from pandas import read_csv
from csv import reader
from sys import argv
from openpyxl import Workbook
from logging import getLogger, INFO
from commons import clibs

logger = getLogger(__file__)
logger.setLevel(INFO)


def find_point(bof, step, pos, data_file, flag, df, row, w2t):
    # bof: backward or forward
    # pos: used for debug
    # flag: greater than or lower than
    if flag == 'gt':
        while 0 < row < df.index[-1]-100:
            _value = df.iloc[row, 2]
            if _value > 2:
                if bof == 'backward':
                    row -= step
                elif bof == 'forward':
                    row += step
                continue
            else:
                if bof == 'backward':
                    row_target = row - step
                elif bof == 'forward':
                    row_target = row + step
                break
        else:
            if bof == 'backward':
                w2t(f"[{pos}] 在 {data_file} 中，无法正确识别数据，需要确认...", 0, 2, 'red')
            elif bof == 'forward':
                row_target = row + 100
    elif flag == 'lt':
        while 0 < row < df.index[-1]-100:
            _value = df.iloc[row, 2]
            if _value < 2:
                if bof == 'backward':
                    row -= step
                elif bof == 'forward':
                    row += step
                continue
            else:
                if bof == 'backward':
                    row_target = row - step
                elif bof == 'forward':
                    row_target = row + step
                break
        else:
            if bof == 'backward':
                w2t(f"[{pos}] 在 {data_file} 中，无法正确识别数据，需要确认...", 0, 3, 'red')
            elif bof == 'forward':
                row_target = row + 100

    return row_target


def get_cycle_info(data_file, df, row, step, w2t):
    # end -> middle: low
    # middle -> start: high
    # 1. 从最后读取数据，无论是大于1还是小于1，都舍弃，找到相反的值的起始点
    # 2. 从起始点，继续往前寻找，找到与之数值相反的中间点
    # 3. 从中间点，继续往前寻找，找到与之数值相反的结束点，至此，得到了高低数值的时间区间以及一轮的周期时间
    if df.iloc[row, 2] < 2:
        row = find_point('backward', step, 'a1', data_file, 'lt', df, row, w2t)

    _row = find_point('backward', step, 'a2', data_file, 'gt', df, row, w2t)
    _row = find_point('backward', step, 'a3', data_file, 'lt', df, _row, w2t)
    row_end = find_point('backward', step, 'a4', data_file, 'gt', df, _row, w2t)
    row_middle = find_point('backward', step, 'a5', data_file, 'lt', df, row_end, w2t)
    row_start = find_point('backward', step, 'a6', data_file, 'gt', df, row_middle, w2t)

    return row_end-row_middle, row_middle-row_start, row_end-row_start


def initialization(path, w2t):
    _, data_files = clibs.traversal_files(path, w2t)

    for data_file in data_files:
        if not data_file.lower().endswith('.csv'):
            w2t(f"{data_file} 文件后缀错误，只允许 .csv 文件，需要确认！", 0, 1, 'red')

    return data_files


def preparation(data_file, wb, w2t):
    shtname = data_file.split('\\')[-1].split('.')[0]
    ws = wb.create_sheet(shtname)
    csv_reader = reader(open(data_file))
    i = 0
    begin = 70
    for row in csv_reader:
        i += 1
        if i == 1:
            begin = int(row[1])
            break
    df = read_csv(data_file, sep=',', encoding='gbk', skip_blank_lines=False, header=begin - 1, on_bad_lines='warn')
    low, high, cycle = get_cycle_info(data_file, df, df.index[-1]-110, 5, w2t)

    return ws, df, low, high, cycle


def single_file_proc(ws, data_file, df, low, high, cycle, w2t):
    _row = _row_lt = _row_gt = count = 1
    _step = 5
    _data = {}
    row_max = df.index[-1]-100
    # print(data_file)
    while _row < row_max:
        if count not in _data.keys():
            _data[count] = []

        _value = df.iloc[_row, 2]
        if _value < 2:
            _row_lt = find_point('forward', _step, 'c'+str(_row), data_file, 'lt', df, _row, w2t)
            _start = int(_row_gt + (_row_lt - _row_gt - 50) / 2)
            _end = _start + 50
            value = df.iloc[_start:_end, 2].mean() + 3 * df.iloc[_start:_end, 2].std()
            _data[count].append(value)
        else:
            _row_gt = find_point('forward', _step, 'c'+str(_row), data_file, 'gt', df, _row, w2t)
            if _row_gt - _row_lt > cycle * 2:
                count += 1

        _row = max(_row_gt, _row_lt)

    for i in range(2, 10):
        ws.cell(row=1, column=i).value = f"第{i-1}次测试"
        ws.cell(row=i, column=1).value = f"第{i-1}次精度变化"

    # print(_data)
    for i in sorted(_data.keys()):
        _row = 2
        _column = i + 1
        for value in _data[i]:
            ws.cell(row=_row, column=_column).value = float(value)
            _row += 1


def execution(data_files, w2t):
    wb = Workbook()
    for data_file in data_files:
        ws, df, low, high, cycle = preparation(data_file, wb, w2t)
        # print(f"low = {low}")
        # print(f"high = {high}")
        # print(f"cycle = {cycle}")
        single_file_proc(ws, data_file, df, low, high, cycle, w2t)

    wd = data_files[0].split('\\')
    del wd[-1]
    wd = '\\'.join(wd)
    filename = wd + '\\result.xlsx'
    wb.save(filename)
    wb.close()
    w2t('----------------------------------------')
    w2t('所有文件均已处理完毕')


def main(path, w2t):
    data_files = initialization(path, w2t)
    execution(data_files, w2t)


if __name__ == '__main__':
    main(path=argv[1], w2t=argv[2])