alayaflow 0.1.2__tar.gz → 0.1.4__tar.gz

alayaflow-0.1.4/.alaya.ai/alayaflow/workflows/alayamem_chat/1.0.0/metadata.json

@@ -0,0 +1,9 @@
+{
+    "id": "alayamem_chat",
+    "name": "AlayaMem Chat",
+    "description": "一个集成 AlayaMem 记忆功能的对话工作流",
+    "version": "1.0.0",
+    "tags": ["chat", "memory", "alayamem"],
+    "entry_file": "workflow.py",
+    "entry_point": "create_graph" 
+}

alayaflow-0.1.4/.alaya.ai/alayaflow/workflows/alayamem_chat/1.0.0/schemas.py

@@ -0,0 +1,61 @@
+from typing import TypedDict, List, Optional
+
+from pydantic import BaseModel, Field
+
+from langchain_core.messages import BaseMessage, AIMessageChunk
+
+
+class WorkflowInitArgs(BaseModel):
+    alayamem_url: str = Field(..., description="AlayaMem URL")
+
+
+class Input(BaseModel):
+    messages: List[BaseMessage] = Field(..., description="List of input messages")
+
+
+class WorkflowContext(BaseModel):
+    user_id: str = Field(..., description="User ID")
+    session_id: str = Field(..., description="Session ID")
+    chat_model_id: str = Field(..., description="Chat Model ID")
+
+    system_prompt: str = Field(
+        default="你是一个有帮助的AI助手。",
+        description="系统基础提示词"
+    )
+    user_profile_prompt: str = Field(
+        default="## 用户画像\n{user_profile}",
+        description="用户画像提示词模板，使用 {user_profile} 作为占位符"
+    )
+    session_summary_prompt: str = Field(
+        default="## 会话摘要\n{session_summary}",
+        description="会话摘要提示词模板，使用 {session_summary} 作为占位符"
+    )
+    retrieved_docs_prompt: str = Field(
+        default="## 相关参考资料\n{retrieved_docs}",
+        description="检索文档提示词模板，使用 {retrieved_docs} 作为占位符"
+    )
+    history_turns_prompt: str = Field(
+        default="## 历史对话\n{history_turns}",
+        description="历史对话提示词模板，使用 {history_turns} 作为占位符"
+    )
+    context_wrapper_prompt: str = Field(
+        default="# 上下文信息\n\n{context_parts}\n\n请基于以上信息来回答用户的问题。",
+        description="上下文包装提示词模板，使用 {context_parts} 作为占位符"
+    )
+
+
+class Output(BaseModel):
+    chat_response: dict = Field(..., description="Chat response")
+
+
+class WorkflowState(TypedDict):
+    messages: List[BaseMessage]
+    memory_initialized: bool
+    retrieved_docs: Optional[List[str]]
+    stream_chunks: List[AIMessageChunk]
+    chat_response: Optional[dict]
+    context: Optional[str]
+    # 新增字段
+    history_turns: Optional[List[dict]]
+    session_summary: Optional[str]
+    user_profile: Optional[dict]

