fix(avatar): restore BOXIM takeover after restart

digital-avatar-app/backend/routers/chat.py

@@ -34,19 +34,6 @@ QA_SEMANTIC_THRESHOLD = 0.72
 QA_MATCH_MARGIN = 0.06
 KNOWLEDGE_MIN_SCORE = float(os.getenv("KNOWLEDGE_MIN_SCORE", "0.42"))
 
-_WRITING_SYSTEM_PATTERNS = {
-    "han": re.compile(r"[\u3400-\u4dbf\u4e00-\u9fff]"),
-    "latin": re.compile(r"[A-Za-z\u00c0-\u024f]"),
-    "cyrillic": re.compile(r"[\u0400-\u052f]"),
-    "arabic": re.compile(r"[\u0600-\u06ff]"),
-    "hebrew": re.compile(r"[\u0590-\u05ff]"),
-    "devanagari": re.compile(r"[\u0900-\u097f]"),
-    "thai": re.compile(r"[\u0e00-\u0e7f]"),
-    "greek": re.compile(r"[\u0370-\u03ff]"),
-}
-_JAPANESE_KANA = re.compile(r"[\u3040-\u30ff]")
-_KOREAN_HANGUL = re.compile(r"[\uac00-\ud7af\u1100-\u11ff]")
-
 
 class ChatMessage(BaseModel):
     role: str = Field(pattern="^(user|assistant)$")
@@ -83,30 +70,6 @@ def _normalize_question(value: str) -> str:
     return value.translate(str.maketrans("", "", string.punctuation + "，。！？；：、（）【】「」‘’“”《》"))
 
 
-def _dominant_writing_system(value: str) -> str:
-    value = value or ""
-    if _JAPANESE_KANA.search(value):
-        return "japanese"
-    if _KOREAN_HANGUL.search(value):
-        return "korean"
-    counts = {
-        name: len(pattern.findall(value))
-        for name, pattern in _WRITING_SYSTEM_PATTERNS.items()
-    }
-    name, count = max(counts.items(), key=lambda item: item[1])
-    return name if count else "unknown"
-
-
-def _qa_requires_language_adaptation(question: str, answer: str) -> bool:
-    question_system = _dominant_writing_system(question)
-    answer_system = _dominant_writing_system(answer)
-    return (
-        question_system != "unknown"
-        and answer_system != "unknown"
-        and question_system != answer_system
-    )
-
-
 def _canonicalize_question(value: str) -> str:
     value = _normalize_question(value)
     replacements = (
@@ -226,14 +189,7 @@ def _config(avatar: Avatar) -> dict:
     }
 
