aigroup-econ-mcp 0.4.2__py3-none-any.whl → 1.4.3__py3-none-any.whl

aigroup_econ_mcp/server.py

@@ -1,452 +0,0 @@
-"""
-AIGroup 计量经济学 MCP 服务器 - 优化版
-使用组件化架构，代码量减少80%，同时自动支持文件输入
-"""
-
-from typing import Dict, Any, Optional, List, Annotated
-from collections.abc import AsyncIterator
-from contextlib import asynccontextmanager
-from dataclasses import dataclass
-
-from pydantic import BaseModel, Field
-from mcp.server.fastmcp import FastMCP, Context
-from mcp.server.session import ServerSession
-from mcp.types import CallToolResult, TextContent
-
-# 导入工具处理器
-from .tools.tool_handlers import (
-    handle_descriptive_statistics,
-    handle_ols_regression,
-    handle_hypothesis_testing,
-    handle_time_series_analysis,
-    handle_correlation_analysis,
-    handle_panel_fixed_effects,
-    handle_panel_random_effects,
-    handle_panel_hausman_test,
-    handle_panel_unit_root_test,
-    handle_var_model,
-    handle_vecm_model,
-    handle_garch_model,
-    handle_state_space_model,
-    handle_variance_decomposition,
-    handle_random_forest,
-    handle_gradient_boosting,
-    handle_lasso_regression,
-    handle_ridge_regression,
-    handle_cross_validation,
-    handle_feature_importance
-)
-
-# 导入装饰器
-from .tools.base import with_file_support_decorator as econometric_tool
-
-
-# 应用上下文
-@dataclass
-class AppContext:
-    """应用上下文，包含共享资源"""
-    config: Dict[str, Any]
-    version: str = "0.2.0"
-
-
-@asynccontextmanager
-async def lifespan(server: FastMCP) -> AsyncIterator[AppContext]:
-    """服务器生命周期管理"""
-    config = {
-        "max_sample_size": 10000,
-        "default_significance_level": 0.05,
-        "supported_tests": ["t_test", "f_test", "chi_square", "adf"],
-        "data_types": ["cross_section", "time_series", "panel"]
-    }
-    try:
-        yield AppContext(config=config, version="0.2.0")
-    finally:
-        pass
-
-
-# 创建MCP服务器实例
-mcp = FastMCP(
-    name="aigroup-econ-mcp",
-    instructions="Econometrics MCP Server - Provides data analysis with automatic file input support",
-    lifespan=lifespan
-)
-
-
-# ============================================================================
-# 基础统计工具 (5个) - 自动支持文件输入
-# ============================================================================
-
-@mcp.tool()
-@econometric_tool('multi_var_dict')
-async def descriptive_statistics(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(default=None, description="CSV/JSON文件路径")] = None,
-    file_content: Annotated[Optional[str], Field(default=None, description="CSV/JSON文件内容")] = None,
-    file_format: Annotated[str, Field(default="auto", description="文件格式(csv/json/auto)")] = "auto",
-    data: Annotated[Optional[Dict[str, List[float]]], Field(default=None, description="数据字典(直接数据输入)")] = None
-) -> CallToolResult:
-    """
-    计算描述性统计量
-    
-    支持三种输入方式（按优先级）：
-    1. file_path: 文件路径 (如 "data.csv")
-    2. file_content: 文件内容字符串
-    3. data: 直接传入数据字典
-    """
-    return await handle_descriptive_statistics(ctx, data=data)
-
-
-@mcp.tool()
-@econometric_tool('regression')
-async def ols_regression(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(default=None, description="CSV/JSON文件路径")] = None,
-    file_content: Annotated[Optional[str], Field(default=None, description="CSV/JSON文件内容")] = None,
-    file_format: Annotated[str, Field(default="auto", description="文件格式")] = "auto",
-    y_data: Annotated[Optional[List[float]], Field(default=None, description="因变量(直接输入)")] = None,
-    x_data: Annotated[Optional[List[List[float]]], Field(default=None, description="自变量(直接输入)")] = None,
-    feature_names: Annotated[Optional[List[str]], Field(default=None, description="特征名称")] = None
-) -> CallToolResult:
-    """
-    OLS回归分析
-    
-    支持文件输入或直接数据输入。文件格式示例：
-    CSV: 最后一列为因变量，其余列为自变量
-    """
-    return await handle_ols_regression(ctx, y_data=y_data, x_data=x_data, feature_names=feature_names)
-
-
-@mcp.tool()
-@econometric_tool('single_var')
-async def hypothesis_testing(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(default=None, description="文件路径")] = None,
