use tulipy and accept an array of technical indicators as an argument (feature request)

[ckhung]

Hi Daniel, thank you for taking over the maintainance of mpl-finance. I agree with you about the new mission statement. Here is a feature request for you to consider. The tulipy project already implemented lots of technical indicators. It would be nice to be able to use them in mpl_finance. Here is a short piece showing how I would like to be able to pass my chosen technical indicators as arguments to candlestick_ohlc():

#!/usr/bin/python3

import numpy as np
import tulipy as ti
from datetime import datetime

def candlestick_ohlc(quotes, ti_array):
    close_prices = quotes['Close'].copy(order='C')
    for ti in ti_array:
        f = ti['f']
        del ti['f']
        print(f.full_name, f(close_prices, **ti))
        # of course in real code it should be plotting here.

dt = [float] * 6
dt[0] = 'datetime64[D]'
convertfunc = lambda x: datetime.strptime(x.decode('utf8'), '%m/%d/%Y')
daily = np.genfromtxt('SP500_NOV2019_Hist.csv',
    delimiter=',', autostrip=True, names=True, dtype=dt,
    converters = {'Date': convertfunc}
)
candlestick_ohlc(daily, [
    {'f':ti.sma, 'period':5},
    {'f':ti.sma, 'period':10},
    {'f':ti.macd, 'short_period':2, 'long_period':5, 'signal_period':9}
] )

Maybe we also need room for options telling candlestick_ohlc what colors and line width, etc. to use for each indicator line. And/or whether some of them should be displayed as solid bar and/or on a separate pane (so that volume can be processed using exactly this same interface without requiring a distinctive/separate interface). I don't have a very good suggestion yet as to how best to design that part of the interface.

[DanielGoldfarb]

Chao-Kuei, Thank you so much for the idea. I had never heard of tulipy, and I like this idea very much. I will try in the next few weeks to make some time to experiment with the code you have provided, and with tulipy in general, and see if I can prototype something for you to play with.

Regarding your other ideas, Volume is already available via the same interface and on a separate pane, and the ability to customize more of the plot is also on my list of things to do. Again, thanks so much for the idea regardy tulipy. I will definitely be looking into it.

[DanielGoldfarb]

The are now a number of ways to add technical indicators to mplfinance plots, most of which are outlined here: https://github.com/matplotlib/mplfinance/blob/master/examples/addplot.ipynb

These techniques allow many different source of technical indicators to be plotted in mplfinance, including tulipy, pandas-ta, and others.

I would prefer not to integrate directly into mplfinance and thereby create unnecessary dependencies.