#!/usr/bin/env python
# -*- coding: utf-8 -*-
# time: 2023/5/11 14:43
# file: cluster_power.py
# author: David
# company: shenyang JY

import os
import re
import numpy as np
import pandas as pd


def read_cfs(cfs, input_path, output_path, is_folder=False):
    if not os.path.exists(output_path):
        os.makedirs(output_path)
    dfs = {}
    for j, ids in cfs.items():
        if is_folder:
            dirname = input_path.split('/')[-1]
            x = re.findall('(?<=Continuous_Turbine_Data_).*?(?=_)',dirname)[0]
            dfs_j = [pd.read_csv(os.path.join(input_path, f"turbine-{id}_{int(x)}.csv")) for id in ids]
        else:
            dfs_j = [pd.read_csv(os.path.join(input_path, f"turbine-{id}.csv")) for id in ids]
        dfj, time_series = dfs_j[0].loc[:, ['C_TIME', 'C_WS', 'C_ACTIVE_POWER']], dfs_j[0]['C_TIME']
        for df in dfs_j[1:]:
            if df['C_TIME'].equals(time_series) is False:
                print("风机之间的日期不一致！")
                raise ValueError
            dfj['C_ACTIVE_POWER'] += df['C_ACTIVE_POWER']
            dfj['C_WS'] += df['C_WS']
        dfj['C_WS'] /= len(dfj)
        dfj.rename(columns=({'C_ACTIVE_POWER':'C_ACTIVE_POWER'+str(j), 'C_WS': 'C_WS'+str(j)}), inplace=True)
        if is_folder:
            dfj.to_csv(os.path.join(output_path, 'cluster_' + str(j) + '.csv'), index=False)
        else:
            dfj[20:].to_csv(os.path.join(output_path, 'cluster_' + str(j) + '.csv'), index=False)
        dfs[j] = dfj
    return dfs

def normalize(input_path, turbine):
    turbines = [pd.read_csv(f) for f in os.listdir(os.path.join(input_path))]
    turbines = [turbine.values[:, 1:].astype(np.float32) for turbine in turbines]
    turbines = np.vstack(turbines)
    mean, std = np.mean(turbines, axis=0), np.std(turbines, axis=0)

    c_time = turbine['C_TIME']
    turbine = (turbine.iloc[:, 1:] - mean) / std
    turbine.insert(loc=0, column='C_TIME', value=c_time)
    return turbine

def get_cfs(cluster, turbine_id):
    cfs = {}
    for j in range(1, max(cluster) + 1):
        arr_j = np.where(cluster == j)[0]  # cluster中聚类j的索引列表
        cfs.setdefault(j, [turbine_id[k] for k in arr_j])
    for key, value in cfs.items():
        print("第{}组：{}".format(key, cfs[key]))
    return cfs


def cluster_power_list_file(cluster, turbine_id, input_path, output_path):
    """
    从turbine-*.csv的文件列表中进行聚类功率相加
    cluster:聚类的结果
    turbine_id:风机ID
    input_path:输入路径 output_filtered_csv_files 所在路径
    output_path:输出每个聚类的功率，和所有聚类的功率cluster_data
    """
    if not os.path.exists(output_path):
        os.makedirs(output_path)

    cfs = get_cfs(cluster, turbine_id)
    dfs = read_cfs(cfs, input_path, output_path)
    dfs_cluster = pd.concat([df.set_index("C_TIME") for df in dfs.values()], join='inner', axis=1)
    dfs_cluster.reset_index().to_csv(os.path.join(output_path, 'cluster_data.csv'), index=False)


def cluster_power_list_folder(cluster, turbine_id, input_path, output_path):
    """
    从嵌套turbine-*.csv的多个文件夹列表中进行聚类功率相加
    cluster:聚类的结果
    turbine_id:风机ID
    input_path:输入路径 continuous_data 所在路径
    output_path:输出每个聚类的功率，和所有聚类的功率cluster_data
    """
    if not os.path.exists(output_path):
        os.makedirs(output_path)
    continuous_list = [os.path.join(input_path, path) for path in os.listdir(input_path)]
    cfs = get_cfs(cluster, turbine_id)
    for con in continuous_list:
        dirname = con.split('/')[-1]
        output = os.path.join(output_path, dirname)
        dfs = read_cfs(cfs, con, output, True)
        dfs_cluster = pd.concat([df.set_index("C_TIME") for df in dfs.values()], join='inner', axis=1)
        dfs_cluster.reset_index().to_csv(os.path.join(output, 'cluster_data.csv'), index=False)



def paint_cluster_power(cluster_path, dfs):
    dfs_cluster = pd.merge(dfs.values, axis=1, join='inner', on='C_TIME')
    dfs = [pd.read_csv(os.path.join(cluster_path, file_path)).rename(columns={'C_ACTIVE_POWER':'power_'+file_path.split('/')[-1][-5], 'C_WS': 'C_WS_'+file_path.split('/')[-1][-5]}) for file_path in os.listdir(cluster_path)]
    df_cluster = pd.DataFrame({df.columns[-1]: df.iloc[:, -1] for df in dfs})
    df_cluster.insert(loc=0, column='C_TIME', value=dfs[0]['C_TIME'])
    df_cluster.insert(loc=len(dfs)+1, column='SUM', value=df_cluster.iloc[:, 1:].sum(axis=1))
    df_cluster.to_csv(os.path.join(cluster_path, 'cluster_data.csv'), index=False)