-    file_content: Annotated[Optional[str], Field(default=None, description="文件内容")] = None,
-    file_format: Annotated[str, Field(default="auto", description="文件格式")] = "auto",
-    data: Annotated[Optional[List[float]], Field(default=None, description="第一组数据")] = None,
-    data2: Annotated[Optional[List[float]], Field(default=None, description="第二组数据")] = None,
-    test_type: Annotated[str, Field(default="t_test", description="检验类型(t_test/adf)")] = "t_test"
-) -> CallToolResult:
-    """假设检验 - 支持文件或直接数据输入"""
-    return await handle_hypothesis_testing(ctx, data1=data, data2=data2, test_type=test_type)
-
-
-@mcp.tool()
-@econometric_tool('single_var')
-async def time_series_analysis(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(default=None, description="文件路径")] = None,
-    file_content: Annotated[Optional[str], Field(default=None, description="文件内容")] = None,
-    file_format: Annotated[str, Field(default="auto", description="文件格式")] = "auto",
-    data: Annotated[Optional[List[float]], Field(default=None, description="时间序列数据")] = None
-) -> CallToolResult:
-    """时间序列分析 - 支持文件或直接数据输入"""
-    return await handle_time_series_analysis(ctx, data=data)
-
-
-@mcp.tool()
-@econometric_tool('multi_var_dict')
-async def correlation_analysis(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(default=None, description="文件路径")] = None,
-    file_content: Annotated[Optional[str], Field(default=None, description="文件内容")] = None,
-    file_format: Annotated[str, Field(default="auto", description="文件格式")] = "auto",
-    data: Annotated[Optional[Dict[str, List[float]]], Field(default=None, description="多变量数据")] = None,
-    method: Annotated[str, Field(default="pearson", description="相关系数类型")] = "pearson"
-) -> CallToolResult:
-    """相关性分析 - 支持文件或直接数据输入"""
-    return await handle_correlation_analysis(ctx, data=data, method=method)
-
-
-# ============================================================================
-# 面板数据工具 (4个) - 自动支持文件输入
-# ============================================================================
-
-@mcp.tool()
-@econometric_tool('panel')
-async def panel_fixed_effects(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(
-        default=None, 
-        description="CSV文件路径。CSV格式要求：必须包含实体ID列(列名含entity_id/id/entity/firm/company/country/region之一)和时间列(列名含time_period/time/date/year/month/period/quarter之一)"
-    )] = None,
-    file_content: Annotated[Optional[str], Field(default=None)] = None,
-    file_format: Annotated[str, Field(default="auto")] = "auto",
-    y_data: Annotated[Optional[List[float]], Field(default=None)] = None,
-    x_data: Annotated[Optional[List[List[float]]], Field(default=None)] = None,
-    entity_ids: Annotated[Optional[List[str]], Field(default=None)] = None,
-    time_periods: Annotated[Optional[List[str]], Field(default=None)] = None,
-    feature_names: Annotated[Optional[List[str]], Field(default=None)] = None,
-    entity_effects: Annotated[bool, Field(default=True)] = True,
-    time_effects: Annotated[bool, Field(default=False)] = False
-) -> CallToolResult:
-    """固定效应模型 - 支持文件输入"""
-    return await handle_panel_fixed_effects(ctx, y_data, x_data, entity_ids, time_periods, 
-                                           feature_names, entity_effects, time_effects)
-
-
-@mcp.tool()
-@econometric_tool('panel')
-async def panel_random_effects(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(
-        default=None, 
-        description="CSV文件路径。CSV格式要求：必须包含实体ID列(列名含entity_id/id/entity/firm/company/country/region之一)和时间列(列名含time_period/time/date/year/month/period/quarter之一)"
-    )] = None,
-    file_content: Annotated[Optional[str], Field(default=None)] = None,
-    file_format: Annotated[str, Field(default="auto")] = "auto",
-    y_data: Annotated[Optional[List[float]], Field(default=None)] = None,
-    x_data: Annotated[Optional[List[List[float]]], Field(default=None)] = None,
-    entity_ids: Annotated[Optional[List[str]], Field(default=None)] = None,
-    time_periods: Annotated[Optional[List[str]], Field(default=None)] = None,
