ZJU_czx
133 lines
5.5 KiB
Python
133 lines
5.5 KiB
Python
from app.tool.python_execute import PythonExecute
|
|
|
|
|
|
class NormalPythonExecute(PythonExecute):
|
|
"""A tool for executing Python code with timeout and safety restrictions."""
|
|
|
|
name: str = "common_python_execute"
|
|
description: str = (
|
|
"""
|
|
Data Analysis Agent Protocol (Non-Visual) v2.1
|
|
|
|
=== Core Requirements ===
|
|
1. Strictly text-based outputs only
|
|
2. Dynamic analysis pipeline with memory
|
|
3. Context-aware processing
|
|
|
|
=== Execution Phases ===
|
|
|
|
1. CONTEXT INITIALIZATION
|
|
- Load historical analysis logs
|
|
- Build data quality baseline
|
|
- Detect previous processing patterns
|
|
|
|
2. ADAPTIVE PIPELINE
|
|
┌───────────────┬──────────────────────────────────────────────┐
|
|
│ Stage │ Enhanced Capabilities │
|
|
├───────────────┼──────────────────────────────────────────────┤
|
|
│ Data Loading │ Auto-select source based on history │
|
|
│ Cleaning │ Context-sensitive null/impute decision │
|
|
│ Transformation│ Dynamic feature engineering with validation │
|
|
│ Validation │ Cross-cycle consistency checks │
|
|
└───────────────┴──────────────────────────────────────────────┘
|
|
|
|
3. ITERATIVE PROCESSING CONTROLLER
|
|
Processing Loop:
|
|
while not convergence():
|
|
current_df = apply_operations(df)
|
|
delta = calculate_improvement(history[-1], current_df)
|
|
if delta < threshold: break
|
|
update_strategy_based_on(delta)
|
|
log_iteration(current_df)
|
|
|
|
Termination Criteria:
|
|
- 数据质量提升率 <2% 连续3次迭代
|
|
- 新增特征解释力 <5%
|
|
- 异常值比例稳定在 ±0.5% 区间
|
|
|
|
=== Enhanced Reporting ===
|
|
|
|
Output 1: dynamic_analysis.md (增量更新)
|
|
┌───────────────────────┬──────────────────────────────┐
|
|
│ Section │ Enhanced Requirements │
|
|
├───────────────────────┼──────────────────────────────┤
|
|
│ Processing History │ 记录每次迭代的操作及影响 │
|
|
│ Data Evolution │ 关键指标跨周期对比 │
|
|
│ Adaptive Findings │ 动态发现的模式变化 │
|
|
└───────────────────────┴──────────────────────────────┘
|
|
|
|
Output 2: intelligent_log.md (智能日志)
|
|
┌───────────────────────┬──────────────────────────────┐
|
|
│ Log Type │ Content │
|
|
├───────────────────────┼──────────────────────────────┤
|
|
│ Decision Log │ 策略调整原因及依据 │
|
|
│ Anomaly Evolution │ 异常值变化轨迹 │
|
|
│ Feature Lifecycle │ 衍生特征的产生/淘汰记录 │
|
|
└───────────────────────┴──────────────────────────────┘
|
|
|
|
=== Implementation Enhancements ===
|
|
|
|
1. Dynamic Code Generation
|
|
- 上下文感知的代码模板：
|
|
def analyze(data_path):
|
|
history = load_analysis_logs()
|
|
df = apply_historical_pipeline(data_path, history)
|
|
|
|
while not convergence_check(df, history):
|
|
df = context_aware_processing(df)
|
|
update_quality_metrics(df)
|
|
generate_incremental_report(df)
|
|
|
|
2. Memory Mechanism
|
|
历史记忆维度：
|
|
- 数据质量变化曲线
|
|
- 异常处理策略有效性
|
|
- 特征工程成功率
|
|
- 资源消耗模式
|
|
|
|
3. Intelligent Validation
|
|
验证增强点：
|
|
- 跨周期统计一致性检查
|
|
- 衍生特征可解释性评估
|
|
- 数据处理操作因果追踪
|
|
|
|
=== Sample Execution Flow ===
|
|
def analyze(data_path):
|
|
'''演进式分析流程'''
|
|
# 阶段1：上下文加载
|
|
df, ctx = initialize_context(data_path)
|
|
|
|
# 阶段2：智能处理循环
|
|
for i in range(MAX_ITERATIONS):
|
|
# 动态策略选择
|
|
ops = select_operations_based_on(ctx)
|
|
|
|
# 执行处理
|
|
df = execute_ops(df, ops)
|
|
|
|
# 生成增量报告
|
|
append_report(f"cycle_{i}_results.md", df)
|
|
|
|
# 收敛检测
|
|
if ctx.convergence_flag:
|
|
break
|
|
|
|
# 阶段3：知识固化
|
|
save_processing_knowledge(ctx)
|
|
"""
|
|
)
|
|
parameters: dict = {
|
|
"type": "object",
|
|
"properties": {
|
|
"code": {
|
|
"type": "string",
|
|
"default": "html",
|
|
"enum": ["process", "report", "others"],
|
|
},
|
|
},
|
|
"required": ["code"],
|
|
}
|
|
|
|
async def execute(self, code: str, code_type: str, timeout=5):
|
|
return await super().execute(code, timeout)
|