From 5a919c24e4ac045c8c14c02e8ef808390c06d2d8 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?=E5=8A=B3=E8=B0=A6?= Date: Fri, 28 Nov 2025 17:05:30 +0800 Subject: [PATCH] desc --- aworld/config/conf.py | 41 ++------------ aworld/core/context/amni/config.py | 89 +++++++++--------------------- 2 files changed, 31 insertions(+), 99 deletions(-) diff --git a/aworld/config/conf.py b/aworld/config/conf.py index 7787be541..d71de7159 100644 --- a/aworld/config/conf.py +++ b/aworld/config/conf.py @@ -1,6 +1,5 @@ # coding: utf-8 # Copyright (c) 2025 inclusionAI. -import copy import os import traceback import uuid @@ -82,12 +81,6 @@ class ClientType(Enum): HTTP = "http" -class HistoryWriteStrategy(Enum): - """History write strategy for memory operations.""" - EVENT_DRIVEN = "event_driven" # Write through message system (default) - DIRECT = "direct" # Direct call to memory handler - - class ConfigDict(dict): """Object mode operates dict, can read non-existent attributes through `get` method.""" __setattr__ = dict.__setitem__ @@ -160,10 +153,10 @@ class OptimizationConfig(BaseConfig): class SummaryPromptConfig(BaseConfig): """Configuration for summary prompt templates.""" - template: str = Field(description="Base template, such as AWORLD_MEMORY_EXTRACT_NEW_SUMMARY") - summary_rule: str = Field(description="Summary rule, used to guide how to generate summaries") - summary_schema: str = Field(description="Summary schema, defines output format and structure") - memory_type: str = Field(default="summary", description="Memory type, used to distinguish different types of summaries") + template: str = Field(description="基础模板，如 AWORLD_MEMORY_EXTRACT_NEW_SUMMARY") + summary_rule: str = Field(description="摘要规则，用于指导如何生成摘要") + summary_schema: str = Field(description="摘要模式，定义输出格式和结构") + memory_type: str = Field(default="summary", description="记忆类型，用于区分不同类型的摘要") class ContextRuleConfig(BaseConfig): @@ -186,9 +179,6 @@ class AgentMemoryConfig(BaseConfig): # short-term config history_rounds: int = Field(default=100, description="rounds of message msg; when the number of messages is greater than the history_rounds, the memory will be trimmed") - history_write_strategy: HistoryWriteStrategy = Field(default=HistoryWriteStrategy.EVENT_DRIVEN, - description="History write strategy: event_driven (through message system) or direct (direct call to handler)") - enable_summary: bool = Field(default=False, description="enable_summary use llm to create summary short-term memory") summary_model: Optional[str] = Field(default=None, description="short-term summary model") @@ -197,8 +187,7 @@ class AgentMemoryConfig(BaseConfig): summary_context_length: Optional[int] = Field(default=40960, description=" when the content length is greater than the summary_context_length, the summary will be created") summary_prompts: Optional[List[SummaryPromptConfig]] = Field(default=[]) - summary_summaried: Optional[bool] = Field(default=True, description="to summary summary message when summary triggered") - summary_role: Optional[str] = Field(default="assistant", description="role for summary memory items") + summary_summaried: Optional[bool] = Field(default=True, description="whether to summarize historical summary messages when summary is triggered") # Long-term memory config enable_long_term: bool = Field(default=False, description="enable_long_term use to store long-term memory") @@ -206,25 +195,6 @@ class AgentMemoryConfig(BaseConfig): # LongTermConfig long_term_config: Optional[BaseModel] = Field(default=None, description="long_term_config") - def __deepcopy__(self, memo=None): - """Support copy.deepcopy for AgentMemoryConfig.""" - if memo is None: - memo = {} - - # Check if already copied (avoid circular references) - if id(self) in memo: - return memo[id(self)] - - # Create a new instance using model_dump and model_validate to avoid recursion - # Use mode='python' to get plain Python objects - data = self.model_dump(mode='python') - # Deep copy the data dict to handle nested objects - copied_data = copy.deepcopy(data, memo) - # Create new instance from copied data - new_instance = self.