def test_get_logger_no_duplicate_handlers(self, mock_read_config):
config = self.config.copy()
# No duplicate hanlders
mock_read_config.return_value = config
logger = get_logger()
number_of_handlers = len(logger.handlers)
logger = get_logger()
self.assertEqual(len(logger.handlers), number_of_handlers)
def test_get_logger_no_config(self, mock_read_config):
config = self.config.copy()
# Add null handler if not configured
del config["app"]["logger"]
mock_read_config.return_value = config
logger = get_logger()
self.assertTrue(
len([h for h in logger.handlers if isinstance(h, logging.NullHandler)]) > 0
)
import torch
import torch.utils.data
import tqdm
from src import config, get_logger, round_probabilities, Dataset, get_metrics
logger = get_logger()
logger.info("Loading train dataset")
trainDataset = Dataset(config["path_to_processed_MAPS"], "train")
trainDatasetMean = trainDataset.mean()
logger.info("Loading test dataset")
testDataset = Dataset(config["path_to_processed_MAPS"], "test")
dataloader = torch.utils.data.DataLoader(testDataset, batch_size=config["mini_batch_size"], shuffle=True)
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
model = torch.load("./models/Dnn3Layers/095.pth")
model = model.to(device)
model = model.eval()
testMetrics = {}
for batchX, batchy in tqdm.tqdm(dataloader):
batchX -= trainDatasetMean
batchX = batchX.to(device)
batchy = batchy.to(device)
prediction = round_probabilities(model(batchX))
def test_get_logger(self, mock_read_config):
config = self.config.copy()
# success flow
mock_read_config.return_value = config
logger = get_logger()
self.assertTrue(len(logger.handlers) > 1)
import os
import re
from src import get_logger
from src.annict import Annict
from src.utils import get_today
from src.aws.sns import SNS
SNS_TOPIC = os.getenv("SNS_TOPIC_ARN")
logger = get_logger(__name__)
def handler(event, context):
annict = Annict()
programs: dict = annict.get_episodes()
today_animes: dict = annict.get_stream_episode_specified_date(
watch_date=get_specified_date(event), programs=programs)
sns = SNS()
sns.send_message(topic_arn=SNS_TOPIC,
message=create_notify_message(today_animes))
def get_specified_date(event):
is_matched = False
if event.get('s_date'):
# Only 0000-00-00 that is date.
is_matched = re.match(r"\d{4}-\d{2}-\d{2}", event.get('s_date'))
default='./output.midi')
args = parser.parse_args()
# Start generating midi workflow
class SongDataset(torch.utils.data.Dataset):
def __init__(self, X):
self._X = X
def __len__(self):
return self._X.size()[0]
def __getitem__(self, index):
return self._X[index]
logger = get_logger(file_silent=True)
logger.info("Performing cqt")
X = cqt(args.wav)
logger.info("Done")
logger.info("Loading model")
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
model = torch.load(os.path.join(config["public_models_folder"], args.model))
model = model.to(device)
model = model.eval()
X = torch.from_numpy(X).float()
trainDatasetMeans = torch.load(os.path.join(config["public_models_folder"], "train_means.pth"))