-    feature_names: Annotated[Optional[List[str]], Field(default=None)] = None,
-    entity_effects: Annotated[bool, Field(default=True)] = True,
-    time_effects: Annotated[bool, Field(default=False)] = False
-) -> CallToolResult:
-    """随机效应模型 - 支持文件输入"""
-    return await handle_panel_random_effects(ctx, y_data, x_data, entity_ids, time_periods,
-                                            feature_names, entity_effects, time_effects)
-
-
-@mcp.tool()
-@econometric_tool('panel')
-async def panel_hausman_test(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(
-        default=None, 
-        description="CSV文件路径。CSV格式要求：必须包含实体ID列(列名含entity_id/id/entity/firm/company/country/region之一)和时间列(列名含time_period/time/date/year/month/period/quarter之一)"
-    )] = None,
-    file_content: Annotated[Optional[str], Field(default=None)] = None,
-    file_format: Annotated[str, Field(default="auto")] = "auto",
-    y_data: Annotated[Optional[List[float]], Field(default=None)] = None,
-    x_data: Annotated[Optional[List[List[float]]], Field(default=None)] = None,
-    entity_ids: Annotated[Optional[List[str]], Field(default=None)] = None,
-    time_periods: Annotated[Optional[List[str]], Field(default=None)] = None,
-    feature_names: Annotated[Optional[List[str]], Field(default=None)] = None
-) -> CallToolResult:
-    """Hausman检验 - 支持文件输入"""
-    return await handle_panel_hausman_test(ctx, y_data, x_data, entity_ids, time_periods, feature_names)
-
-
-@mcp.tool()
-@econometric_tool('panel')  # 保持panel类型以获取entity_ids和time_periods
-async def panel_unit_root_test(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(
-        default=None, 
-        description="CSV文件路径。CSV格式要求：必须包含实体ID列(列名含entity_id/id/entity/firm/company/country/region之一)和时间列(列名含time_period/time/date/year/month/period/quarter之一)。数据量要求：至少3个实体，每个实体至少5个时间点"
-    )] = None,
-    file_content: Annotated[Optional[str], Field(default=None)] = None,
-    file_format: Annotated[str, Field(default="auto")] = "auto",
-    data: Annotated[Optional[List[float]], Field(default=None)] = None,
-    y_data: Annotated[Optional[List[float]], Field(default=None)] = None,  # 从panel转换来的
-    x_data: Annotated[Optional[List[List[float]]], Field(default=None)] = None,  # 从panel转换来的，忽略
-    entity_ids: Annotated[Optional[List[str]], Field(default=None)] = None,
-    time_periods: Annotated[Optional[List[str]], Field(default=None)] = None,
-    feature_names: Annotated[Optional[List[str]], Field(default=None)] = None,  # 从panel转换来的，忽略
-    test_type: Annotated[str, Field(default="levinlin")] = "levinlin"
-) -> CallToolResult:
-    """面板单位根检验 - 支持文件输入"""
-    # 传递所有参数给handler
-    return await handle_panel_unit_root_test(
-        ctx, 
-        data=data, 
-        y_data=y_data,
-        entity_ids=entity_ids, 
-        time_periods=time_periods, 
-        test_type=test_type
-    )
-
-
-# ============================================================================
-# 高级时间序列工具 (5个) - 自动支持文件输入
-# ============================================================================
-
-@mcp.tool()
-@econometric_tool('time_series')
-async def var_model_analysis(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(default=None, description="文件路径")] = None,
-    file_content: Annotated[Optional[str], Field(default=None)] = None,
-    file_format: Annotated[str, Field(default="auto")] = "auto",
-    data: Annotated[Optional[Dict[str, List[float]]], Field(default=None)] = None,
-    max_lags: Annotated[int, Field(default=5)] = 5,
-    ic: Annotated[str, Field(default="aic")] = "aic"
-) -> CallToolResult:
-    """VAR模型分析 - 支持文件输入"""
-    return await handle_var_model(ctx, data, max_lags, ic)
-
-
-@mcp.tool()
-@econometric_tool('time_series')
-async def vecm_model_analysis(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(default=None, description="文件路径")] = None,
-    file_content: Annotated[Optional[str], Field(default=None)] = None,
-    file_format: Annotated[str, Field(default="auto")] = "auto",
-    data: Annotated[Optional[Dict[str, List[float]]], Field(default=None)] = None,
