Source code for neurokin.utils.features_extraction.core

from abc import ABC, abstractmethod
from typing import List, Union

from typeguard import typechecked

from neurokin.constants.features_extraction import VALID_EXTRACTION_TARGETS




class FeatureExtraction(ABC):

    def __new__(cls, *args, **kwargs):
        if not isinstance(cls.extraction_target, str):
            raise TypeError(f"Extraction target type not valid, expected str got {type(cls.extraction_target)}")

        if cls.extraction_target not in VALID_EXTRACTION_TARGETS:
            raise ValueError(
                f"{cls.extraction_target} in not a valid target. Valid targets are "
                f"{', '.join(str(i) for i in VALID_EXTRACTION_TARGETS)}")
        return super().__new__(cls)



    @abstractmethod
    def compute_feature(self):
        pass




    def run_feat_extraction(self, *args, **kwargs):
        feat = self.compute_feature(*args, **kwargs)
        return feat






class MyFeat(FeatureExtraction):
    extraction_target = "markers"

    @typechecked
    def __init__(self, param_a: List[Union[int, str]], param_b: List[int]):
        self.param_a = param_a
        self.param_b = param_b



    def compute_feature(self):
        print("ehi look im doing cool stuff")
        return






class MyFeat2(FeatureExtraction):
    extraction_target = "joints"

    def __init__(self, param_a: ABC, param_b: List[int]):
        self.param_a = param_a
        self.param_b = param_b



    def compute_feature(self):
        pass




if __name__ == "__main__":
    feat = MyFeat(param_a=[1, "str"], param_b=[1, 2, 3])
    feat2 = MyFeat2(param_a=3.0, param_b=[1, 2, 3])
    # feat2.run_feat_extraction()
    feat.run_feat_extraction()
    print(feat.extraction_target)
    # feat.extraction_target = "joints"
    print(feat2.extraction_target)
    # print(feat.extraction_target)