refactor: 单图方案重构 + 动态模型选择 + chat_services优化
All checks were successful
构建并部署 AI Agent 服务 / deploy (push) Successful in 12m9s

## 核心改动

### 1. 单图方案重构
- 删除了多图(self.graphs),改为单图(self.graph)
- 新增 MainGraphState.current_model 字段用于运行时注入模型
- llm_call 节点改为动态选择模型(create_dynamic_llm_call_node)

### 2. chat_services 优化
- 添加 _cached_services 缓存,避免重复初始化
- 新增 get_cached_chat_services() 函数,用于单图注入
- 新增 _check_http_service_available() 统一HTTP探测逻辑
- 减少重复代码,LocalVLLMChatProvider和LocalSmallModelProvider共用探测方法

### 3. AIAgentService 重构
- initialize() 只构建一次图,传入 chat_services 字典
- 新增 _resolve_model() 模型回退逻辑
- 新增 _build_invocation() 统一构建调用参数
- process_message() 和 process_message_stream() 改为注入 current_model
- 流式处理代码拆分,增加可读性

### 4. 新增和删除文件
- 新增:backend/app/main_graph/main_graph_builder.py(图构建)
- 新增:backend/app/main_graph/subgraph_wrapper.py(子图封装)
- 新增:tools/test/test_tavily_search.py(测试)
- 删除:backend/app/main_graph/graph.py(旧图)
- 删除:backend/app/main_graph/utils/main_graph_builder.py(旧构建器)
- 删除:backend/app/main_graph/utils/__init__.py

### 5. 其他更新
- README.md:新增模型服务使用情况详解章节
- backend/app/model_services/__init__.py:新增 get_cached_chat_services 导出

## 方案优势

- 内存优化:N张图 → 1张图
- 灵活性:运行时动态选择模型,支持同会话不同模型
- 性能:模型服务缓存,初始化仅一次
- 可维护性:减少重复代码,统一HTTP探测逻辑
This commit is contained in:
2026-05-05 17:30:55 +08:00
parent 8b5fbbd395
commit b5c15ef445
25 changed files with 1225 additions and 830 deletions

View File

@@ -0,0 +1,149 @@
#!/usr/bin/env python3
"""测试 Tavily 搜索功能 - 直接调用 API"""
import sys
from pathlib import Path
from dataclasses import dataclass
from datetime import datetime
from typing import List, Optional
from dotenv import load_dotenv
# 路径设置
project_root = Path(__file__).resolve().parent.parent.parent
sys.path.insert(0, str(project_root))
load_dotenv(project_root / ".env")
import os
TAVILY_API_KEY = os.getenv("TAVILY_API_KEY")
TAVILY_MAX_RESULTS = int(os.getenv("TAVILY_MAX_RESULTS") or "5")
@dataclass
class SearchResult:
"""搜索结果数据类"""
title: str
url: str
snippet: str
source: str = "DuckDuckGo"
timestamp: datetime = None
def __post_init__(self):
if self.timestamp is None:
self.timestamp = datetime.now()
def test_tavily_api_key():
"""测试 API Key 配置"""
print("=" * 60)
print("测试 1: 检查 Tavily API Key")
print("=" * 60)
if TAVILY_API_KEY:
print(f"✓ TAVILY_API_KEY 已配置: {TAVILY_API_KEY[:15]}...")
else:
print("✗ TAVILY_API_KEY 未配置")
print()
def test_tavily_search_direct():
"""直接测试 Tavily API"""
print("=" * 60)
print("测试 2: 直接调用 Tavily API")
print("=" * 60)
if not TAVILY_API_KEY:
print("✗ 未配置 API Key跳过测试")
return
from tavily import TavilyClient
client = TavilyClient(api_key=TAVILY_API_KEY)
test_queries = [
"Python 编程语言最新版本",
"LangGraph AI 框架",
]
for query in test_queries:
print(f"\n搜索: {query}")
print("-" * 40)
try:
response = client.search(
query=query,
max_results=3,
include_answer=True,
include_raw_content=False
)
print(f"✓ 搜索成功")
print(f" - 结果数量: {len(response.get('results', []))}")
# 打印结果
for i, item in enumerate(response.get("results", []), 1):
print(f"\n [{i}] {item.get('title', '')}")
print(f" URL: {item.get('url', '')}")
print(f" 摘要: {item.get('content', '')[:100]}...")
# 如果有 answer
if response.get("answer"):
print(f"\n 🤖 AI 摘要: {response['answer'][:200]}...")
except Exception as e:
print(f"✗ 搜索失败: {e}")
print()
def test_web_search_integration():
"""测试 web_search 模块集成"""
print("=" * 60)
print("测试 3: 测试 web_search 模块集成")
print("=" * 60)
# 直接导入 web_search 模块(避免循环依赖)
web_search_path = project_root / "backend" / "app" / "core" / "web_search.py"
if not web_search_path.exists():
print(f"✗ web_search.py 不存在于 {web_search_path}")
return
print(f"✓ 找到 web_search.py: {web_search_path}")
# 使用 exec 动态加载模块
import importlib.util
spec = importlib.util.spec_from_file_location("web_search_module", web_search_path)
web_search_module = importlib.util.module_from_spec(spec)
try:
spec.loader.exec_module(web_search_module)
print("✓ web_search 模块加载成功")
except Exception as e:
print(f"✗ 模块加载失败: {e}")
return
# 测试搜索
print("\n执行搜索测试:")
try:
result = web_search_module.web_search("今天天气怎么样", max_results=3)
print(f"✓ 搜索成功,返回 {len(result)} 字符")
print("-" * 40)
print(result[:800] + "..." if len(result) > 800 else result)
except Exception as e:
print(f"✗ 搜索失败: {e}")
print()
def main():
print("\n" + "=" * 60)
print("🚀 Tavily 搜索功能测试")
print("=" * 60 + "\n")
test_tavily_api_key()
test_tavily_search_direct()
test_web_search_integration()
print("=" * 60)
print("✅ 测试完成")
print("=" * 60)
if __name__ == "__main__":
main()