import numpy as np
import matplotlib.pyplot as plt
import pandas as pd
from sklearn.neighbors import NearestNeighbors
from sklearn.metrics import roc_auc_score
from sklearn.linear_model import LogisticRegression
from scipy.io import loadmat
sys.path.insert(0, '../')
from ecg_AAAI.parse_dataset.readECG import loadECG
from ecg_AAAI.models.supervised.ecg_fi_model_keras import build_fi_model
from ecg_AAAI.models.supervised.ecg_fc import build_fc_model
from ecg_AAAI.models.gpu_utils import restrict_GPU_keras
from ecg_AAAI.models.supervised.ablation_helpers import *
restrict_GPU_keras("3")
import warnings
warnings.filterwarnings("error")
y_modes = ["mi", "cvd"]
splits = ["0", "1", "2", "3", "4"]
day_threshs = [30, 60, 90, 365]
pred_fs = [np.mean, np.median, top_10_mean, top_20_mean]
pred_f_names = ['mean', 'median', 'top_10_mean', 'top_20_mean']
n_unit_opts = [1, 2, 3]
instances = ['one', 'two', 'three', 'four']
fig_dir = "/home/divyas/ecg_AAAI/models/supervised/figs"
n_train_opts = [.1 * i for i in range(1, 10)]
batch_size = 90
import matplotlib.pyplot as plt
from sklearn.neighbors import NearestNeighbors
from sklearn.metrics import roc_auc_score
from sklearn.linear_model import LogisticRegression
from scipy.io import loadmat
sys.path.insert(0, '../')
from ecg_AAAI.parse_dataset.readECG import loadECG
from ecg_AAAI.models.ecg_utils import get_all_adjacent_beats
from ecg_AAAI.models.supervised.ecg_fi_model_keras import build_fi_model
from ecg_AAAI.models.supervised.ecg_fc import build_fc_model
from ecg_AAAI.models.gpu_utils import restrict_GPU_keras
from ecg_AAAI.models.supervised.eval import evaluate_AUC, evaluate_HR, risk_scores
import tftables
restrict_GPU_keras("1")
mode = sys.argv[1]
m_type = sys.argv[2]
y_mode = "cvd"
splits = ["0", "1", "2", "3", "4"]
split_num = "2"
split_dir = "./datasets/splits/split_" + split_num
# Load Y
# hf = h5py.File('datasets/data.h5', 'r')
# y_train = np.array(hf.get('y_train'))
# y_test = np.array(hf.get('y_test'))
# hf.close()
# Load Y
import matplotlib.pyplot as plt
import pandas as pd
from sklearn.neighbors import NearestNeighbors
from sklearn.metrics import roc_auc_score
from sklearn.linear_model import LogisticRegression
from scipy.io import loadmat
sys.path.insert(0, '../')
from ecg_AAAI.parse_dataset.readECG import loadECG
from ecg_AAAI.models.supervised.ecg_fi_model_keras import build_fi_model
from ecg_AAAI.models.supervised.ecg_fc import build_fc_model
from ecg_AAAI.models.supervised.ecg_cnn import build_cnn, build_small_f_cnn
from ecg_AAAI.models.gpu_utils import restrict_GPU_keras
from ecg_AAAI.models.supervised.ablation_helpers import *
restrict_GPU_keras("0")
import warnings
warnings.filterwarnings("error")
warnings.simplefilter("ignore", DeprecationWarning)
y_modes = ["mi", "cvd"]
splits = ["4", "3", "0", "1", "2"]
day_threshs = [365, 90, 30, 60]
pred_fs = [np.mean, np.median, top_10_mean, top_20_mean]
pred_f_names = ['mean', 'median', 'top_10_mean', 'top_20_mean']
instances = ['four', 'one', 'two', 'three', 'four']
split_prefix = "/home/divyas/ecg_AAAI/datasets/split_"
fig_dir = "/home/divyas/ecg_AAAI/models/supervised/figs"
model_name = "small_filter_cnn"
batch_size = 90