def mode(mode):
# train 1 file
window_size = 20
if mode == "1":
# file = './transposeDataSet/bach/C_bach_850_format0.mid'
file = './transposeDataSet/bach/C_bach_847_format0.mid'
# file = './transposeDataSet/chopin/C_chpn_op10_e12_format0.mid'
# file = './transposeDataSet/chopin/C_chpn-p10_format0.mid'
# file = './new_song.mid'
print("Preparing Data.....")
prepare = PrepareData()
prepare.buildCorpus(file)
trainX, trainY = prepare.buildXY(window_size)
# print(trainX)
prepare.save_data()
m = Model()
m.createModel(trainX, trainY)
m.compileModel()
m.trainModel(trainX, trainY)
m.save_model()
m.testModel(trainX, trainY)
# model = "./model/model.json"
# weights = "./model/model.h5"
# m.load_model(model, weights)
# train all files
elif mode == "2":
print("Preparing Data.....")
prepare = PrepareData()
for file in glob.iglob('./transposeDataSet/bach/*.mid'):
prepare.buildCorpus(file)
trainX, trainY = prepare.buildXY(window_size)
prepare.save_data()
m = Model()
m.createModel(trainX, trainY)
m.compileModel()
m.trainModel(trainX, trainY)
m.save_model()
m.testModel(trainX, trainY)
# model = "./model/model.json"
# weights = "./model/model.h5"
# m.load_model(model, weights)
# test model
elif mode == "3":
# file = './transposeDataSet/bach/C_bach_850_format0.mid'
# file = './transposeDataSet/bach/C_bach_847_format0.mid'
# file = './transposeDataSet/chopin/C_chpn_op10_e12_format0.mid'
# file = './transposeDataSet/chopin/C_chpn-p10_format0.mid'
# file = "./SONG.mid"
for file in glob.iglob('./transposeDataSet/bach/*.mid'):
prepare = PrepareData()
prepare.buildCorpus(file)
testX, testY = prepare.buildXY(window_size)
m = Model()
model = "./model/model.json"
weights = "./model/model.h5"
m.load_model(model, weights)
m.compileModel()
m.testModel(testX, testY)
# transpose midi
elif mode == "4":
for file in glob.iglob('./DataSet/*/*.mid'):
t = Transpose(file)
t.transpose()
# predict
elif mode == "5":
prepare = PrepareData()
dataX = prepare.load_dataX()
# print(len(dataX))
dist_time1, dist_time2 = prepare.load_dist_time()
# print(dist_time)
m = Model()
model = "./model/model.json"
weights = "./model/model.h5"
m.load_model(model, weights)
p = Predictor(m.get_model())
p.predict_on_corpus(dataX, window_size)
type_and_note = p.buildNotetype()
song = p.buildSong(type_and_note, dist_time1, dist_time2)
midi = MidiMananger()
midi.creatMidiSong(song)
print("Created Song!! name: Song.midi")
else:
print("your selecting mode is not found!!!")