alayaflow-0.1.4/.alaya.ai/alayaflow/workflows/alayamem_chat/1.0.0/workflow.py

@@ -0,0 +1,246 @@
+from langgraph.graph import StateGraph, START, END
+from langgraph.runtime import Runtime
+
+from alayaflow.component.memory import query_file, commit_turn, turns, summary, profile
+from alayaflow.component.model import ModelManager
+
+from .schemas import WorkflowInitArgs, WorkflowState, WorkflowContext, Input, Output
+
+def mk_turns_node(alayamem_url: str):
+    def turns_node(state: WorkflowState, runtime: Runtime[WorkflowContext]):
+        session_id = runtime.context.session_id
+        user_id = runtime.context.user_id
+        original_result = turns(alayamem_url, session_id, user_id)
+
+        # 打印返回字段
+        print(f"\n[1. turns 接口返回]")
+        print(f"  返回类型: {type(original_result)}")
+        print(f"  返回值: {original_result}")
+
+        updated_state = state.copy()
+        # turns 接口直接返回列表
+        updated_state["history_turns"] = original_result if isinstance(original_result, list) else []
+        return updated_state
+    return turns_node
+
+def mk_query_file_node(alayamem_url: str):
+    def query_file_node(state: WorkflowState, runtime: Runtime[WorkflowContext]):
+        user_id = runtime.context.user_id
+        messages = state.get("messages", [])
+        message = messages[-1] if messages else None
+        original_result = query_file(alayamem_url, user_id, message)
+
+        # 打印返回字段
+        print(f"\n[2. query_file 接口返回]")
+        print(f"  返回类型: {type(original_result)}")
+        if isinstance(original_result, dict):
+            print(f"  字段列表: {list(original_result.keys())}")
+        print(f"  返回值: {original_result}")
+
+        updated_state = state.copy()
+        # query_file 返回 dict，results 是 dict，包含 documents 字段
+        if isinstance(original_result, dict):
+            results = original_result.get("results", {})
+            updated_state["retrieved_docs"] = results.get("documents", []) if isinstance(results, dict) else []
+        else:
+            updated_state["retrieved_docs"] = []
+        return updated_state
+    return query_file_node
+
+def mk_summary_node(alayamem_url: str):
+    def summary_node(state: WorkflowState, runtime: Runtime[WorkflowContext]):
+        session_id = runtime.context.session_id
+        user_id = runtime.context.user_id
+        original_result = summary(alayamem_url, session_id, user_id)
+
+        # 打印返回字段
+        print(f"\n[2. summary 接口返回]")
+        print(f"  返回类型: {type(original_result)}")
+        if isinstance(original_result, dict):
+            print(f"  字段列表: {list(original_result.keys())}")
+        print(f"  返回值: {original_result}")
+
+        updated_state = state.copy()
+        # summary 可能是 None，需要处理
+        if isinstance(original_result, dict):
+            summary_value = original_result.get("summary", "")
+            updated_state["session_summary"] = summary_value if summary_value is not None else ""
+        else:
+            updated_state["session_summary"] = ""
+        return updated_state
+    return summary_node
+
+def mk_profile_node(alayamem_url: str):
+    def profile_node(state: WorkflowState, runtime: Runtime[WorkflowContext]):
+        user_id = runtime.context.user_id
+        original_result = profile(alayamem_url, user_id)
+
+        # 打印返回字段
+        print(f"\n[3. profile 接口返回]")
+        print(f"  返回类型: {type(original_result)}")
+        if isinstance(original_result, dict):
+            print(f"  字段列表: {list(original_result.keys())}")
+        print(f"  返回值: {original_result}")
+
+        updated_state = state.copy()
+        # profile 可能是 None，需要处理
+        if isinstance(original_result, dict):
+            profile_value = original_result.get("profile", {})
+            updated_state["user_profile"] = profile_value if profile_value is not None else {}
+        else:
+            updated_state["user_profile"] = {}
+        return updated_state
+    return profile_node
+
+def mk_commit_turn_node(alayamem_url: str):
+    def commit_turn_node(state: WorkflowState, runtime: Runtime[WorkflowContext]):
+        session_id = runtime.context.session_id
+        user_id = runtime.context.user_id
+        messages = state.get("messages", [])
+        chat_response = state.get("chat_response")
+
+        # 提取用户消息和助手回复
+        user_text = ""
+        if messages:
+            last_msg = messages[-1]
+            if hasattr(last_msg, 'content'):
+                user_text = last_msg.content
+            elif isinstance(last_msg, dict):
+                user_text = last_msg.get('content', '')
+
+        assistant_text = ""
+        if chat_response:
+            if hasattr(chat_response, 'content'):
+                assistant_text = chat_response.content
+            elif isinstance(chat_response, dict):
+                assistant_text = chat_response.get('content', '')
+
