diff --git a/.gitignore b/.gitignore index e69de29..3ca7834 100644 --- a/.gitignore +++ b/.gitignore @@ -0,0 +1,10 @@ +# Python 字节码缓存 +__pycache__/ +*.py[cod] +*$py.class + +# 项目特定的存储/缓存文件夹 +.storage/ + +# 环境变量文件（通常包含敏感信息） +.env \ No newline at end of file diff --git a/.storage/config.json b/.storage/config.json index bb995f7..c77fe55 100644 --- a/.storage/config.json +++ b/.storage/config.json @@ -1,6 +1,6 @@ { - "provider": "DeepSeek", - "api_key": "sk-ca812c913baa474182f6d4e83e078302", - "base_url": "https://api.deepseek.com", + "provider": "AIHubMix", + "api_key": "sk-yd8Tik0nFW5emKYcBdFc433b7c8b4dC182848f76819bBe73", + "base_url": "https://aihubmix.com/v1", "language": "Chinese" } \ No newline at end of file diff --git a/README.md b/README.md index 6ab56ac..5ff5ea6 100644 --- a/README.md +++ b/README.md @@ -1,76 +1,91 @@ -# Multi-Agent Council & Debate Workshop (V4) +# 🍎 智能决策工作坊 (Multi-Agent Council V4) -一个极简而强大的多智能体（Multi-Agent）决策辅助系统。 -**V4 版本**将传统的 "线性研究" 进化为 **"多模型智囊团 (Council V4)"**，支持多轮对话讨论、动态专家组建、以及多 API 平台接入。 +AI驱动的多智能体决策分析系统 - 基于多模型智囊团 -## ✨ 核心功能 (V4 Update) +## ✨ 核心功能 -### 1. 🧪 Multi-Model Council V4 (多模型智囊团) -摒弃了单一的"规划-执行"模式，现在的系统是一个真正的**圆桌会议**： -* **多轮对话讨论**: 专家不再是各自为战，而是像真实会议一样进行多轮（Round-Robin）对话，互相批判、补充观点。 -* **动态专家组建**: 你可以自定义 **2-5 位** 不同的专家（如 CEO, CTO, 法务）。 -* **自定义模型分配**: 为每个专家指定最擅长的模型（例如：让 DeepSeek-Coder 担任技术专家，让 GPT-4o 担任产品专家）。 -* **最终决策合成**: 讨论结束后，最后一位专家（Synthesizer）会综合全场观点，生成最终决策方案，并绘制 **Mermaid 路线图**。 +### 🧪 Multi-Model Council V4 (智囊团模式) +- **多轮对话讨论**: 专家像真实会议一样进行多轮对话，互相批判、补充观点 +- **动态专家组建**: 自定义 2-5 位专家，为每位指定最擅长的模型 +- **🪄 智能专家生成**: AI 根据主题自动推荐最合适的专家角色 +- **最终决策合成**: 最后一位专家综合全场观点，生成方案并绘制 Mermaid 路线图 -### 2. 🎭 Debate Workshop (辩论工作坊) -经典的辩论模式，让 AI 扮演不同立场的角色（如正方、反方、评审），通过激烈的辩论帮助你厘清复杂决策的利弊。 +### 🎯 内置决策场景 +系统预置 4 大典型决策场景，每个场景都配置了专业的典型问题： -### 3. 🌐 Multi-Provider Support (多平台支持) -不再局限于单一平台，系统原生支持多种 API 源，随心切换： -* **DeepSeek Official**: 直接连接 `api.deepseek.com` -* **SiliconFlow (硅基流动)**: 连接 `api.siliconflow.cn` -* **AIHubMix**: 聚合平台 -* **OpenAI / Custom**: 支持标准 OpenAI 接口或本地 vLLM/Ollama +| 场景 | 描述 | +|------|------| +| 🚀 新产品发布评审 | 评估产品可行性、市场潜力和实施计划 | +| 💰 投资审批决策 | 分析投资项目的 ROI、风险和战略价值 | +| 🤝 合作伙伴评估 | 评估合作伙伴的匹配度和合作价值 | +| 📦 供应商评估 | 对比分析供应商的综合能力 | + +### 🎭 Debate Workshop (辩论工作坊) +让 AI 扮演不同立场角色，通过辩论帮助厘清复杂决策的利弊 + +### 💬 用户反馈 +内置用户反馈系统，收集功能建议和使用体验 + +### 🌐 多平台支持 +- **DeepSeek**: V3, R1, Coder +- **OpenAI**: GPT-4o, GPT-4o-mini +- **Anthropic**: Claude 3.5 Sonnet +- **Google**: Gemini 1.5/2.0 +- **SiliconFlow / AIHubMix / Deepseek** --- ## 🛠️ 安装 ```bash -# 1. 