#!/usr/bin/env python3
import requests
from pyAudioAnalysis import audioTrainTest as aT
from pyAudioAnalysis.audioTrainTest import load_model, load_model_knn
from utils import config
from .abcpreprocessor import AbcPreProcessor

"""
Abstract base class for Sender
"""

__author__ = "@tormakris"
__copyright__ = "Copyright 2020, Birbnetes Team"
__module_name__ = "soundpreprocessor"
__version__text__ = "1"


class SoundPreProcessor(AbcPreProcessor):
    """
    SoundPreProcessor class, responsible for detecting birb chirps in sound sample.
    """

    def preprocesssignal(self, signal: str) -> bool:
        """
        Classify a sound sample.
        :param signal: Access path of the sound sample up for processing.
        :return:
        """
        # TODO: Dynamic model injection?
        r = requests.get(f"http://model-service/model/{config.MODEL_ID}/details")
        r.raise_for_status()

        model_details = r.json()

        if model_details['type'] == 'knn':
            classifier, mean, std, classes, mid_window, mid_step, short_window, short_step, compute_beat \
                = load_model_knn(config.MODEL_ID + "MEANS")

        else:
            classifier, mean, std, classes, mid_window, mid_step, short_window, short_step, compute_beat \
                = load_model(config.MODEL_ID + "MEANS")

        target_id = classes.index(config.TARGET_NAME)

        class_id, probability = aT.file_classification(signal, config.MODEL_ID, "svm")
        return class_id == target_id