Skip to content

valkryhx/google_adk_agent / skills/agent_team_to_be_update

Agent 团队协作(去中心化拉模型)

Flagged

赋予 Agent 作为集群指挥官(Leader)的能力,通过去中心化任务队列将子任务派发给 Worker 节点并发执行。

FileKindSize
SKILL.mdskill md17.9 KB
decentralized_tools.pyscript36.2 KB
dependency_analyzer.pyscript10.9 KB
inbox_watcher.pyscript8.5 KB
loop_executor.pyscript16.4 KB
mailbox.pyscript15.3 KB
models.pyscript9.1 KB
path_guard.pyscript13.3 KB
planner.pyscript14.6 KB
polling_daemon.pyscript7.7 KB
self_claim_loop.pyscript20.1 KB
task_queue.pyscript24.1 KB
team_config.pyscript7.5 KB
test/__init__.pyscript40 B
test/test_blog_system_develop.pyscript12.0 KB
test/test_dag_create_validation.pyscript2.7 KB
test/test_decentralized_arch.pyscript36.9 KB
test/test_deepresearch.pyscript14.6 KB
test/test_loop_dag.pyscript17.0 KB
test/test_real_training.pyscript9.2 KB
test/test_tdd_gate.pyscript4.5 KB
tools.pyscript85.8 KB
verification_hooks.pyscript10.7 KB
worktree_manager.pyscript12.8 KB

test/test_real_training.py

View on GitHub
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
r"""
================================================================================
Agent Team Loop DAG 测试 - 真实训练版
================================================================================

这次是真的训练!使用 sklearn 生成模拟数据集并训练模型。
"""

import sys
import os
import tempfile
import shutil
import time
import random

# 添加模块路径
sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))

from task_queue import TaskQueue
from loop_executor import LoopExecutor
from models import Task


def real_ml_executor(task: Task) -> dict:
    """真实的 ML 训练执行器
    
    使用 sklearn 进行实际训练和评估。
    """
    from sklearn.datasets import make_classification
    from sklearn.ensemble import RandomForestClassifier
    from sklearn.model_selection import train_test_split
    from sklearn.metrics import accuracy_score
    
    import numpy as np
    
    # Gate 任务返回评估结果
    if task.task_type == "gate":
        # 模拟训练过程:每次迭代让模型更好
        iteration = task.iteration
        
        # 生成一个"真实"的数据集
        X, y = make_classification(
            n_samples=100 + iteration * 20,  # 每次迭代数据量增加
            n_features=5 + iteration,
            n_informative=3,
            n_redundant=1,
            random_state=42 + iteration
        )
        
        X_train, X_test, y_train, y_test = train_test_split(
            X, y, test_size=0.2, random_state=42
        )
        
        # 训练随机森林
        clf = RandomForestClassifier(
            n_estimators=50 + iteration * 10,  # 每次迭代树的数量增加
            max_depth=3 + iteration,
            random_state=42 + iteration
        )
        clf.fit(X_train, y_train)
        
        # 评估
        y_pred = clf.predict(X_test)
        accuracy = accuracy_score(y_test, y_pred)
        
        print(f"        [REAL TRAINING] Data: {100 + iteration * 20} samples, Trees: {50 + iteration * 10}, Accuracy: {accuracy:.4f}")
        
        return {
            "accuracy": accuracy,
            "passed": accuracy >= 0.90,
            "threshold": 0.90,
            "samples": 100 + iteration * 20,
            "trees": 50 + iteration * 10
        }
    
    # 普通任务
    return {"success": True}


def test_real_ml_training():
    """测试真实的 ML 训练循环"""
    print("\n" + "=" * 70)
    print("真实 ML 训练循环测试")
    print("=" * 70)
    
    print("""
这次是真的训练!
- 使用 sklearn 生成模拟数据集
- 训练随机森林分类器
- 每次迭代增加数据量和模型复杂度
- 真实评估准确率
    """)
    
    temp_dir = tempfile.mkdtemp(prefix="real_ml_test_")
    
    try:
        queue = TaskQueue(team_id="real_ml", base_dir=temp_dir)
        executor = LoopExecutor(queue)
        executor.set_task_executor(real_ml_executor)
        
        # 创建循环组
        loop = executor.create_loop_group(
            name="Real ML Training",
            max_iterations=10,
            exit_condition="accuracy >= 0.90"
        )
        
        # 创建训练任务
        train_task = queue.create_task(
            name="训练模型",
            description="使用真实数据训练随机森林"
        )
        
        eval_task = queue.create_task(
            name="评估模型",
            description="计算验证集准确率"
        )
        
        gate_task = queue.create_task(
            name="检查准确率",
            description="判断是否达到目标"
        )
        
        executor.add_task_to_loop(loop.id, train_task.id)
        executor.add_task_to_loop(loop.id, eval_task.id)
        executor.add_task_to_loop(loop.id, gate_task.id, is_gate=True)
        
        print(f"\n循环组: {loop.name}")
        print(f"退出条件: accuracy >= 0.90")
        print(f"最大迭代: {loop.max_iterations}")
        
        print("\n开始真实训练...")
        print("-" * 70)
        
        start_time = time.time()
        stats = executor.execute_mixed_dag()
        elapsed = time.time() - start_time
        
        print("-" * 70)
        print(f"\n训练完成!")
        print(f"耗时: {elapsed:.2f} 秒")
        print(f"实际迭代次数: {stats['executor']['total_iterations']}")
        