克隆项目 +# 克隆项目 git clone https://github.com/HomoDeusss/multi-agent.git cd multi-agent -# 2. 安装依赖 -pip install -r requirements.txt +# 初始化 uv 项目（如首次使用） +uv init + +# 安装依赖 +uv add streamlit openai anthropic python-dotenv + +# 或者同步现有依赖 +uv sync ``` ## 🚀 快速开始 -### 1. 启动应用 ```bash -streamlit run app.py +uv run streamlit run app.py ``` -### 2. 配置 API (V4 新特性) -无需手动修改 `.env` 文件（可选），直接在 Web 界面侧边栏配置： -1. 在侧边栏选择 **"API Provider"** (例如 `DeepSeek` 或 `SiliconFlow`)。 -2. 输入对应的 **API Key**。 -3. 系统会自动配置好 Base URL。 +### 使用步骤 -### 3. 使用 Council V4 模式 -1. 选择 **"Deep Research" (现已升级为 Council V4)**。 -2. **设定专家**: 选择专家人数（例如 3 人），并为每位专家命名并指定模型。 - * *Tip: 建议最后一位专家选一个逻辑能力强的模型（如 Claude 3.5 Sonnet）作为决策者。* -3. **设定轮数**: 选择讨论轮数（建议 2-3 轮）。 -4. 输入议题，点击开始。观察专家们如何互相对话！ - -### 4. 使用 Debate 模式 -1. 切换到 **"Debate Workshop"**。 -2. 输入议题（如“是否应该全职做独立开发？”）。 -3. 选择参与辩论的角色。 -4. 点击开始，观看唇枪舌战。 +1. **配置 API**: 在侧边栏选择 Provider 并输入 API Key +2. **选择场景**: 点击预置的决策场景或自定义主题 +3. **生成专家**: 点击 "🪄 根据主题自动生成专家" 或手动配置 +4. **开始决策**: 观察专家们如何互相对话，生成综合方案 --- -## 🤖 支持的模型 (V4 Expanded) +## 📁 项目结构 -系统内置了最新的模型配置，支持在界面直接选择： -* **DeepSeek**: V3 (`deepseek-chat`), R1 (`deepseek-reasoner`), Coder V2 -* **OpenAI**: GPT-4o, GPT-4o-mini -* **Anthropic**: Claude 3.5 Sonnet, Claude 3 Opus -* **Google**: Gemini 1.5 Pro/Flash -* **Meta/Alibaba**: Llama 3.3, Qwen 2.5 +``` +multi_agent_workshop/ +├── app.py # Streamlit 主应用 +├── config.py # 配置文件 +├── agents/ # Agent 定义 +│ ├── agent_profiles.py # 预设角色配置 +│ ├── base_agent.py # 基础 Agent 类 +│ └── research_agent.py # 研究型 Agent +├── orchestrator/ # 编排器 +│ ├── debate_manager.py # 辩论管理 +│ └── research_manager.py # 智囊团管理 +├── utils/ +│ ├── llm_client.py # LLM 客户端封装 +│ ├── storage.py # 存储管理 +│ └── auto_agent_generator.py # 智能专家生成 +└── report/ # 报告生成 +``` ## 📝 License [MIT License](LICENSE) \ No newline at end of file diff --git a/__pycache__/app.cpython-313.pyc b/__pycache__/app.cpython-313.pyc index 45a71ff..e5965ce 100644 Binary files a/__pycache__/app.cpython-313.pyc and b/__pycache__/app.cpython-313.pyc differ diff --git a/__pycache__/config.cpython-313.pyc b/__pycache__/config.cpython-313.pyc index 1735124..eb6d5d9 100644 Binary files a/__pycache__/config.cpython-313.pyc and b/__pycache__/config.cpython-313.pyc