def train(start, end, label, label_values, model_dir, receipt):
train_path = create_train_path()
start_date = datetime.strptime(start, '%d-%m-%Y')
end_date = datetime.strptime(end, '%d-%m-%Y')
next_date = end_date + pd.DateOffset(years=1)
if end_date > datetime.now():
end_date = datetime.now()
learning_cfg = get_learning_cfg(model_dir)
train_filename = "train-players-" + start_date.strftime(
"%d-%m-%Y") + '-' + end_date.strftime("%d-%m-%Y") + ".csv"
evaluate_filename = "train-players-" + end_date.strftime(
"%d-%m-%Y") + '-' + next_date.strftime("%d-%m-%Y") + ".csv"
train_file_path = local_dir + train_path + train_filename
evaluate_file_path = local_dir + train_path + evaluate_filename
has_data = model_utils.create_csv(url=model_utils.EVENT_MODEL_URL,
filename=train_file_path,
start_date=start_date,
end_date=end_date,
aws_path=train_path)
if learning_cfg['evaluate']:
has_test_data = model_utils.create_csv(url=model_utils.EVENT_MODEL_URL,
filename=evaluate_file_path,
start_date=end_date,
end_date=next_date,
aws_path=train_path)
if has_data == True and has_test_data == False:
evaluate_filename = None
else:
logger.info('we can evaluate')
if has_data:
train_filename = train_path + train_filename
if evaluate_filename is not None:
evaluate_filename = train_path + evaluate_filename
match_model.create(train=True,
label=label,
label_values=label_values,
model_dir=model_dir,
train_filename=train_filename,
test_filename=evaluate_filename,
init=True)
else:
logger.info('no data to train')
if receipt is not None:
receipt_service.put_receipt(receipt_service.TRAIN_RECEIPT_URL, receipt,
None)
def predict(data, country, label, label_values, model_dir, receipt):
# def create(type, country, train, label, label_values, model_dir, train_filename, test_filename, outcome, previous_vocab_date):
# there is no guarantee the predict is on same day as the train. so we need the history
classifier = match_model.create(country=country,
train=False,
label=label,
label_values=label_values,
model_dir=model_dir,
train_filename='',
test_filename='')
home = []
away = []
outcomes = []
# Generate predictions from the model
home.append(data['home'])
away.append(data['away'])
# print(data)
expected = [0]
predict_x = {'home': home, 'away': away}
response = model_utils.predict(classifier=classifier,
predict_x=predict_x,
label_values=label_values)
match_model.tidy_up(local_dir + '/models/' + model_dir + '/' + country,
None, None, None)
receipt_service.put_receipt(receipt_service.PREDICT_RECEIPT_URL, receipt,
response)
def predict(data, init, label, label_values, model_dir, receipt):
# def create(type, country, train, label, label_values, model_dir, train_filename, test_filename, outcome, previous_vocab_date):
# there is no guarantee the predict is on same day as the train. so we need the history
classifier = match_model.create(train=False,
label=label,
label_values=label_values,
model_dir=model_dir,
train_filename='',
test_filename='',
init=init)
player = []
home = []
opponent = []
# Generate predictions from the model
opponent.append(data['opponent'])
home.append(data['home'])
player.append(data['player'])
predict_x = {'player': player, 'opponent': opponent, 'home': home}
response = model_utils.predict(classifier=classifier,
predict_x=predict_x,
label_values=label_values)
if init:
logger.info('tidying up')
match_model.tidy_up(local_dir + '/models/' + model_dir, None, None,
None)
receipt_service.put_receipt(receipt_service.PREDICT_RECEIPT_URL, receipt,
response)
def train(data_range, label, label_values, model_dir, train_path, receipt,
history, history_file):
for data in data_range:
learning_cfg = get_learning_cfg(model_dir)
train_filename = "train-players" + data.replace('/', '-') + ".csv"
evaluate_filename = "train-players" + get_next_in_range(
data_range, data).replace('/', '-') + ".csv"
train_file_path = local_dir + train_path + train_filename
evaluate_file_path = local_dir + train_path + evaluate_filename
has_data = model_utils.create_csv(url=model_utils.EVENT_MODEL_URL,
filename=train_file_path,
range=data,
aws_path=train_path)
if learning_cfg['evaluate']:
has_test_data = model_utils.create_csv(
url=model_utils.EVENT_MODEL_URL,
filename=evaluate_file_path,
range=get_next_in_range(data_range, data),
aws_path=train_path)
if has_data == True and has_test_data == False:
evaluate_filename = None
else:
logger.info('we can evaluate')
if has_data:
train_filename = train_path + train_filename
if evaluate_filename is not None:
evaluate_filename = train_path + evaluate_filename
##take a copy of our file if it doesnt exist.
#if not is_on_file(test_file_path):
# copyfile(train_file_path,
# test_file_path)
# put_aws_file_with_path(train_path,test_filename)
# write_filenames_index_from_filename(test_file_path)
# else:
# get_aws_file(train_path, test_filename)
match_model.create(train=True,
label=label,
label_values=label_values,
model_dir=model_dir,
train_filename=train_filename,
test_filename=evaluate_filename,
init=True)
else:
logger.info('no data to train')
#write the history...
start_day, start_month, start_year, end_day, end_month, end_year = get_range_details(
data)
history = train_history_utils.create_history('Success - Partial',
start_day, start_month,
start_year, end_day,
end_month, end_year)
train_history_utils.add_history(history_file, 'default', history)
if receipt is not None:
receipt_utils.put_receipt(receipt_utils.TRAIN_RECEIPT_URL, receipt,
None)
history['status'] = "Success - Full"
train_history_utils.add_history(history_file, 'default', history)