+        # 调用 commit_turn
+        original_result = commit_turn(alayamem_url, session_id, user_text,
+                                     assistant_text, user_id, window_size=3)
+
+        # 打印返回字段
+        print(f"\n[4. commit_turn 接口返回]")
+        print(f"  返回类型: {type(original_result)}")
+        if isinstance(original_result, dict):
+            print(f"  字段列表: {list(original_result.keys())}")
+        print(f"  返回值: {original_result}")
+
+        return state.copy()
+    return commit_turn_node
+
+def mk_chat_node():
+    model_manager = ModelManager()
+
+    def chat_node(state: WorkflowState, runtime: Runtime[WorkflowContext]):
+        model_id = runtime.context.chat_model_id
+        chat_model = model_manager.get_model(model_id)
+        if not chat_model:
+            raise ValueError(f"无法找到模型ID为 '{model_id}' 的模型配置")
+
+        from langchain_core.messages import SystemMessage, HumanMessage, AIMessage
+
+        # 获取 prompt 模板配置
+        ctx = runtime.context
+        system_prompt = ctx.system_prompt
+        user_profile_prompt = ctx.user_profile_prompt
+        session_summary_prompt = ctx.session_summary_prompt
+        retrieved_docs_prompt = ctx.retrieved_docs_prompt
+        history_turns_prompt = ctx.history_turns_prompt
+        context_wrapper_prompt = ctx.context_wrapper_prompt
+
+        # 构建完整的消息列表
+        full_messages = []
+
+        # 添加系统基础提示词
+        if system_prompt:
+            full_messages.append(SystemMessage(content=system_prompt))
+
+        # 构建上下文信息部分
+        context_parts = []
+
+        # 添加用户画像
+        # user_profile = state.get("user_profile", {})
+        # if user_profile:
+        #     profile_text = user_profile_prompt.format(user_profile=str(user_profile))
+        #     context_parts.append(profile_text)
+
+        # 添加会话摘要
+        session_summary = state.get("session_summary", "")
+        if session_summary:
+            summary_text = session_summary_prompt.format(session_summary=session_summary)
+            context_parts.append(summary_text)
+
+        # 添加检索文档
+        retrieved_docs = state.get("retrieved_docs", [])
+        if retrieved_docs:
+            docs_content = "\n\n".join([str(doc) for doc in retrieved_docs])
+            docs_text = retrieved_docs_prompt.format(retrieved_docs=docs_content)
+            context_parts.append(docs_text)
+
+        # 添加历史对话
+        history_turns = state.get("history_turns", [])
+        if history_turns:
+            history_content = "\n".join([
+                f"用户: {turn.get('user_text', '')}\n助手: {turn.get('assistant_text', '')}"
+                for turn in history_turns
+            ])
+            history_text = history_turns_prompt.format(history_turns=history_content)
+            context_parts.append(history_text)
+
+        # 如果有上下文信息，添加上下文系统消息
+        if context_parts:
+            context_content = context_wrapper_prompt.format(
+                context_parts="\n\n".join(context_parts)
+            )
+            full_messages.append(SystemMessage(content=context_content))
+
+        # 添加当前用户消息
+        current_messages = state["messages"]
+        full_messages.extend(current_messages)
+
+        # 打印最终输入到模型的消息
+        print(f"\n[5. 模型输入消息]")
+        print(f"  消息数量: {len(full_messages)}")
+        print("-" * 50)
+        full_content = ""
+        for msg in full_messages:
+            msg_type = type(msg).__name__
+            content = msg.content if hasattr(msg, 'content') else str(msg)
+            full_content += f"[{msg_type}]\n{content}\n\n"
+        print(full_content)
+        print("-" * 50)
+
+        # 调用模型
+        response = chat_model.invoke(full_messages)
+
+        updated_state = state.copy()
+        updated_state['chat_response'] = response
+        return updated_state
+    return chat_node
+
+def create_graph(init_args: WorkflowInitArgs | dict):
+    if isinstance(init_args, dict):
+        init_args = WorkflowInitArgs(**init_args)
+    alayamem_url = init_args.alayamem_url
+
+    graph = StateGraph(WorkflowState, WorkflowContext, input_type=Input, output_type=Output)
+
+    # 添加节点
+    graph.add_node("turns_node", mk_turns_node(alayamem_url))
+    graph.add_node("query_file_node", mk_query_file_node(alayamem_url))
+    # 
+    graph.add_node("summary_node", mk_summary_node(alayamem_url))
+    graph.add_node("profile_node", mk_profile_node(alayamem_url))
+    graph.add_node("chat_node", mk_chat_node())
+    graph.add_node("commit_turn_node", mk_commit_turn_node(alayamem_url))
+
+    # 添加边：获取历史 -> 获取摘要 -> 获取画像 -> 聊天 -> 提交记录
+    graph.add_edge(START, "turns_node")
+    graph.add_edge("turns_node",  "summary_node")
+    graph.add_edge("summary_node", "profile_node")
+    graph.add_edge("profile_node", "chat_node")
+    graph.add_edge("chat_node", "commit_turn_node")
+    graph.add_edge("commit_turn_node", END)
+
+    return graph.compile()

