import keras.backend as K
import numpy as np

inp = Input(shape=(x_train_show.shape[1],1000))
nnet = Dense(64, activation='relu')(inp)
nnet = Dropout(0.1)(nnet)
nnet = Dense(128, activation='relu')(nnet)
output = Dense(1, activation='sigmoid')(nnet)
model = Model(inputs=inp, outputs=output)

spec_loss = specificity_loss_wrapper()
model.compile(loss=spec_loss,optimizer=Adam(lr=0.00001))


def compute_specificity(y_pred, y_true):
    """Compute the confusion matrix for a set of predictions.

    Parameters
    ----------
    y_pred   : predicted values for a batch if samples (must be binary: 0 or 1)
    y_true   : correct values for the set of samples used (must be binary: 0 or 1)

    Returns
    -------
    out : the specificity
    """

    num_classes = 2
    conf_mat = np.zeros((num_classes, num_classes))

    for i in range(K.flatten(y_true).shape[0].value):
        conf_mat[y_true[i]][y_pred[i]] += 1

    # Convert the conf_mat to a Keras Tensor
    conf_mat = K.variable(conf_mat, dtype=np.int32))

    # Compute the actual specificity
    specificity = conf_mat[0,0] / (conf_mat[0,0] + conf_mat[0, 1])

    # Could also compute other metrics from the conf_mat, such as:
    # sensitivity, false discovery rate, F1 score, Jaccard index etc.

    return specificity

def specificity_loss_wrapper():
    """A wrapper to create and return a function which computes the specificity loss, as (1 - specificity)

    """
    # Define the function for your loss
    def specificity_loss(y_true, y_pred):
        return 1.0 - compute_specificity(y_true, y_pred)

    return specificity_loss    # we return this function object