Python function 时间序列数据模板

Question

I want to process some array/list/timeseries data, and want to use many different filters for this.

This led to two problems: I don't want to copy-paste the function every time, especially if I change something. Also with different dependencies (there might be a dependency on the previous, or n-th previous element, or n-th following element), the array that is looped over can go out of bounds, if I don't adjust the ranges.

The conditions for the filters could be arbitrarily complex, but always involve relative position in the data.

Here is a minimal example:

import random as r
data = [r.random() for _ in range(100)]

def example_filter(data):
    counter = 0
    for i in range(1, len(data)):
        if((data[i-1]>0.8) and (data[i]<0.5)):
            counter +=1
            #might want to change something here
            #right now I would need to do this in all filters separately
    return counter

def example_filter_2(data):
    counter = 0
    for i in range(2, len(data)):
        if((data[i-2]>0.8) or ((data[i-1]>0.9) and (data[i]<0.2))):
            counter +=1
    return counter

My idea was to somehow compress the conditions (they are more complicated in the real example), use a converter function to make the real condition out of them, pass it as a string to a template function, and then use the condition, like this:

def filter_template(condition):
    def instance_of_filter(data):
        counter = 0
        #problem: the range isn't adjusted to account for out of bounds here
        for i in range(len(data)):
            #problem: condition will be passed as a string, so how can I evaluate it
            #also, I can't evaluate condition before I know what 'data' is, so I need to keep the dependency
            if condition:
                counter += 1
        return counter
    return instance_of_filter

Any ideas?

Answer 1

You can use your last code idea, just change the condition from variable to a predicate function based on data and index.

Example:

def filter_template(condition_func, start_at=0):
    def instance_of_filter(data):
        counter = 0
        for i in range(start_at, len(data)):
            if condition_func(data, i):
                counter += 1
        return counter
    return instance_of_filter


def condition1(data, i):
    return (data[i-1]>0.8) and (data[i]<0.5)

def condition2(data, i):
    return ((data[i-2]>0.8) or ((data[i-1]>0.9) and (data[i]<0.2)))

# usage
filter_template(condition1, 1)
filter_template(condition2, 2)