__class__.model_validate(copied_data) - memo[id(self)] = new_instance - return new_instance - class AgentConfig(BaseConfig): llm_config: ModelConfig = ModelConfig() @@ -246,7 +216,6 @@ class AgentConfig(BaseConfig): exit_on_failure: bool = False human_tools: List[str] = [] skill_configs: Dict[str, Any] = None - ext: dict = {} def __init__(self, **kwargs): super().__init__(**kwargs) diff --git a/aworld/core/context/amni/config.py b/aworld/core/context/amni/config.py index 04fe75c9d..fb0123a0e 100644 --- a/aworld/core/context/amni/config.py +++ b/aworld/core/context/amni/config.py @@ -1,18 +1,23 @@ # coding: utf-8 # Copyright (c) 2025 inclusionAI. +import logging import os +import queue +import threading from enum import Enum -from typing import Optional, List, Union, Dict +from logging.handlers import TimedRotatingFileHandler +from typing import Optional, List, Union, Dict, Any from pydantic import BaseModel, Field from aworld.config import ModelConfig -from aworld.config.conf import AgentMemoryConfig, SummaryPromptConfig, HistoryWriteStrategy -from aworld.core.memory import MemoryConfig +from aworld.config.conf import AgentMemoryConfig, SummaryPromptConfig +from aworld.core.memory import MemoryConfig, MemoryLLMConfig, EmbeddingsConfig from aworld.memory.db.sqlite import SQLiteMemoryStore # from aworld.memory.db import SQLiteMemoryStore # Temporarily commented out to avoid import errors from aworld.memory.main import MemoryFactory from .retrieval.base import RetrieverFactory +from .retrieval.graph.base import GraphDBConfig from ...event.base import TopicType @@ -88,8 +93,6 @@ class AgentContextConfig(BaseConfig): # Context Reduce - Purge history_rounds: int = Field(default=100, description="rounds of message msg; when the number of messages is greater than the history_rounds, the memory will be trimmed") - history_write_strategy: HistoryWriteStrategy = Field(default=HistoryWriteStrategy.EVENT_DRIVEN, - description="History write strategy: event_driven (through message system) or direct (direct call to handler)") # Context Reduce - Compress enable_summary: bool = Field(default=False, @@ -100,8 +103,7 @@ class AgentContextConfig(BaseConfig): summary_context_length: Optional[int] = Field(default=40960, description=" when the content length is greater than the summary_context_length, the summary will be created") summary_prompts: Optional[List[SummaryPromptConfig]] = Field(default=[]) - summary_summaried: Optional[bool] = Field(default=True, description="summary_summaried use to store summary memory") - summary_role: Optional[str] = Field(default="user", description="role for summary memory items") + summary_summaried: Optional[bool] = Field(default=True, description="whether to summarize historical summary messages when summary is triggered") # Context Offload tool_result_offload: bool = Field(default=False, description="tool result offload") @@ -117,61 +119,22 @@ class AgentContextConfig(BaseConfig): def to_memory_config(self) -> AgentMemoryConfig: return AgentMemoryConfig( history_rounds=self.history_rounds, - history_write_strategy=self.history_write_strategy, enable_summary=self.enable_summary, summary_rounds=self.summary_rounds, summary_context_length=self.summary_context_length, summary_prompts=self.summary_prompts, - summary_summaried=self.summary_summaried, - summary_role=self.summary_role + summary_summaried=self.summary_summaried ) DEFAULT_AGENT_CONFIG = AgentContextConfig() -class WorkingDirOssConfig(BaseModel): - """OSS configuration for