enose_data_loader.py

import json
import numpy as np


def enose_data_load():
    ## Enose spikes
    model = "gas_minus_heater"
    delta = 0.1
    with open(f"data/events_train_{model}_{str(delta).replace('.', 'p')}.json", 'r', encoding='utf-8-sig') as fp:
        all_events_train = json.load(fp)
    with open(f"data/events_test_{model}_{str(delta).replace('.', 'p')}.json", 'r', encoding='utf-8-sig') as fp:
        all_events_test = json.load(fp)

    print(all_events_train.keys())


    labels2num = {
        'EB':0, 
        '2H':1, 
        'IA':2, 
        'Eu':3, 
        'blank':4,
    }

    y_train = []
    X_train = []
    for label, cycles in all_events_train.items():
        for cycle in cycles:
            sample = {"x":[], "t":[]}
            for i in range(8):
                sample["x"].extend([i for c in cycle[i]])
                sample["t"].extend([c for c in cycle[i]])

            index = np.argsort(sample["t"])
            sample["t"] = np.array(sample["t"])[index]
            sample["x"] = np.array(sample["x"])[index]
        
            y_train.append(labels2num[label])
            X_train.append(sample)

    y_test = []
    X_test = []
    for label, cycles in all_events_test.items():
        for cycle in cycles:
            sample = {"x":[], "t":[]}
            for i in range(8):
                sample["x"].extend([i for c in cycle[i]])
                sample["t"].extend([c for c in cycle[i]])

            index = np.argsort(sample["t"])
            sample["t"] = np.array(sample["t"])[index]
            sample["x"] = np.array(sample["x"])[index]
        
            y_test.append(labels2num[label])
            X_test.append(sample)
    y_train = np.array(y_train).astype(np.int8)
    X_train = np.array(X_train)
    y_test = np.array(y_test).astype(np.int8)
    X_test = np.array(X_test)

    print(len(y_train))
    print(len(X_train))        
    print(len(y_test))
    print(len(X_test))
    return (X_train, y_train, X_test, y_test)

if __name__ == '__main__':
    X_train, y_train, X_test, y_test = enose_data_load()
    print(y_train)