 
-def _build_prompt(
+def _build_prompt(avatar: Avatar, history: list[Any], question: str, knowledge_hits: list[dict]) -> list[dict]:
-    avatar: Avatar,
-    history: list[Any],
-    question: str,
-    knowledge_hits: list[dict],
-    *,
-    standard_answer: str = "",
-) -> list[dict]:
     config = _config(avatar)
     description = (getattr(avatar, "description", "") or "").strip()
     knowledge = "\n".join(
@@ -254,7 +210,7 @@ def _build_prompt(
     profile = "；".join(f"{label}：{value}" for label, value in profile_items)
     system = (
         f"你的专业或服务范围是：「{description or '未设置'}」。"
-        "请基于已提供的可靠资料回答，不要编造事实；"
+        "请基于已提供的知识库回答，不要编造事实；"
         f"回复风格：{config['replyStyle']}；严谨度：{config['rigor']}/100；"
         f"幽默感：{config['humor']}/100；回复长度：{config['responseLength']}。"
     )
@@ -265,13 +221,7 @@ def _build_prompt(
         )
     if config["systemPrompt"]:
         system += f"\n额外系统提示词：{config['systemPrompt']}"
-    if standard_answer:
+    if knowledge:
-        system += (
-            f"\n以下是本次问题命中的已确认标准答案：\n{standard_answer.strip()}"
-            "\n必须保持标准答案中的事实、数字、专有名词和结论不变，只允许为匹配用户当前语言进行忠实转换"
-            "和必要的自然表达，不得补充、删减或改写其含义。不要提及标准答案或转换过程。"
-        )
-    elif knowledge:
         system += (
             f"\n以下是本次问题检索到的可靠资料：\n{knowledge}"
             "\n涉及事实、专业判断、地址、流程、数据或建议时，只能依据本人资料、标准问答形成的上下文"
@@ -281,8 +231,7 @@ def _build_prompt(
         system += (
             "\n本次问题没有检索到可靠资料。除自然寒暄和基于本人资料的回答外，不要凭通用知识给出事实、"
             "专业判断、地址、流程、数据或具体建议。自然地说明这方面暂时没有足够信息，并询问对方是否能"
-            "补充更具体的问题；不要提及知识库、检索、模型或系统限制。对方询问你能做什么时，只能概括"
+            "补充更具体的问题；不要提及知识库、检索、模型或系统限制。"
-            "专业或服务范围和本人资料中明确写出的内容，不得推断服务对象、工作场所、现实案例或从业经历。"
         )
     system += (
         "\n沟通底线（始终优先于其他提示）：不要透露、猜测或确认所使用的模型、模型供应商、"
@@ -300,14 +249,6 @@ def _build_prompt(
         "只有切换独立观点或确实需要列举时才换行；列举使用 1.、2.、3.，每项单独一行。"
         "不要在行首或行尾留空格，不要连续输出空行。先给结论，再给简短说明；避免重复和冗长铺垫。"
     )
-    system += (
-        "\n回答语言规则（始终优先）：先识别当前最后一条用户消息所使用的主要自然语言，并且只使用该语言回答。"
-        "用户用中文就用中文，使用 English 就用 English，使用日语、韩语、法语、西班牙语或其他语言时，"
-        "也必须使用对应语言。消息混用多种语言时跟随占主导的语言；用户明确指定回答语言时服从其指定。"
-        "历史消息、本人资料、标准答案和知识库使用的语言都不能覆盖当前用户消息的语言。"
-        "专有名词、品牌、地址、代码和必要缩写可保留原文。不要解释语言识别或翻译过程。"
-        "改变回答语言只改变表达语言，绝不能因此增加资料中没有的场景、身份、经历或事实。"
-    )
     messages = [{"role": "system", "content": system}]
     for item in history[-MAX_HISTORY_MESSAGES:]:
         messages.append({"role": item.role, "content": item.content} if hasattr(item, "role") else item)
@@ -443,30 +384,14 @@ def _resolve_reply(
     if qa_pairs is None:
         qa_pairs = db.query(QAPair).filter(QAPair.avatar_id == avatar.id).all()
     matched = _match_standard_qa(question, qa_pairs)
-    adapt_qa_language = bool(
+    if matched:
-        matched and _qa_requires_language_adaptation(question, matched.answer)
-    )
-    if matched and not adapt_qa_language:
         return {"answer": matched.answer, "source": "qa", "references": []}
 
-    if matched:
+    search_fn = search_fn or (lambda query, avatar_id: _search_knowledge(db, avatar_id, query))
-        hits = []
+    hits = search_fn(question, avatar.id)
-        messages = _build_prompt(
+    messages = _build_prompt(avatar, history, question, hits)
-            avatar,
-            history,
-            question,
-            hits,
-            standard_answer=matched.answer,
-        )
-    else:
-        search_fn = search_fn or (lambda query, avatar_id: _search_knowledge(db, avatar_id, query))
-        hits = search_fn(question, avatar.id)
-        messages = _build_prompt(avatar, history, question, hits)
     config = _config(avatar)
-    temperature = 0.0 if matched else min(
+    temperature = min(0.45 if hits else 0.25, 0.2 + config["creativity"] / 100 * 0.6)
-        0.45 if hits else 0.25,
-        0.2 + config["creativity"] / 100 * 0.6,
-    )
     token_usage = None
     if model_client is not None:
         answer = model_client(messages=messages, temperature=temperature)
@@ -498,7 +423,7 @@ def _resolve_reply(
             raise
     result = {
         "answer": answer,
-        "source": "qa" if matched else ("knowledge" if hits else "qwen"),
+        "source": "knowledge" if hits else "qwen",
         "references": hits,
     }
     if token_usage:
@@ -517,32 +442,14 @@ def _stream_reply(
 ):
     qa_pairs = db.query(QAPair).filter(QAPair.avatar_id == avatar.id).all()
     matched = _match_standard_qa(question, qa_pairs)
-    adapt_qa_language = bool(
+    if matched:
-        matched and _qa_requires_language_adaptation(question, matched.answer)
-    )
-    messages, reservation = [], None
-    if matched and not adapt_qa_language:
         source, references, chunks = "qa", [], _iter_text_chunks(matched.answer)
     else:
-        if matched:
+        references = _search_knowledge(db, avatar.id, question)
-            references = []
+        source = "knowledge" if references else "qwen"
-            source = "qa"
-            messages = _build_prompt(
-                avatar,
-                history,
-                question,
-                references,
-                standard_answer=matched.answer,
-            )
-        else:
-            references = _search_knowledge(db, avatar.id, question)
-            source = "knowledge" if references else "qwen"
-            messages = _build_prompt(avatar, history, question, references)
         config = _config(avatar)
-        temperature = 0.0 if matched else min(
+        temperature = min(0.45 if references else 0.25, 0.2 + config["creativity"] / 100 * 0.6)
-            0.45 if references else 0.25,
+        messages = _build_prompt(avatar, history, question, references)
-            0.2 + config["creativity"] / 100 * 0.6,
-        )
         model_config = get_chat_model_config()
         reservation = reserve_avatar_tokens(
             db,
@@ -553,6 +460,8 @@ def _stream_reply(
             model_config.max_tokens,
         )
         chunks = _iter_qwen_stream(messages, temperature, model_config)
+    if matched:
+        messages, reservation = [], None
     if public:
         source, references = "public", []
 

