Question

试图弄清楚所有这些如何工作之后，我设法将Bitmex的存储桶式API提要加载到DataFrame中。现在，尝试遵循我能找到的所有逻辑（包括文档），以了解如何将其解析为talib以吐出（正确的）数字，这就是我遇到的问题。我不需要绘制图形（尽管这很酷），但这是在以后的阶段。到目前为止，我有：

  import requests
  import numpy
  import talib
  import talib as ta
  import panda as pd

    dfURI = 'https://www.bitmex.com/api/v1/trade/bucketed?binSize=1h&partial=false&symbol=XBT&columns=timestamp%2C%20open%2C%20high%2C%20low%2C%20close%2C%20volume&count=96&reverse=true'
    df = pd.read_json(dfURI, orient="columns")
    df.columns = ["high", "close", "low", "open", "symbol", "timestamp", "volume"]
    df = df.set_index('timestamp')

    period = "4H" #1D or 1H
    data_rows = "1"

    if period == "4H" :
        data_period = df.resample('4H').mean()
        data_rows = 8
    elif period == "1D" :
        data_period = df.resample('1D').mean()
        data_rows = 2
    elif period == "1H" :
        data_period = df.resample('1H').mean()
        data_rows = 12

    #daily_close =  df.set_index('timestamp').resample('1D').mean()
    print (data_period.tail(data_rows))
    close = data_period['close'] #.values
    low = data_period['low'] #.values
    high = data_period['high'] #.values

    #real = talib.WILLR(high, low, close.values, timeperiod=14)
    realK, realD = talib.STOCHRSI(close)
    logger.info("TALib output is: \n %s %s " % (realK, realD))

JSON读取的输出是正确的（尽管它不是基于时间戳排序的，但稍后会介绍）

                     high     close       low      open        volume
timestamp                                                                
2018-09-10 12:00:00  6294.500  6315.375  6273.750  6292.125  1.052005e+08
2018-09-10 16:00:00  6268.250  6289.875  6243.875  6273.625  9.917725e+07
2018-09-10 20:00:00  6289.625  6298.625  6270.625  6288.000  4.804955e+07
2018-09-11 00:00:00  6337.125  6369.875  6301.875  6321.375  1.142635e+08
2018-09-11 04:00:00  6326.500  6342.250  6321.625  6331.875  5.114736e+07
2018-09-11 08:00:00  6309.500  6322.000  6297.250  6314.250  6.218389e+07
2018-09-11 12:00:00  6251.125  6286.625  6220.750  6276.250  1.322033e+08
2018-09-11 16:00:00  6247.500  6248.000  6180.000  6204.500  2.071930e+08

现在，当尝试使用talib.stochrsi（Close）或talib.willr（按代码中的设置）调用数据时，我得到了完全错误的数字（相对于交易视图）和NaN代码，就像这样：

2018-09-10 04:00:00          NaN
2018-09-10 08:00:00          NaN
2018-09-10 12:00:00          NaN
2018-09-10 16:00:00          NaN
2018-09-10 20:00:00          NaN
2018-09-11 00:00:00    52.523650
2018-09-11 04:00:00    76.982179
2018-09-11 08:00:00    75.890822
2018-09-11 12:00:00    46.515123
2018-09-11 16:00:00    22.056595

我在这里做错了什么？任何帮助将非常感激。很想获得正确的STOCHRSI值！

需要一些指导：将数据框架转换为TA-lib

0 个答案: