from pandas import read_csv
from numpy import array
from numpy import argmax
from pandas import concat
from pandas import DataFrame
from keras.models import Sequential
from keras.layers import LSTM
from keras.layers import Dense

# generate a sequence of random numbers in [0, 99]
pitches = read_csv('goro.csv', squeeze=True)

print(pitches)

# one hot encode sequence
def one_hot_encode(sequence, n_unique=1000):
    encoding = list()
    for value in sequence:
        vector = [0 for _ in range(n_unique)]
        vector[value] = 1
        encoding.append(vector)
    return array(encoding)

# decode a one hot encoded string
def one_hot_decode(encoded_seq):
    return [argmax(vector) for vector in encoded_seq]

# generate data for the lstm
def generate_data():
    # generate sequence
    sequence = pitches()
    # one hot encode
    encoded = one_hot_encode(sequence)
    # create lag inputs
    df = DataFrame(encoded)
    df = concat([df.shift(4), df.shift(3), df.shift(2), df.shift(1), df], axis=1)
    # remove non-viable rows
    values = df.values
    values = values[5:,:]
    # convert to 3d for input
    X = values.reshape(len(values), 5, 1000)
    # drop last value from y
    y = encoded[4:-1,:]
    return X, y

# define model
model = Sequential()
model.add(LSTM(50, batch_input_shape=(5, 5, 1000), stateful=True))
model.add(Dense(100, activation='softmax'))
model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
# fit model
for i in range(2000):
    X, y = generate_data()
    model.fit(X, y, epochs=1, batch_size=5, verbose=2, shuffle=False)
    model.reset_states()
# evaluate model on new data
X, y = generate_data()
yhat = model.predict(X, batch_size=5)
print('Expected:  %s' % one_hot_decode(y))
print('Predicted: %s' % one_hot_decode(yhat))