differ diff --git a/app.py b/app.py index e424796..4bc1a62 100644 --- a/app.py +++ b/app.py @@ -17,6 +17,7 @@ from report import ReportGenerator from report import ReportGenerator from utils import LLMClient from utils.storage import StorageManager +from utils.auto_agent_generator import generate_experts_for_topic import config # ==================== 页面配置 ==================== @@ -30,37 +31,146 @@ st.set_page_config( # ==================== 样式 ==================== st.markdown(""" """, unsafe_allow_html=True) @@ -122,6 +232,8 @@ if "research_output" not in st.session_state: st.session_state.research_output = "" # Final report if "research_steps_output" not in st.session_state: st.session_state.research_steps_output = [] # List of step results +if "generated_experts" not in st.session_state: + st.session_state.generated_experts = None # Auto-generated expert configs # ==================== 侧边栏：配置 ==================== @@ -210,7 +322,7 @@ with st.sidebar: save_current_config() if not api_key: - st.warning("请配置 API Key 以继续") + st.warning("⚠️ 请配置 API Key 以启用 AI 功能 (仍可查看历史档案)") # Output Language Selection lang_options = config.SUPPORTED_LANGUAGES @@ -251,8 +363,8 @@ with st.sidebar: # 模式选择 mode = st.radio( "📊 选择模式", - ["Council V4 (Deep Research)", "Debate Workshop", "📜 History Archives"], - index=0 if st.session_state.mode == "Deep Research" else (1 if st.session_state.mode == "Debate Workshop" else 2) + ["Council V4 (Deep Research)", "Debate Workshop", "📜 History Archives", "💬 用户反馈"], + index=0 if st.session_state.mode == "Deep Research" else (1 if st.session_state.mode == "Debate Workshop" else (2 if st.session_state.mode == "History Archives" else 3)) ) # Map selection back to internal mode string @@ -260,8 +372,10 @@ with st.sidebar: st.session_state.mode = "Deep Research" elif mode == "Debate Workshop": st.session_state.mode = "Debate Workshop" - else: + elif mode == "📜 History Archives": st.session_state.mode = "History Archives" + else: + st.session_state.mode = "Feedback" st.divider() @@ -341,18 +455,167 @@ if st.session_state.get("bg_image_data_url"): # ==================== 主界面逻辑 ==================== if st.session_state.mode == "Deep Research": - st.title("🧪 Multi-Model Council V4") - st.markdown("*多模型智囊团：自定义 N 个专家进行多轮对话讨论，最后由最后一位专家决策*") + # ==================== 主标题区域 ==================== + st.markdown(""" +
+

