None = None,
+    language: str = "zh",
 ) -> tuple[str | None, bool]:
     """
     Unified narrative unit executor: generate narrative, apply fidelity/safety,
@@ -740,7 +753,9 @@ def _execute_narrative_unit(
     """
     t0 = time.perf_counter()
     oral_norm = (oral_text or "").strip()
-    new_content_input = format_narrative_user_content(oral_text, evidence_text)
+    new_content_input = format_narrative_user_content(
+        oral_text, evidence_text, language=language
+    )
 
     raw_gen = narrative_agent.generate_narrative(
         stage=chapter_category,
@@ -753,6 +768,7 @@ def _execute_narrative_unit(
         background_voice=background_voice,
         occupation=occupation,
         fallback_plain_oral=oral_norm,
+        language=language,
     )
     json_invalid = False
     s0 = (raw_gen or "").strip()
@@ -816,7 +832,7 @@ def _execute_narrative_unit(
         sid_log = target_story_id
         is_append = True
     else:
-        story_title = _placeholder_title(chapter_category)
+        story_title = _placeholder_title(chapter_category, language=language)
         st = create_story_with_version_sync(
             session,
             user_id=user_id,
@@ -905,6 +921,7 @@ def _run_batch_plan_writes(
     occupation: str = "",
     memoir_correlation_id: str | None = None,
     fidelity_llm: Any | None = None,
+    language: str = "zh",
 ) -> set[str]:
     dispatch_ids: set[str] = set()
     for unit in plan.units:
@@ -951,6 +968,7 @@ def _run_batch_plan_writes(
             occupation=occupation,
             memoir_correlation_id=memoir_correlation_id,
             fidelity_llm=fidelity_llm,
+            language=language,
         )
         if sid:
             dispatch_ids.add(sid)
@@ -972,6 +990,7 @@ def run_story_pipeline_for_category_batch(
     memoir_correlation_id: str | None = None,
     llm_fast: Any | None = None,
     memory_evidence: dict | None = None,
+    language: str = "zh",
 ) -> StoryPipelineResult:
     """运行某 chapter_category 的 Phase2 写入管线。
 
@@ -1064,7 +1083,9 @@ def run_story_pipeline_for_category_batch(
         user_id=user_id,
     )
 
-    title = chapter.title if chapter else _placeholder_title(chapter_category)
+    title = chapter.title if chapter else _placeholder_title(
+        chapter_category, language=language
+    )
 
     # 仅同 chapter_category（story.stage）的 Story 可作为 append 候选，避免跨章节链接导致多章内容相同
     all_stories = list_active_stories_for_user_sync(session, user_id)
@@ -1167,6 +1188,7 @@ def run_story_pipeline_for_category_batch(
             occupation=occupation,
             memoir_correlation_id=memoir_correlation_id,
             fidelity_llm=llm_fidelity,
+            language=language,
         )
     else:
         route = single_route
@@ -1215,6 +1237,7 @@ def run_story_pipeline_for_category_batch(
             occupation=occupation,
             memoir_correlation_id=memoir_correlation_id,
             fidelity_llm=llm_fidelity,
+            language=language,
         )
         if sid:
             dispatch_ids.add(sid)