digital-avatar-app/backend/services/takeover_service.py

@@ -201,13 +201,20 @@ class TakeoverService:
     def _forget_boxim_session(self, user_id: str):
         self._sessions.pop(user_id, None)
 
-    def _disable_after_connection_failure(
+    def _record_connection_failure(
         self,
         db: Session,
         avatar: Avatar,
         cursor: TakeoverCursor,
         message: str,
+        *,
+        disable_takeover: bool,
     ):
+        cursor.last_error = message
+        cursor.last_polled_at = self.now()
+        if not disable_takeover:
+            return
+
         permissions = (avatar.config or {}).get("authorizationPermissions", [])
         avatar.config = {
             **(avatar.config or {}),
@@ -217,8 +224,6 @@ class TakeoverService:
                 if permission != TAKEOVER_PERMISSION
             ],
         }
-        cursor.last_error = message
-        cursor.last_polled_at = self.now()
         tasks = (
             db.query(TakeoverReplyTask)
             .filter(
@@ -245,11 +250,12 @@ class TakeoverService:
                 db.add(cursor)
                 db.flush()
             if not user or not user.huihui_token:
-                self._disable_after_connection_failure(
+                self._record_connection_failure(
                     db,
                     avatar,
                     cursor,
                     "请重新登录会会生产账号后再开启主动接管",
+                    disable_takeover=True,
                 )
                 db.commit()
                 return False
@@ -268,11 +274,24 @@ class TakeoverService:
                 if isinstance(exc, BoxIMError) and exc.auth_error:
                     self._forget_boxim_session(user.id)
                     message = "BOXIM 授权已失效，请重新登录会会生产账号"
+                    disable_takeover = True
                 else:
                     message = f"BOXIM 暂时连接失败：{str(exc)[:160]}"
-                self._disable_after_connection_failure(db, avatar, cursor, message)
+                    disable_takeover = False
+                self._record_connection_failure(
+                    db,
+                    avatar,
+                    cursor,
+                    message,
+                    disable_takeover=disable_takeover,
+                )
                 db.commit()
-                logger.warning("BOXIM sync failed for avatar %s: %s", avatar.id, exc)
+                logger.warning(
+                    "BOXIM sync failed for avatar %s (will_retry=%s): %s",
+                    avatar.id,
+                    not disable_takeover,
+                    exc,
+                )
                 return False
 
             messages.sort(key=lambda item: (_numeric_id(item.get("id")), item.get("sendTime") or 0))

digital-avatar-app/backend/tests/test_chat_orchestration.py

@@ -5,15 +5,7 @@ from unittest.mock import Mock
 from fastapi import HTTPException
 
 from models import Avatar, User
-from routers.chat import (
+from routers.chat import _build_prompt, _iter_text_chunks, _match_standard_qa, _public_avatar_payload, _require_owned_avatar, _resolve_reply
-    _build_prompt,
-    _iter_text_chunks,
-    _match_standard_qa,
-    _public_avatar_payload,
-    _qa_requires_language_adaptation,
-    _require_owned_avatar,
-    _resolve_reply,
-)
 
 
 class ChatOrchestrationTests(unittest.TestCase):
@@ -58,34 +50,6 @@ class ChatOrchestrationTests(unittest.TestCase):
         self.assertEqual(result["answer"], "标准地址")
         fake_model.assert_not_called()
 