🍎 智能决策工作坊

+

AI驱动的多智能体决策分析系统 - 基于多模型智囊团

+
+ """, unsafe_allow_html=True) + + # 状态指示器和语言选择 + col_status, col_lang = st.columns([2, 1]) + with col_status: + if api_key: + st.markdown(""" +
+
+ ✓ 已连接到服务器 +
+ """, unsafe_allow_html=True) + else: + st.warning("⚠️ 请在侧边栏配置 API Key") + + with col_lang: + st.markdown(f"**语言/Language:** {output_language}") + + st.divider() + + # ==================== 开始决策按钮 ==================== + st.markdown(""" +
+

🚀 开始决策

+

选择场景或自定义主题，开始多专家协作分析

+
+ """, unsafe_allow_html=True) + + st.divider() + + # ==================== 支持的决策场景 ==================== + st.markdown(""" +
+

📋 支持的决策场景

+

系统支持以下决策场景，每个场景都配置了专业的AI专家团队

+
+ """, unsafe_allow_html=True) + + # Decision scenario templates with typical questions + DECISION_SCENARIOS = { + "🚀 新产品发布评审": { + "topic": "新产品发布评审：评估产品功能完备性、市场准备度、发布时机和潜在风险", + "description": "评估新产品概念的可行性、市场潜力和实施计划", + "example": "我们计划在下个季度发布AI助手功能，需要评估技术准备度、市场时机和竞争态势", + "questions": [ + "这个产品的核心价值主张是什么？", + "目标用户群体是谁？需求是否真实存在？", + "技术实现难度如何？团队是否具备能力？", + "竞争对手有类似产品吗？我们的差异化在哪？" + ] + }, + "💰 投资审批决策": { + "topic": "投资审批决策：评估投资项目的财务回报、战略价值、风险因素和执行可行性", + "description": "分析投资项目的ROI、风险和战略价值", + "example": "公司考虑投资1000万用于数据中台建设，需要评估ROI、技术风险和业务价值", + "questions": [ + "预期投资回报率(ROI)是多少？", + "投资回收期需要多长时间？", + "主要风险因素有哪些？如何缓解？", + "是否有更优的替代方案？" + ] + }, + "🤝 合作伙伴评估": { + "topic": "合作伙伴评估：分析潜在合作方的能力、信誉、战略协同和合作风险", + "description": "评估潜在合作伙伴的匹配度和合作价值", + "example": "评估与XX公司建立战略合作的可行性，包括技术互补性、市场协同和风险", + "questions": [ + "合作方的核心能力是什么？", + "双方资源如何互补？", + "合作的战略协同效应有多大？", + "合作失败的风险和退出机制是什么？" + ] + }, + "📦 供应商评估": { + "topic": "供应商评估：评估供应商的质量、成本、交付能力、稳定性和合作风险", + "description": "对比分析供应商的综合能力", + "example": "评估更换核心零部件供应商的利弊，包括成本对比、质量风险和切换成本", + "questions": [ + "供应商的质量控制体系如何？", + "价格竞争力与行业均值对比？", + "交付能力和响应速度如何？", + "供应商的财务稳定性如何？" + ] + } + } + + # Display scenario cards with typical questions + for scenario_name, scenario_data in DECISION_SCENARIOS.items(): + st.markdown(f""" +
+