alayaflow-0.1.4/.alaya.ai/alayaflow/workflows/autotable/1.0.0/metadata.json

@@ -0,0 +1,9 @@
+{
+    "id": "autotable",
+    "name": "Auto Table Filler",
+    "description": "基于知识库的智能表格自动填写工作流",
+    "version": "1.0.0",
+    "tags": ["extraction", "table", "rag"],
+    "entry_file": "workflow.py",
+    "entry_point": "create_graph"
+}

alayaflow-0.1.4/.alaya.ai/alayaflow/workflows/autotable/1.0.0/metadata.py

@@ -0,0 +1,17 @@
+from pathlib import Path
+
+from alayaflow.workflow import WorkflowInfo
+
+def get_metadata():
+    meta = {
+        "id": "autotable",
+        "name": "Auto Table Filler",
+        "description": "基于知识库的智能表格自动填写工作流",
+        "version": "1.0.0",
+        "tags": ["extraction", "table", "rag"],
+        "entry_file": "workflow.py",
+        "entry_point": "create_graph",
+        "wf_dir": Path(__file__).parent
+    }
+    return WorkflowInfo(**meta)
+

alayaflow-0.1.4/.alaya.ai/alayaflow/workflows/autotable/1.0.0/requirements.txt

@@ -0,0 +1,14 @@
+# LangGraph 核心依赖
+langgraph>=0.2.0
+
+# LangChain Core
+langchain-core>=0.3.0
+
+# LangChain Community (用于 ChatOpenAI)
+langchain-community>=0.3.0
+
+# OpenAI SDK (DeepSeek API 兼容 OpenAI 格式)
+openai>=1.0.0
+
+# Langfuse (可选，用于追踪)
+langfuse>=3.0.0,<4.0.0

alayaflow-0.1.4/.alaya.ai/alayaflow/workflows/autotable/1.0.0/schemas.py

@@ -0,0 +1,44 @@
+from __future__ import annotations
+
+from typing import Dict, List, Any, Union, TypeAlias, TypedDict, Annotated
+from dataclasses import dataclass
+from pydantic import BaseModel, Field
+from .utils import merge_dicts, deep_merge
+
+FieldSpec: TypeAlias = Union[str, Dict[str, List[Any]]]
+
+@dataclass(frozen=True)
+class GroupTask:
+    path: tuple[str, ...]
+    keys: tuple[str, ...]
+
+
+class WorkflowInitArgs(BaseModel):
+    alayamem_url: str = Field(..., description="AlayaMem 服务地址")
+
+
+class Input(BaseModel):
+    fields: List[FieldSpec] = Field(..., description="要填写的表格字段模板（支持嵌套结构）")
+
+
+class WorkflowContext(BaseModel):
+    collection_name: str = Field(default="file_watcher_collection", description="检索的集合名称")
+
+
+class Output(BaseModel):
+    final_result: Dict[str, Any] = Field(..., description="填写完成的表格数据（嵌套结构）")
+    context_by_task: Dict[str, List[str]] = Field(default_factory=dict, description="每个任务检索到的文档片段（调试用）")
+    errors: Dict[str, str] = Field(default_factory=dict, description="错误信息")
+    tasks: List[GroupTask] = Field(default_factory=list, description="规划的任务列表")
+
+
+class OverallState(TypedDict):
+    fields: List[FieldSpec]
+    final_result: Annotated[Dict[str, Any], deep_merge]
+    context_by_task: Annotated[Dict[str, List[str]], merge_dicts]
+    errors: Annotated[Dict[str, str], merge_dicts]
+    tasks: List[GroupTask]
+
+
+class TaskState(TypedDict):
+    task: GroupTask

alayaflow-0.1.4/.alaya.ai/alayaflow/workflows/autotable/1.0.0/utils.py

@@ -0,0 +1,35 @@
+from typing import Any, Dict, List, Tuple
+
+
+def make_patch(path: Tuple[str, ...], kv: Dict[str, str]) -> Dict[str, Any]:
+    """
+    path=("个人信息","联系方式"), kv={"电话":"..","邮箱":".."} =>
+    {"个人信息":{"联系方式":{"电话":"..","邮箱":".."}}}
+    """
+    node: Dict[str, Any] = dict(kv)
+    for p in reversed(path):
+        node = {p: node}
+    return node
+
+
+def slim_docs(docs: List[str], max_doc_chars: int) -> List[str]:
+        out = []
+        for d in docs or []:
+            s = str(d)
+            if len(s) > max_doc_chars:
+                s = s[:max_doc_chars] + "…"
+            out.append(s)
+        return out
+
+
+def merge_dicts(a: Dict, b: Dict) -> Dict:
+    return {**a, **b}
+
+def deep_merge(a: Dict[str, Any], b: Dict[str, Any]) -> Dict[str, Any]:
+    out = dict(a or {})
+    for k, v in (b or {}).items():
+        if k in out and isinstance(out[k], dict) and isinstance(v, dict):
+            out[k] = deep_merge(out[k], v)
+        else:
+            out[k] = v
+    return out