-    coint_rank: Annotated[int, Field(default=1)] = 1,
-    deterministic: Annotated[str, Field(default="co")] = "co",
-    max_lags: Annotated[int, Field(default=5)] = 5
-) -> CallToolResult:
-    """VECM模型分析 - 支持文件输入"""
-    return await handle_vecm_model(ctx, data, coint_rank, deterministic, max_lags)
-
-
-@mcp.tool()
-@econometric_tool('single_var')
-async def garch_model_analysis(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(default=None, description="文件路径")] = None,
-    file_content: Annotated[Optional[str], Field(default=None)] = None,
-    file_format: Annotated[str, Field(default="auto")] = "auto",
-    data: Annotated[Optional[List[float]], Field(default=None)] = None,
-    order: Annotated[tuple, Field(default=(1, 1))] = (1, 1),
-    dist: Annotated[str, Field(default="normal")] = "normal"
-) -> CallToolResult:
-    """GARCH模型分析 - 支持文件输入"""
-    return await handle_garch_model(ctx, data, order, dist)
-
-
-@mcp.tool()
-@econometric_tool('single_var')
-async def state_space_model_analysis(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(default=None, description="文件路径")] = None,
-    file_content: Annotated[Optional[str], Field(default=None)] = None,
-    file_format: Annotated[str, Field(default="auto")] = "auto",
-    data: Annotated[Optional[List[float]], Field(default=None)] = None,
-    state_dim: Annotated[int, Field(default=1)] = 1,
-    observation_dim: Annotated[int, Field(default=1)] = 1,
-    trend: Annotated[bool, Field(default=True)] = True,
-    seasonal: Annotated[bool, Field(default=False)] = False,
-    period: Annotated[int, Field(default=12)] = 12
-) -> CallToolResult:
-    """状态空间模型分析 - 支持文件输入"""
-    return await handle_state_space_model(ctx, data, state_dim, observation_dim, trend, seasonal, period)
-
-
-@mcp.tool()
-@econometric_tool('time_series')
-async def variance_decomposition_analysis(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(default=None, description="文件路径")] = None,
-    file_content: Annotated[Optional[str], Field(default=None)] = None,
-    file_format: Annotated[str, Field(default="auto")] = "auto",
-    data: Annotated[Optional[Dict[str, List[float]]], Field(default=None)] = None,
-    periods: Annotated[int, Field(default=10)] = 10,
-    max_lags: Annotated[int, Field(default=5)] = 5
-) -> CallToolResult:
-    """方差分解分析 - 支持文件输入"""
-    return await handle_variance_decomposition(ctx, data, periods, max_lags)
-
-
-# ============================================================================
-# 机器学习工具 (6个) - 自动支持文件输入
-# ============================================================================
-
-@mcp.tool()
-@econometric_tool('regression')
-async def random_forest_regression_analysis(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(default=None, description="文件路径")] = None,
-    file_content: Annotated[Optional[str], Field(default=None)] = None,
-    file_format: Annotated[str, Field(default="auto")] = "auto",
-    y_data: Annotated[Optional[List[float]], Field(default=None)] = None,
-    x_data: Annotated[Optional[List[List[float]]], Field(default=None)] = None,
-    feature_names: Annotated[Optional[List[str]], Field(default=None)] = None,
-    n_estimators: Annotated[int, Field(default=100)] = 100,
-    max_depth: Annotated[Optional[int], Field(default=None)] = None
-) -> CallToolResult:
-    """随机森林回归 - 支持文件输入"""
-    return await handle_random_forest(ctx, y_data, x_data, feature_names, n_estimators, max_depth)
-
-
-@mcp.tool()
-@econometric_tool('regression')
-async def gradient_boosting_regression_analysis(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(default=None, description="文件路径")] = None,
-    file_content: Annotated[Optional[str], Field(default=None)] = None,
-    file_format: Annotated[str, Field(default="auto")] = "auto",
-    y_data: Annotated[Optional[List[float]], Field(default=None)] = None,
-    x_data: Annotated[Optional[List[List[float]]], Field(default=None)] = None,
-    feature_names: Annotated[Optional[List[str]], Field(default=None)] = None,
-    n_estimators: Annotated[int, Field(default=100)] = 100,