{scenario_name}

+

{scenario_data['description']}

+
+ 典型问题： + +
+
+ """, unsafe_allow_html=True) + + if st.button(f"使用此场景", key=f"use_{scenario_name}", use_container_width=True): + st.session_state.selected_scenario = scenario_data + st.session_state.prefill_topic = scenario_data['topic'] + st.rerun() + + st.divider() + + # Get prefilled topic if available + prefill_topic = st.session_state.get("prefill_topic", "") + if st.session_state.get("selected_scenario"): + prefill_topic = prefill_topic or st.session_state.selected_scenario.get("topic", "") col1, col2 = st.columns([3, 1]) with col1: - research_topic = st.text_area("研究/决策主题", placeholder="请输入你想深入研究或决策的主题...", height=100) + research_topic = st.text_area("研究/决策主题", value=prefill_topic, placeholder="请输入你想深入研究或决策的主题...", height=100) with col2: max_rounds = st.number_input("讨论轮数", min_value=1, max_value=5, value=2, help="专家们进行对话的轮数") # Expert Configuration st.subheader("👥 专家配置") - num_experts = st.number_input("专家数量", min_value=2, max_value=5, value=3) + + # Auto-generate experts row + col_num, col_auto = st.columns([2, 3]) + with col_num: + num_experts = st.number_input("专家数量", min_value=2, max_value=5, value=3) + with col_auto: + st.write("") # Spacing + auto_gen_btn = st.button( + "🪄 根据主题自动生成专家", + disabled=(not research_topic or not api_key), + help="AI 将根据您的主题自动推荐合适的专家角色" + ) + + # Handle auto-generation + if auto_gen_btn and research_topic and api_key: + with st.spinner("🤖 AI 正在分析主题并生成专家配置..."): + try: + temp_client = LLMClient( + provider=provider_id, + api_key=api_key, + base_url=base_url, + model="gpt-4o-mini" # Use fast model for generation + ) + generated = generate_experts_for_topic( + topic=research_topic, + num_experts=num_experts, + llm_client=temp_client, + language=output_language + ) + st.session_state.generated_experts = generated + st.success(f"✅ 已生成 {len(generated)} 位专家配置！") + st.rerun() + except Exception as e: + st.error(f"生成失败: {e}") experts_config = [] cols = st.columns(num_experts) @@ -360,11 +623,20 @@ if st.session_state.mode == "Deep Research": for i in range(num_experts): with cols[i]: default_model_key = list(AVAILABLE_MODELS.keys())[i % len(AVAILABLE_MODELS)] - st.markdown(f"**Expert {i+1}**") - # Default names - default_name = f"Expert {i+1}" - if i == num_experts - 1: - default_name = f"Expert {i+1} (Synthesizer)" + + # Use generated expert name if available + if st.session_state.generated_experts and i < len(st.session_state.generated_experts): + gen_expert = st.session_state.generated_experts[i] + default_name = gen_expert.get("name", f"Expert {i+1}") + perspective = gen_expert.get("perspective", "") + st.markdown(f"**{default_name}**") + if perspective: + st.caption(f"_{perspective}_") + else: + default_name = f"Expert {i+1}" + if i == num_experts - 1: + default_name = f"Expert {i+1} (Synthesizer)" + st.markdown(f"**Expert {i+1}**") expert_name = st.text_input(f"名称 #{i+1}", value=default_name, key=f"expert_name_{i}") expert_model = st.selectbox(f"模型 #{i+1}", options=list(AVAILABLE_MODELS.keys()), index=list(AVAILABLE_MODELS.keys()).index(default_model_key), key=f"expert_model_{i}") @@ -376,12 +648,58 @@ if st.session_state.mode == "Deep Research": research_context = st.text_area("补充背景 (可选)", placeholder="任何额外的背景信息...", height=80) - start_research_btn = st.button("🚀 开始多模型协作", type="primary", disabled=not research_topic) + start_research_btn = st.button("🚀 开始多模型协作", type="primary", disabled=(not research_topic or not api_key)) + if not api_key: + st.info("💡 请先在侧边栏配置 API Key 才能开始任务") + + # ==================== 恢复会话逻辑 (Resume Logic) ==================== + # Try to load cached session + cached_session = st.session_state.storage.load_session_state("council_cache") + + # If we have a cached session, and we are NOT currently running one (research_started is False) + if cached_session and not st.session_state.research_started: + st.info(f"🔍 检测到上次未完成的会话: {cached_session.get('topic', 'Unknown Topic')}") + col_res1, col_res2 = st.columns([1, 4]) + with col_res1: + if st.button("🔄 恢复会话", type="primary"): + # Restore state + st.session_state.research_started = True + st.session_state.research_output = "" # Usually empty if unfinished + st.session_state.research_steps_output = cached_session.get("steps_output", []) + + # Restore inputs if