Own indicator: "if function" with min max is very slow

Jonny8

Hi everyone,

I am having a hard time creating / adapting an indicator.
I made it in the most slow way and I wonder whether there is a better solution.

The idea is to have an if positive RoC the 1 else 0.

class Momentumplus_org(bt.Indicator):
    lines = ('trend',)
    params = (('period', 190), ('rperiod', 30))
    
    def __init__(self):
        self.addminperiod(self.params.period)
        self.roc = bt.ind.ROC(self.data, period=self.p.rperiod)

    def next(self):
        returns = np.log(self.data.get(size=self.p.period))
        x = np.arange(len(returns))
        slope, _, rvalue, _, _ = linregress(x, returns)
        annualized = (1 + slope) ** 252
        test = math.ceil(self.roc[0])
        rate = bt.Min(1, bt.Max(0, test))
        self.lines.trend[0] = annualized * (rvalue ** 2) * rate

This takes a hell of a time to get calculated. I tried to get the test and rate into the init ...

I also tried to keep the original design:

def momentum_func(the_array):
    r = np.log(the_array)
    slope, _, rvalue, _, _ = linregress(np.arange(len(r)), r)
    annualized = (1 + slope) ** 252
    return annualized * (rvalue ** 2)


class Momentum(bt.ind.OperationN):
    lines = ('trend',)
    params = dict(period=50)
    func = momentum_func

But I couldn't make it work with two parameters...

Are there better ways than math.ceil / min / max?

Thank you very much in advance!

Best

crunchypickle

@Jonny8 said in Own indicator: "if function" with min max is very slow:

test

are you sure it is slow where you said it was? Have you profiled the code?

Jonny8

@crunchypickle

Good call.

I used cProfile to do it. https://towardsdatascience.com/how-to-profile-your-code-in-python-e70c834fad89

There is a execution difference of 20% but it does not seem to come from min max or ceil, I assume it is due to the roc calculation in the init.

Thank you chrunchypickle!