-    learning_rate: Annotated[float, Field(default=0.1)] = 0.1,
-    max_depth: Annotated[int, Field(default=3)] = 3
-) -> CallToolResult:
-    """梯度提升树回归 - 支持文件输入"""
-    return await handle_gradient_boosting(ctx, y_data, x_data, feature_names, n_estimators, learning_rate, max_depth)
-
-
-@mcp.tool()
-@econometric_tool('regression')
-async def lasso_regression_analysis(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(default=None, description="文件路径")] = None,
-    file_content: Annotated[Optional[str], Field(default=None)] = None,
-    file_format: Annotated[str, Field(default="auto")] = "auto",
-    y_data: Annotated[Optional[List[float]], Field(default=None)] = None,
-    x_data: Annotated[Optional[List[List[float]]], Field(default=None)] = None,
-    feature_names: Annotated[Optional[List[str]], Field(default=None)] = None,
-    alpha: Annotated[float, Field(default=1.0)] = 1.0
-) -> CallToolResult:
-    """Lasso回归 - 支持文件输入"""
-    return await handle_lasso_regression(ctx, y_data, x_data, feature_names, alpha)
-
-
-@mcp.tool()
-@econometric_tool('regression')
-async def ridge_regression_analysis(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(default=None, description="文件路径")] = None,
-    file_content: Annotated[Optional[str], Field(default=None)] = None,
-    file_format: Annotated[str, Field(default="auto")] = "auto",
-    y_data: Annotated[Optional[List[float]], Field(default=None)] = None,
-    x_data: Annotated[Optional[List[List[float]]], Field(default=None)] = None,
-    feature_names: Annotated[Optional[List[str]], Field(default=None)] = None,
-    alpha: Annotated[float, Field(default=1.0)] = 1.0
-) -> CallToolResult:
-    """Ridge回归 - 支持文件输入"""
-    return await handle_ridge_regression(ctx, y_data, x_data, feature_names, alpha)
-
-
-@mcp.tool()
-@econometric_tool('regression')
-async def cross_validation_analysis(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(default=None, description="文件路径")] = None,
-    file_content: Annotated[Optional[str], Field(default=None)] = None,
-    file_format: Annotated[str, Field(default="auto")] = "auto",
-    y_data: Annotated[Optional[List[float]], Field(default=None)] = None,
-    x_data: Annotated[Optional[List[List[float]]], Field(default=None)] = None,
-    feature_names: Annotated[Optional[List[str]], Field(default=None)] = None,
-    model_type: Annotated[str, Field(default="random_forest")] = "random_forest",
-    cv_folds: Annotated[int, Field(default=5)] = 5,
-    scoring: Annotated[str, Field(default="r2")] = "r2"
-) -> CallToolResult:
-    """交叉验证 - 支持文件输入"""
-    return await handle_cross_validation(ctx, y_data, x_data, model_type, cv_folds, scoring)
-
-
-@mcp.tool()
-@econometric_tool('regression')
-async def feature_importance_analysis_tool(
-    ctx: Context[ServerSession, AppContext],
-    file_path: Annotated[Optional[str], Field(default=None, description="文件路径")] = None,
-    file_content: Annotated[Optional[str], Field(default=None)] = None,
-    file_format: Annotated[str, Field(default="auto")] = "auto",
-    y_data: Annotated[Optional[List[float]], Field(default=None)] = None,
-    x_data: Annotated[Optional[List[List[float]]], Field(default=None)] = None,
-    feature_names: Annotated[Optional[List[str]], Field(default=None)] = None,
-    method: Annotated[str, Field(default="random_forest")] = "random_forest",
-    top_k: Annotated[int, Field(default=5)] = 5
-) -> CallToolResult:
-    """特征重要性分析 - 支持文件输入"""
-    return await handle_feature_importance(ctx, y_data, x_data, feature_names, method, top_k)
-
-
-def create_mcp_server() -> FastMCP:
-    """创建并返回MCP服务器实例"""
-    return mcp

aigroup_econ_mcp/tools/__init__.py

@@ -1,18 +0,0 @@
-"""
-计量经济学工具模块
-"""
-
-from . import regression, statistics, time_series, machine_learning, panel_data
-from . import validation, cache, monitoring, file_parser
-
-__all__ = [
-    "regression", 
-    "statistics", 
-    "time_series", 
-    "machine_learning", 
-    "panel_data",
-    "validation",
-    "cache", 
-    "monitoring",
-    "file_parser"
-]