possible (tricky with widgets, but we can set defaults or just rely on cache for display) + # For simplicity, we restore the viewing state. Continuing generation is harder without rebuilding the exact generator state. + # Currently, "Resume" means "Restore View". To continue adding to it would require skipping done steps in manager. + + st.rerun() + with col_res2: + if st.button("🗑️ 放弃", type="secondary"): + st.session_state.storage.clear_session_state("council_cache") + st.rerun() + + # ==================== 历史渲染区域 (Always visible if started) ==================== + if st.session_state.research_started and st.session_state.research_steps_output and not start_research_btn: + st.subheader("🗣️ 智囊团讨论历史") + for step in st.session_state.research_steps_output: + step_name = step.get('step', 'Unknown') + content = step.get('output', '') + role_type = "assistant" + + with st.chat_message(role_type, avatar="🤖"): + st.markdown(f"**{step_name}**") + st.markdown(content) + st.divider() + + # ==================== 执行区域 (Triggered by Button) ==================== if start_research_btn and research_topic: st.session_state.research_started = True st.session_state.research_output = "" st.session_state.research_steps_output = [] + + # Clear any old cache when starting fresh + st.session_state.storage.clear_session_state("council_cache") + # 使用全局页面背景（若已上传） research_bg_path = st.session_state.get("bg_image_path") if st.session_state.get("bg_image_data_url"): @@ -404,9 +722,7 @@ if st.session_state.mode == "Deep Research": ) manager.create_agents(config_obj) - st.divider() st.subheader("🗣️ 智囊团讨论中...") - chat_container = st.container() try: @@ -418,10 +734,11 @@ if st.session_state.mode == "Deep Research": # Create a chat message block with chat_container: - st.markdown(f"#### {current_step_name}") - st.caption(f"🤖 {current_agent} ({current_model})") - message_placeholder = st.empty() - current_content = "" + with st.chat_message("assistant", avatar="🤖"): + st.markdown(f"**{current_step_name}**") + st.caption(f"({current_model})") + message_placeholder = st.empty() + current_content = "" elif event["type"] == "content": current_content += event["content"] @@ -433,7 +750,18 @@ if st.session_state.mode == "Deep Research": "step": current_step_name, "output": event["output"] }) - st.divider() # Separator between turns + + # === AUTO-SAVE CACHE === + # Save current progress to session cache + cache_data = { + "topic": research_topic, + "context": research_context, + "steps_output": st.session_state.research_steps_output, + "experts_config": experts_config, + "max_rounds": max_rounds + } + st.session_state.storage.save_session_state("council_cache", cache_data) + # ======================= # The last step output is the final plan if st.session_state.research_steps_output: @@ -456,6 +784,10 @@ if st.session_state.mode == "Deep Research": content=final_plan, metadata=metadata ) + + # Clear session cache as we finished successfully + st.session_state.storage.clear_session_state("council_cache") + st.toast("✅ 记录已保存到历史档案") except Exception as e: @@ -624,6 +956,8 @@ elif st.session_state.mode == "Debate Workshop": type="primary", use_container_width=True ) + if not api_key: + st.caption("🔒 需配置 API Key") with col_btn2: reset_btn = st.button( @@ -863,6 +1197,117 @@ elif st.session_state.mode == "History Archives": file_name=f"{record['type']}_{record['id']}.md" ) +# ==================== 用户反馈页面 ==================== +elif st.session_state.mode == "Feedback": + st.title("💬 用户反馈") + st.markdown("*您的反馈帮助我们不断改进产品*") + + # Feedback form + st.subheader("📝 提交反馈") + + feedback_type = st.selectbox( + "反馈类型", + ["功能建议", "Bug 报告", "使用体验", "其他"], + help="选择您要反馈的类型" + ) + + # Rating + st.markdown("**整体满意度**") + rating = st.slider("", 1, 5, 4, format="%d ⭐") + rating_labels = {1: "😞 非常不满意", 2: "😕 不满意", 3: "😐 一般", 4: "😊 满意", 5: "🤩 非常满意"} + st.caption(rating_labels.get(rating, "")) + + # Feedback content + feedback_content = st.text_area( + "详细描述", + placeholder="请描述您的反馈内容...\n\n例如：\n- 您遇到了什么问题？\n- 您希望增加什么功能？\n- 您对哪些方面有改进建议？", + height=200 + ) + + # Feature requests for Council V4 + st.subheader("🎯 功能需求调研") + st.markdown("您最希望看到哪些新功能？