working directory.""" - access_key_id: Optional[str] = Field( - default=None, - description="OSS access key ID. Priority: config > WORKING_DIR_OSS_ACCESS_KEY_ID > OSS_ACCESS_KEY_ID" - ) - access_key_secret: Optional[str] = Field( - default=None, - description="OSS access key secret. Priority: config > WORKING_DIR_OSS_ACCESS_KEY_SECRET > OSS_ACCESS_KEY_SECRET" - ) - endpoint: Optional[str] = Field( - default=None, - description="OSS endpoint. Priority: config > WORKING_DIR_OSS_ENDPOINT > OSS_ENDPOINT" - ) - bucket_name: Optional[str] = Field( - default=None, - description="OSS bucket name. Priority: config > WORKING_DIR_OSS_BUCKET_NAME > OSS_BUCKET_NAME" - ) - class ContextEnvConfig(BaseModel): """Represents environment configuration for an agent team.""" isolate: bool = Field(default=False, description="One Task, One Isolate Env") env_type: str = Field(default="local", description="Env Type, local|remote") env_mount_path: str = Field(default="~/workspace", description="Env Working directory for share") env_config: dict = Field(default_factory=dict, description="Env Config") - - # Working directory path configuration - working_dir_base_path: Optional[str] = Field( - default=None, - description="Base path for working directory. Priority: config > WORKING_DIR_BASE_PATH > WORKING_DIR_OSS_BASE_PATH > WORKSPACE_PATH" - ) - working_dir_path_template: Optional[str] = Field( - default=None, - description="Template for working directory path. Supports placeholders: {base_path}, {session_id}. " - "Example: '{base_path}/custom/{session_id}/workspace' or '{base_path}/{session_id}/files'. " - "Priority: config > WORKING_DIR_PATH_TEMPLATE > default" - ) - - # OSS configuration for remote working directory - working_dir_oss_config: Optional[WorkingDirOssConfig] = Field( - default=None, - description="OSS configuration for working directory. Priority: config > WORKING_DIR_OSS_* > OSS_* environment variables" - ) class AmniContextConfig(BaseConfig): """AmniContext configs""" @@ -219,6 +182,7 @@ def init_middlewares(init_memory: bool = True, init_retriever: bool = True) -> N RetrieverFactory.init() def build_memory_config(): + from aworld.core.memory import VectorDBConfig return MemoryConfig( provider="aworld", llm_config=ModelConfig( @@ -227,19 +191,19 @@ def build_memory_config(): api_key=os.getenv("LLM_API_KEY"), base_url=os.getenv("LLM_BASE_URL") ), - # embedding_config=EmbeddingsConfig( - # base_url=os.getenv('EMBEDDING_BASE_URL'), - # api_key=os.getenv('EMBEDDING_API_KEY'), - # model_name=os.getenv('EMBEDDING_MODEL_NAME'), - # dimensions=int(os.getenv('EMBEDDING_MODEL_DIMENSIONS', '1024')) - # ), - # vector_store_config=VectorDBConfig( - # provider="chroma", - # config={ - # "chroma_data_path": os.getenv('CHROMA_PATH', "./data/chroma_db"), - # "collection_name": "aworld_memory", - # } - # ) + embedding_config=EmbeddingsConfig( + base_url=os.getenv('EMBEDDING_BASE_URL'), + api_key=os.getenv('EMBEDDING_API_KEY'), + model_name=os.getenv('EMBEDDING_MODEL_NAME'), + dimensions=int(os.getenv('EMBEDDING_MODEL_DIMENSIONS', '1024')) + ), + vector_store_config=VectorDBConfig( + provider="chroma", + config={ + "chroma_data_path": os.getenv('CHROMA_PATH', "./data/chroma_db"), + "collection_name": "aworld_memory", + } + ) ) @@ -274,6 +238,7 @@ def get_default_config() -> AmniContextConfig: priority=0 ) ], + debug_mode=True ) @@ -311,9 +276,7 @@ def create(level: Optional[AmniConfigLevel] = None, if not level or level == AmniConfigLevel.PILOT or level == AmniConfigLevel.COPILOT: config = get_default_config() - config.agent_config = AgentContextConfig(neuron_names=neuron_names) - if neuron_names: - config.agent_config.enable_system_prompt_augment = True + config.agent_config = AgentContextConfig() config.debug_mode = debug_mode config.env_config = env_config return config