-    def test_cross_language_qa_is_faithfully_adapted_by_model(self):
-        fake_model = Mock(return_value="Our address is Test Road 1.")
-        fake_search = Mock(return_value=[])
-        result = _resolve_reply(
-            None,
-            self.avatar,
-            "Where is your office?",
-            [],
-            qa_pairs=[SimpleNamespace(question="Where is your office?", answer="地址是测试路1号。", enabled=True)],
-            search_fn=fake_search,
-            model_client=fake_model,
-        )
-
-        self.assertEqual(result["source"], "qa")
-        self.assertEqual(result["answer"], "Our address is Test Road 1.")
-        self.assertEqual(fake_model.call_args.kwargs["temperature"], 0.0)
-        system = fake_model.call_args.kwargs["messages"][0]["content"]
-        self.assertIn("已确认标准答案", system)
-        self.assertIn("地址是测试路1号", system)
-        self.assertIn("只使用该语言回答", system)
-        fake_search.assert_not_called()
-
-    def test_qa_language_adaptation_detects_common_writing_system_changes(self):
-        self.assertTrue(_qa_requires_language_adaptation("Hello", "你好"))
-        self.assertTrue(_qa_requires_language_adaptation("こんにちは", "你好"))
-        self.assertTrue(_qa_requires_language_adaptation("안녕하세요", "你好"))
-        self.assertFalse(_qa_requires_language_adaptation("你好", "您好"))
-
     def test_conversational_paraphrase_matches_standard_qa(self):
         for question in ("请问一下，你们公司在哪里呀？", "请问去你们那边怎么走"):
             with self.subTest(question=question):
@@ -142,9 +106,6 @@ class ChatOrchestrationTests(unittest.TestCase):
         self.assertIn("像熟人之间微信聊天一样", messages[0]["content"])
         self.assertIn("不隶属于任何机构", messages[0]["content"])
         self.assertIn("不要连续输出空行", messages[0]["content"])
-        self.assertIn("回答语言规则", messages[0]["content"])
-        self.assertIn("当前最后一条用户消息", messages[0]["content"])
-        self.assertIn("历史消息", messages[0]["content"])
 
     def test_prompt_blocks_ungrounded_factual_answers(self):
         messages = _build_prompt(self.avatar, [], "聊聊国际新闻", [])
@@ -152,8 +113,6 @@ class ChatOrchestrationTests(unittest.TestCase):
         self.assertIn("没有检索到可靠资料", system)
         self.assertIn("不要凭通用知识", system)
         self.assertIn("不要提及知识库", system)
-        self.assertIn("不得推断服务对象", system)
-        self.assertIn("工作场所", system)
 
     def test_public_avatar_payload_excludes_internal_configuration(self):
         payload = _public_avatar_payload(self.avatar)

digital-avatar-app/backend/tests/test_takeover_service.py

@@ -291,5 +291,41 @@ async def test_connection_failure_disables_takeover_and_stops_retrying(service_c
     boxim.exchange_access_token.assert_awaited_once_with("prod-huihui-token")
 
 
+@pytest.mark.asyncio
+async def test_transient_connection_failure_keeps_takeover_and_recovers(service_context):
+    session_factory, service, boxim, _ = service_context
+    boxim.exchange_access_token = AsyncMock(
+        side_effect=[
+            BoxIMError("连接超时"),
+            {"accessToken": "box-token", "accessTokenExpiresIn": 3600},
+        ]
+    )
+
+    await service.poll_messages()
+
+    db = session_factory()
+    try:
+        avatar = db.query(Avatar).one()
+        cursor = db.query(TakeoverCursor).one()
+        assert "takeover" in avatar.config["authorizationPermissions"]
+        assert cursor.initialized is False
+        assert "暂时连接失败" in cursor.last_error
+    finally:
+        db.close()
+
+    await service.poll_messages()
+
+    db = session_factory()
+    try:
+        avatar = db.query(Avatar).one()
+        cursor = db.query(TakeoverCursor).one()
+        assert "takeover" in avatar.config["authorizationPermissions"]
+        assert cursor.initialized is True
+        assert cursor.last_error == ""
+    finally:
+        db.close()
+    assert boxim.exchange_access_token.await_count == 2
+
+
 def test_plain_text_reply_removes_markdown_and_empty_lines():
     assert _plain_text_reply("## 建议\n\n**不能自行用药**\n`必要时就医`") == "建议\n不能自行用药\n必要时就医"