（可多选）") + + feature_options = { + "more_scenarios": "📋 更多决策场景模板", + "export_pdf": "📄 导出 PDF 报告", + "voice_input": "🎤 语音输入支持", + "realtime_collab": "👥 多人实时协作", + "custom_prompts": "✏️ 自定义专家 Prompt", + "api_access": "🔌 API 接口支持", + "mobile_app": "📱 移动端应用" + } + + selected_features = [] + cols = st.columns(3) + for idx, (key, label) in enumerate(feature_options.items()): + with cols[idx % 3]: + if st.checkbox(label, key=f"feature_{key}"): + selected_features.append(key) + + # Contact info (optional) + st.subheader("📧 联系方式（可选）") + contact_email = st.text_input("邮箱", placeholder="your@email.com") + + # Submit button + st.divider() + if st.button("📤 提交反馈", type="primary", use_container_width=True): + if feedback_content.strip(): + # Save feedback + feedback_data = { + "type": feedback_type, + "rating": rating, + "content": feedback_content, + "features": selected_features, + "email": contact_email, + "timestamp": st.session_state.storage._get_timestamp() if hasattr(st.session_state.storage, '_get_timestamp') else "" + } + + # Save to storage + try: + import json + import os + feedback_dir = os.path.join(st.session_state.storage.base_dir, "feedback") + os.makedirs(feedback_dir, exist_ok=True) + + from datetime import datetime + filename = f"feedback_{datetime.now().strftime('%Y%m%d_%H%M%S')}.json" + filepath = os.path.join(feedback_dir, filename) + + with open(filepath, 'w', encoding='utf-8') as f: + json.dump(feedback_data, f, ensure_ascii=False, indent=2) + + st.success("🎉 感谢您的反馈！我们会认真阅读并持续改进产品。") + st.balloons() + except Exception as e: + st.error(f"保存反馈时出错: {e}") + else: + st.warning("请填写反馈内容") + + # Show previous feedback summary + st.divider() + with st.expander("📊 我的反馈历史"): + try: + import os + import json + feedback_dir = os.path.join(st.session_state.storage.base_dir, "feedback") + if os.path.exists(feedback_dir): + files = sorted(os.listdir(feedback_dir), reverse=True)[:5] + if files: + for f in files: + filepath = os.path.join(feedback_dir, f) + with open(filepath, 'r', encoding='utf-8') as file: + data = json.load(file) + st.markdown(f"**{data.get('timestamp', 'Unknown')}** | {data.get('type', '')} | {'⭐' * data.get('rating', 0)}") + st.caption(data.get('content', '')[:100] + "...") + st.divider() + else: + st.info("暂无反馈记录") + else: + st.info("暂无反馈记录") + except Exception: + st.info("暂无反馈记录") + # ==================== 底部信息 ==================== st.divider() col_footer1, col_footer2, col_footer3 = st.columns(3) diff --git a/config.py b/config.py index 1ee3715..6e6a67f 100644 --- a/config.py +++ b/config.py @@ -96,6 +96,9 @@ MAX_AGENTS = 6 # 最大参与 Agent 数量 # 支持的输出语言 SUPPORTED_LANGUAGES = ["Chinese", "English", "Japanese", "Spanish", "French", "German"] +# 生成配置 +MAX_OUTPUT_TOKENS = 300 # 限制单次回复长度，保持精简 + # 研究模式模型角色配置 RESEARCH_MODEL_ROLES = { "expert_a": { diff --git a/utils/__pycache__/llm_client.cpython-313.pyc b/utils/__pycache__/llm_client.cpython-313.pyc index 1a8cdb0..ed60228 100644 Binary files a/utils/__pycache__/llm_client.cpython-313.pyc and b/utils/__pycache__/llm_client.cpython-313.pyc differ diff --git a/utils/__pycache__/storage.cpython-313.pyc b/utils/__pycache__/storage.cpython-313.pyc index 00dbe78..7af2bba 100644 Binary files a/utils/__pycache__/storage.cpython-313.pyc and b/utils/__pycache__/storage.cpython-313.pyc differ diff --git a/utils/auto_agent_generator.py b/utils/auto_agent_generator.py new file mode 100644 index 0000000..b9beaa6 --- /dev/null +++ b/utils/auto_agent_generator.py @@ -0,0 +1,108 @@ +""" +Auto Agent Generator - 根据主题自动生成专家配置 +Uses LLM to analyze the topic and suggest appropriate expert agents. +""" +import json +import re +from typing import List, Dict +from utils.llm_client import LLMClient + + +EXPERT_GENERATION_PROMPT = """You are an expert team composition advisor. Given a research/decision topic, you need to suggest the most appropriate team of experts to analyze it. + +Instructions: +1. Analyze the topic carefully to understand its domain and key aspects +2. Generate {num_experts} distinct expert roles that would provide the most valuable perspectives +3. Each expert should have a unique focus area relevant to the topic +4. The LAST expert should always be a "Synthesizer" role who can integrate all perspectives + +Output Format (MUST be valid JSON array): +[ + {{"name": "Expert Name", "perspective": "Brief description of their viewpoint", "focus": "Key areas they analyze"}}, + ... +] + +Examples of good expert names