import os
import pandas as pd

# 定义文件夹路径
def data_get(folder_path, algo_pack_param):
    folder_path1 = f"{folder_path}/1"
    folder_path2 = f"{folder_path}/2"

    # 获取文件夹中的所有文件和文件夹
    all_files1 = os.listdir(folder_path1)
    all_files2 = os.listdir(folder_path2)
    all_data = pd.DataFrame()
    for file in all_files1:
        df = pd.read_excel(f"{folder_path1}/{file}")
        if not df.empty:
            all_data = all_data.append(df, ignore_index=True)
    for file in all_files2:
        df = pd.read_excel(f"{folder_path2}/{file}")
        if not df.empty:
            all_data = all_data.append(df, ignore_index=True)
    try:
        all_data['time'] = pd.to_datetime(all_data['time'], format='%Y-%m-%d %H:%M:%S')
    except:
        mask = all_data['time'].apply(lambda x: isinstance(x, str))
        all_data = all_data[~mask]
        all_data['time'] = pd.to_datetime(all_data['time'], format='%Y-%m-%d %H:%M:%S')
    all_data.sort_values(by=['time'], inplace=True)
    
    #数据处理
    CellVoltNums = len([col for col in df.columns if 'vol' in col])-6
    CellTempNums = len([col for col in df.columns if 'temp' in col])-3
    # CellVoltNums=int(algo_pack_param['CellVoltTotalCount'])
    # CellTempNums = int(algo_pack_param['CellTempTotalCount'])
    cellvolt_name=['cell_voltage'+str(x) for x in range(1, CellVoltNums+1)]
    celltemp_name=['cell_temp'+str(x) for x in range(1, CellTempNums+1)]
    all_data.columns = ['id','time','time1','pack_volt','pack_crnt','pack_soc'] + cellvolt_name + celltemp_name + \
                        ['diff_vol','cell_volt_max','max_vol_num','cell_volt_min','min_vol_num','diff_temp','cell_temp_max',
                         'max_temp_num','odo','insulation_resistance','speed','bms_sta','dcdc']
    all_data['cell_temp_min'] = all_data[celltemp_name].min(axis=1)
    all_data['pack_volt'] = all_data['pack_volt']/1000
    all_data['pack_crnt'] = all_data['pack_crnt']/1000
    return all_data, cellvolt_name, celltemp_name