China Stock MCP

Instructions

Implementation Reference

• src/china_stock_mcp/server.py:288-311 (handler)

  The main handler function for the 'get_realtime_data' tool, including registration decorator (@mcp.tool), input schema (Annotated parameters), and execution logic. It uses fallback data sources to fetch realtime stock data via the external 'ako.get_realtime_data' function and formats the output.
  @mcp.tool( name="get_realtime_data", description="获取指定的股票实时行情数据" ) def get_realtime_data( symbol: Annotated[str, Field(description="股票代码 (例如: '000001')")], output_format: Annotated[ Literal["json", "csv", "xml", "excel", "markdown", "html"], Field(description="输出数据格式: json, csv, xml, excel, markdown, html。默认: markdown"), ] = "markdown" ) -> str: """获取实时股票行情数据. 'eastmoney_direct' """ # 定义内部 fetch_func def realtime_data_fetcher(source: str, **kwargs: Any) -> pd.DataFrame: return ako.get_realtime_data(source=source, **kwargs) df = _fetch_data_with_fallback( fetch_func=realtime_data_fetcher, primary_source="eastmoney", fallback_sources=["eastmoney_direct","xueqiu"], symbol=symbol, ) return _format_dataframe_output(df, output_format)
• src/china_stock_mcp/server.py:15-72 (helper)

  Helper function that implements fallback logic for fetching data from multiple sources (eastmoney, eastmoney_direct, xueqiu), used by the get_realtime_data handler.
  def _fetch_data_with_fallback( fetch_func: Callable[..., pd.DataFrame], primary_source: str, fallback_sources: List[str], **kwargs: Any, ) -> pd.DataFrame: """ 通用的数据源故障切换辅助函数。 按优先级尝试数据源，直到获取到有效数据或所有数据源都失败。 Args: fetch_func: 实际调用 akshare 或 akshare_one 获取数据的函数。 这个函数应该接受 'source' 参数，或者在内部处理 source 的映射。 primary_source: 用户指定的首选数据源。 fallback_sources: 备用数据源列表，按优先级排序。 **kwargs: 传递给 fetch_func 的其他参数。 Returns: pd.DataFrame: 获取到的数据。 Raises: RuntimeError: 如果所有数据源都未能获取到有效数据。 """ if primary_source is None: return fetch_func(**kwargs) data_source_priority = [primary_source] + fallback_sources # 移除重复项并保持顺序 seen = set() unique_data_source_priority = [] for x in data_source_priority: if x not in seen: unique_data_source_priority.append(x) seen.add(x) df = None errors = [] for current_source in unique_data_source_priority: try: # 假设 fetch_func 能够接受 source 参数 # 或者 fetch_func 内部根据 kwargs 中的 source 参数进行逻辑判断 temp_df = fetch_func(source=current_source, **kwargs) if temp_df is not None and not temp_df.empty: print(f"成功从数据源 '{current_source}' 获取数据。") df = temp_df break else: errors.append(f"数据源 '{current_source}' 返回空数据。") except Exception as e: errors.append(f"从数据源 '{current_source}' 获取数据失败: {str(e)}") if df is None or df.empty: raise RuntimeError( f"所有数据源都未能获取到有效数据。详细错误: {'; '.join(errors)}" ) return df
• src/china_stock_mcp/server.py:89-118 (helper)

  Helper function to format the fetched DataFrame into various output formats (json, csv, etc.), called by the handler.
  def _format_dataframe_output( df: pd.DataFrame, output_format: Literal["json", "csv", "xml", "excel", "markdown", "html"], ) -> str: """ 根据指定的格式格式化 DataFrame 输出。 """ if df.empty: return json.dumps([]) if output_format == "json": return df.to_json(orient="records", force_ascii=False) elif output_format == "csv": return df.to_csv(index=False) elif output_format == "xml": return df.to_xml(index=False) elif output_format == "excel": # 使用 BytesIO 将 Excel 写入内存 output = io.BytesIO() df.to_excel(output, index=False, engine='openpyxl') # 返回 base64 编码的二进制数据，或者直接返回字节流 # 为了兼容性，这里尝试返回 utf-8 编码的字符串，但对于二进制文件，通常直接传输字节流更合适 return output.getvalue().decode("utf-8", errors="ignore") elif output_format == "markdown": return df.to_markdown(index=False) elif output_format == "html": return df.to_html(index=False) else: return df.to_json(orient="records", force_ascii=False)
• src/china_stock_mcp/server.py:288-290 (registration)

  The @mcp.tool decorator registers the get_realtime_data function as an MCP tool.
  @mcp.tool( name="get_realtime_data", description="获取指定的股票实时行情数据" )