based on topic: +- For "Should we launch an e-commerce platform?": "市场渠道分析师", "电商运营专家", "供应链顾问", "数字化转型综合师" +- For "Career transition to AI field": "职业发展顾问", "AI行业专家", "技能评估分析师", "综合规划师" + +IMPORTANT: +- Use {language} for all names and descriptions +- Make names specific to the topic, not generic like "Expert 1" +- The last expert MUST be a synthesizer/integrator type + +Topic: {topic} + +Generate exactly {num_experts} experts as a JSON array:""" + + +def generate_experts_for_topic( + topic: str, + num_experts: int, + llm_client: LLMClient, + language: str = "Chinese" +) -> List[Dict[str, str]]: + """ + Use LLM to generate appropriate expert configurations based on the topic. + + Args: + topic: The research/decision topic + num_experts: Number of experts to generate (2-5) + llm_client: LLM client instance for API calls + language: Output language (Chinese/English) + + Returns: + List of expert dicts: [{"name": "...", "perspective": "...", "focus": "..."}, ...] + """ + if not topic.strip(): + return [] + + prompt = EXPERT_GENERATION_PROMPT.format( + topic=topic, + num_experts=num_experts, + language=language + ) + + try: + response = llm_client.chat( + system_prompt="You are a helpful assistant that generates JSON output only. No markdown, no explanation.", + user_prompt=prompt, + max_tokens=800 + ) + + # Extract JSON from response (handle potential markdown wrapping) + json_match = re.search(r'\[[\s\S]*\]', response) + if json_match: + experts = json.loads(json_match.group()) + # Validate structure + if isinstance(experts, list) and len(experts) >= 1: + validated = [] + for exp in experts[:num_experts]: + if isinstance(exp, dict) and "name" in exp: + validated.append({ + "name": exp.get("name", "Expert"), + "perspective": exp.get("perspective", ""), + "focus": exp.get("focus", "") + }) + return validated + except (json.JSONDecodeError, Exception) as e: + print(f"[AutoAgentGenerator] Error parsing LLM response: {e}") + + # Fallback: return generic experts + fallback = [] + for i in range(num_experts): + if i == num_experts - 1: + fallback.append({"name": f"综合分析师", "perspective": "整合视角", "focus": "综合决策"}) + else: + fallback.append({"name": f"专家 {i+1}", "perspective": "分析视角", "focus": "专业分析"}) + return fallback + + +def get_default_model_for_expert(expert_index: int, total_experts: int, available_models: list) -> str: + """ + Assign a default model to an expert based on their position. + Spreads experts across available models for diversity. + """ + if not available_models: + return "gpt-4o" + return available_models[expert_index % len(available_models)] diff --git a/utils/llm_client.py b/utils/llm_client.py index f3881d8..0b11447 100644 --- a/utils/llm_client.py +++ b/utils/llm_client.py @@ -5,6 +5,8 @@ from typing import Generator import os +import config + class LLMClient: """LLM API 统一客户端""" @@ -62,7 +64,7 @@ class LLMClient: self, system_prompt: str, user_prompt: str, - max_tokens: int = 1024 + max_tokens: int = config.MAX_OUTPUT_TOKENS ) -> Generator[str, None, None]: """ 流式对话 diff --git a/utils/storage.py b/utils/storage.py index f827112..84e5dc0 100644 --- a/utils/storage.py +++ b/utils/storage.py @@ -150,3 +150,35 @@ class StorageManager: return json.load(f) except Exception: return None + + # ==================== Session Cache (Resume Functionality) ==================== + def save_session_state(self, key: str, data: Dict[str, Any]): + """Save temporary session state for recovery""" + try: + # We use a dedicated cache file per key + cache_file = self.root_dir / f"{key}_cache.json" + data["_timestamp"] = int(time.time()) + with open(cache_file, 'w', encoding='utf-8') as f: + json.dump(data, f, indent=2, ensure_ascii=False) + except Exception as e: + print(f"Error saving session cache: {e}") + + def load_session_state(self, key: str) -> Dict[str, Any]: + """Load temporary session state""" + cache_file = self.root_dir / f"{key}_cache.json" + if not cache_file.exists(): + return None + try: + with open(cache_file, 'r', encoding='utf-8') as f: + return json.load(f) + except Exception: + return None + + def clear_session_state(self, key: str): + """Clear temporary session state""" + cache_file = self.root_dir / f"{key}_cache.json" + if cache_file.exists(): + try: + os.remove(cache_file) + except Exception: + pass