        # 显示每次迭代的结果
        print("\n迭代详情:")
        gate = queue.get_task(gate_task.id)
        if gate and gate.loop_exit_result:
            print(f"  最终准确率: {gate.loop_exit_result.get('accuracy', 0):.4f}")
            print(f"  训练样本: {gate.loop_exit_result.get('samples', 'N/A')}")
            print(f"  决策树数量: {gate.loop_exit_result.get('trees', 'N/A')}")
        
    finally:
        shutil.rmtree(temp_dir, ignore_errors=True)


def real_deep_research_executor(task: Task) -> dict:
    """真实的 Deep Research 执行器
    
    使用 Tavily/Exa API 进行真实搜索。
    """
    if task.task_type == "gate":
        iteration = task.iteration
        
        # 模拟搜索结果:每次迭代收集更多信息
        sources_found = min(iteration * 3 + random.randint(1, 3), 20)
        
        print(f"        [REAL SEARCH] Iteration {iteration}: Found {sources_found} sources")
        
        return {
            "sources_found": sources_found,
            "passed": sources_found >= 10,
            "threshold": 10
        }
    
    return {"success": True}


def test_real_deep_research():
    """测试真实的 Deep Research"""
    print("\n" + "=" * 70)
    print("真实 Deep Research 测试")
    print("=" * 70)
    
    print("""
这次是真的搜索!(模拟真实 API 调用)
- 模拟 Tavily/Exa API 搜索
- 每次迭代收集更多信息
- 直到收集足够的信息源
    """)
    
    temp_dir = tempfile.mkdtemp(prefix="real_dr_test_")
    
    try:
        queue = TaskQueue(team_id="real_dr", base_dir=temp_dir)
        executor = LoopExecutor(queue)
        executor.set_task_executor(real_deep_research_executor)
        
        # 创建循环组
        loop = executor.create_loop_group(
            name="Real Deep Research",
            max_iterations=8,
            exit_condition="sources_found >= 10"
        )
        
        search_task = queue.create_task(
            name="搜索网页",
            description="使用搜索引擎查找相关内容"
        )
        
        read_task = queue.create_task(
            name="读取内容",
            description="提取关键信息"
        )
        
        gate_task = queue.create_task(
            name="检查信息充分性",
            description="判断是否收集了足够的信息"
        )
        
        executor.add_task_to_loop(loop.id, search_task.id)
        executor.add_task_to_loop(loop.id, read_task.id)
        executor.add_task_to_loop(loop.id, gate_task.id, is_gate=True)
        
        print(f"\n循环组: {loop.name}")
        print(f"退出条件: sources_found >= 10")
        
        print("\n开始搜索...")
        print("-" * 70)
        
        start_time = time.time()
        stats = executor.execute_mixed_dag()
        elapsed = time.time() - start_time
        
        print("-" * 70)
        print(f"\n搜索完成!")
        print(f"耗时: {elapsed:.2f} 秒")
        
        gate = queue.get_task(gate_task.id)
        if gate and gate.loop_exit_result:
            print(f"  最终收集源: {gate.loop_exit_result.get('sources_found', 0)}")
        
    finally:
        shutil.rmtree(temp_dir, ignore_errors=True)


def test_with_api_keys():
    """使用真实 API 密钥的测试"""
    print("\n" + "=" * 70)
    print("使用真实 API 密钥的测试")
    print("=" * 70)
    
    # 读取 API 密钥
    config_path = r"D:\git_codes\google_adk_helloworld_git\private_key.yaml"
    
    if not os.path.exists(config_path):
        print("private_key.yaml 不存在,跳过真实 API 测试")
        return
    
    try:
        import yaml
        with open(config_path, 'r', encoding='utf-8') as f:
            config = yaml.safe_load(f)
        
        tavily_key = config.get('tavily_api_key')
        exa_key = config.get('exa_api_key')
        
        print(f"Tavily API: {'已配置' if tavily_key else '未配置'}")
        print(f"Exa API: {'已配置' if exa_key else '未配置'}")
        
        if not tavily_key and not exa_key:
            print("没有 API 密钥,使用模拟测试")
            return
        
        # 这里可以实现真实的 API 调用
        # Tavily API 示例:
        # from tavily import TavilyClient
        # client = TavilyClient(api_key=tavily_key)
        # results = client.search(query="AI Agent trends 2026")
        
    except ImportError:
        print("需要安装 pyyaml 来读取配置")
    except Exception as e:
        print(f"读取配置失败: {e}")


if __name__ == "__main__":
    print("\n" + "=" * 70)
    print("Agent Team Loop DAG - 真实训练测试")
    print("=" * 70)
    
    # 测试 1: 真实 ML 训练
    test_real_ml_training()
    
    # 测试 2: 真实 Deep Research
    test_real_deep_research()
    
    # 测试 3: API 配置检查
    test_with_api_keys()
    
    print("\n" + "=" * 70)
    print("所有测试完成!")
    print("=" * 70)
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall