def main():
allSong, targetList = dataset.main()
allSong = np.array(allSong)
targetList = np.array(targetList)
allSong = listNormalizer(allSong)
model(allSong, targetList)
def predictionSong(songUri: str):
# Girilin uri "spotify:track:URI" şeklinde ise URI kısmını ayırmak için kullanılır.
if songUri.find("spotify") != -1:
songUri = songUri[14:]
# yeni şarkının veri setine eklenmesi ve normalize edilmesi
artistName, songName, songInfo = SpotifyConnection.getSongInfo(songUri)
allSong, targetList = dataset.main()
allSong.append(songInfo)
allSong = np.array(allSong)
allSong = allSong / allSong.max(axis=0)
# Modelin hazırlanması
model = readModelFromJSON()
#veri setinden şarkının alınması ve tahmini
mysong = allSong[-1:]
predict = model.predict(mysong)
print("*************HIT SONG PREDICTION*************")
print("Song Name: " +songName)
print("Artist Name:" +artistName)
print("Hit Rate: %" + "%.2f" % (predict[0][0] * 100))
if predict >= 0.7:
print("SONG IS HIT!")
else:
print("SONG IS NOT HIT!")
def main():
parser = argparse.ArgumentParser(description='PyTorch Kaggle')
parser.add_argument('--jobtype', '-M', type=str, default='evaluate', help='what are you going to do on this function')
parser.add_argument('--setting', '-S', type=str, default='setting1', help='which setting file are you going to use for training.')
#parser.add_argument('--model', '-M', type=str, default='', help='')
args = parser.parse_args()
#cfg = setting.Config()
cfg = cfgs.get(args.setting)
if args.jobtype == 'preprocess':
dataset.main(cfg)
if args.jobtype == 'train':
train.main(cfg)
if args.jobtype == 'evaluate':
evaluate.main(cfg)
def predictionSong(songUri: str, songNameLabel: Label, artistNameLabel: Label,
hitRateLabel: Label, resultLabel: Label,
decisionHitLabel: Label, decisionResult: Label,
knnHitLabel: Label, knnResult: Label, svmHitLabel: Label,
svmResult: Label, bayesHitLabel: Label, bayesResult: Label):
if songUri.find("spotify") != -1:
songUri = songUri[14:]
# yeni şarkının veri setine eklenmesi ve normalize edilmesi
artistName, songName, songInfo = SpotifyConnection.getSongInfo(songUri)
allSong, targetList = dataset.main()
allSong.append(songInfo)
allSong = np.array(allSong)
allSong = allSong / allSong.max(axis=0)
# veri setinden şarkının alınması ve tahmini
mysong = allSong[-1:]
print("*************HIT SONG PREDICTION*************")
print("Song Name: " + songName)
print("Artist Name:" + artistName)
songNameLabel['text'] = "ŞARKI İSMİ: " + songName
artistNameLabel['text'] = "SANATÇI İSMİ: " + artistName
predict, predictions, result = neuralPrediction(mysong)
hitRateLabel['text'] = "POPÜLER OLMA İHTİMALİ: % " + "%.2f" % (
predict[0][0] * 100) + " " + str(predictions[0])
resultLabel['text'] = "SONUÇ: " + result
print("Hit Rate: %" + "%.2f" % (predict[0][0] * 100))
predict, result = decisionTreePrediction(mysong)
decisionHitLabel['text'] = "POPÜLERLİK : " + str(predict)
decisionResult['text'] = "SONUÇ: " + result
predict, result = knnPrediction(mysong)
knnHitLabel['text'] = "POPÜLERLİK : " + str(predict)
knnResult['text'] = "SONUÇ: " + result
predict, result = svmPrediction(mysong)
svmHitLabel['text'] = "POPÜLERLİK : " + str(predict)
svmResult['text'] = "SONUÇ: " + result
predict, result = bayesPrediction(mysong)
bayesHitLabel['text'] = "POPÜLERLİK : " + str(predict)
bayesResult['text'] = "SONUÇ: " + result
def predictionSong():
songUri = "spotify:track:1aaI0imelqLqye35922oMD"
if songUri.find("spotify") != -1:
songUri = songUri[14:]
artistName, songName, songInfo = SpotifyConnection.getSongInfo(songUri)
allSong, targetList = dataset.main()
allSong.append(songInfo)
allSong = np.array(allSong)
allSong = allSong / allSong.max(axis=0)
mySong = allSong[-1:]
model = joblib.load('BNB.pkl', mmap_mode='r')
y_pred = model.predict(mySong)
print(y_pred)
print("Sanatçı:" + artistName)
print("Şarkı Adı:" + songName)
if (y_pred == [0]):
print("THIS SONG IS NOT HIT")
else:
print("THIS SONG IS HIT")