From d47f9e214a84e061ebf3262c18fc88a48cff45c6 Mon Sep 17 00:00:00 2001 From: zk Date: Wed, 13 May 2026 15:41:56 +0800 Subject: [PATCH] =?UTF-8?q?=E8=A1=A5=E5=85=85=E6=97=A5=E5=BF=97?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- app/ai/skill_gap_analyzer/analyzer.py | 6 +++--- app/services/skill_gap_service.py | 13 +++++++++++++ 2 files changed, 16 insertions(+), 3 deletions(-) diff --git a/app/ai/skill_gap_analyzer/analyzer.py b/app/ai/skill_gap_analyzer/analyzer.py index daf1253..4d6ff5f 100644 --- a/app/ai/skill_gap_analyzer/analyzer.py +++ b/app/ai/skill_gap_analyzer/analyzer.py @@ -94,7 +94,7 @@ async def optimize_module(job_title: str, job_description: str, module_data: str _plan_chain = ( ChatPromptTemplate.from_messages([("system", AGENT_PLAN_PROMPT), ("human", "请分析用户指令。")]) - | LLM.DOUBAO_PRO_32K.create(temperature=0) + | LLM.ZM_GPT_5_4_MINI.create(temperature=0) | StrOutputParser() ) @@ -119,7 +119,7 @@ async def plan_edit(job_title: str, job_description: str, resume_json: str, _record_edit_chain = ( ChatPromptTemplate.from_messages([("system", AGENT_MODULE_EDIT_PROMPT), ("human", "请执行修改。")]) - | LLM.ZM_GPT_5_4_MINI.create(temperature=0.3) + | LLM.DOUBAO_PRO_32K.create(temperature=0.3) | StrOutputParser() ) @@ -144,7 +144,7 @@ async def execute_record_edit(job_title: str, job_description: str, instruction: _record_add_chain = ( ChatPromptTemplate.from_messages([("system", AGENT_MODULE_ADD_PROMPT), ("human", "请生成新记录。")]) - | LLM.ZM_GPT_5_4_MINI.create(temperature=0.3) + | LLM.DOUBAO_PRO_32K.create(temperature=0.3) | StrOutputParser() ) diff --git a/app/services/skill_gap_service.py b/app/services/skill_gap_service.py index 1240af0..e5eae19 100644 --- a/app/services/skill_gap_service.py +++ b/app/services/skill_gap_service.py @@ -177,6 +177,8 @@ class SkillGapService: async def ai_edit_customize_resume(self, user_id: int, job_id: int, instruction: str, chat_history: list) -> dict: """AI 对话式编辑定制简历（原子化操作版）""" + log.info(f"AI编辑定制简历开始 [user={user_id}, job={job_id}, instruction={instruction[:50]}]") + t_start = time.monotonic() # 1. 取当前定制简历 cr_data = await customize_resume_store.get(user_id, job_id) if not cr_data: @@ -188,7 +190,9 @@ class SkillGapService: job_desc = f"{job.description or ''}\n{job.requirement or ''}" # 3. 规划 AI（意图识别 + 操作原子化） history_str = json.dumps(chat_history, ensure_ascii=False) if chat_history else "无" + t0 = time.monotonic() plan = await plan_edit(job.title or "", job_desc, resume_json, history_str, instruction) + log.info(f"AI编辑规划完成 ({round(time.monotonic() - t0, 2)}s), plan_action={plan.get('action') if plan else None} , 详情计划:{plan}") if not plan: return {"type": "message", "message": "抱歉，我没有理解你的意思，请再描述一下。"} if plan.get("action") == "chat": @@ -197,14 +201,19 @@ class SkillGapService: operations = plan.get("operations", []) if not operations: return {"type": "message", "message": plan.get("message", "请再描述一下你的需求。")} + log.info(f"AI编辑操作列表: {len(operations)}个操作 {[op.get('type') + '/' + op.get('module', '') for op in operations]}") # 截取最近10条对话历史 recent_history = chat_history[-10:] if len(chat_history) > 10 else chat_history recent_history_str = json.dumps(recent_history, ensure_ascii=False) if recent_history else "无" # 5. 按操作类型分发执行 # 先处理 delete（零 AI 开销） + delete_count = 0 for op in operations: if op.get("type") == "delete": self._apply_delete(cr, op.get("module", ""), op.get("id", "")) + delete_count += 1 + if delete_count: + log.info(f"AI编辑删除操作完成: {delete_count}条记录") # 并发执行 update 和 add ai_tasks = [] for op in operations: @@ -232,8 +241,11 @@ class SkillGapService: )) # 并发执行 if ai_tasks: + log.info(f"AI编辑执行开始: {len(ai_tasks)}个并发任务") + t0 = time.monotonic() coros = [t[3] for t in ai_tasks] results = await asyncio.gather(*coros, return_exceptions=True) + log.info(f"AI编辑执行完成, 耗时={round(time.monotonic() - t0, 2)}s") for (op_type, mod_name, record_id, _), result in zip(ai_tasks, results): if isinstance(result, Exception): log.warning(f"AI编辑[{op_type}/{mod_name}/{record_id}]失败: {result}") @@ -249,6 +261,7 @@ class SkillGapService: # 拼接更新模块标签 updated_modules = list(dict.fromkeys(op.get("module", "") for op in operations)) label = "、".join(_MODULE_LABELS.get(m, m) for m in updated_modules if m) + log.info(f"AI编辑定制简历完成 [user={user_id}, job={job_id}], 总耗时={round(time.monotonic() - t_start, 2)}s, 更新模块={label}") return {"type": "updated", "message": f"完成！已更新：{label or '简历内容'}"} @staticmethod