def __init__(self,
config,
stockList,
featureList,
start_date,
end_date,
fake_data,
restore_dir=None,
save_path=None,
agent=None,
device="cpu"):
config["training"]["buffer_biased"] = config["trading"][
"buffer_biased"]
config["training"]["learning_rate"] = config["trading"][
"learning_rate"]
TraderTrainer.__init__(self,
config,
stockList,
featureList,
start_date,
end_date,
fake_data,
restore_dir=restore_dir,
save_path=save_path,
agent=agent,
device=device)
def train_main(device='cpu'):
# logging level 설정
console_level = logging.INFO
logfile_level = logging.DEBUG
# net_config.json 파일이 위치한 디렉토리
config_dir = "train_package"
# 디렉토리명을 현재 시간으로 설정 및 생성
now_dt = datetime.datetime.today()
now_str = now_dt.strftime('%y%m%d_%H%M')
now_str = 'run' + now_str
base_log_dir = config_dir + '/' + now_str
os.makedirs(base_log_dir)
# 프로그램 로그파일(programlog)와 콘솔 로그 레벨 설정
# 파일에 출력할 로그와 콘솔에 출력할 로그를 구분한다는 의미임
logging.basicConfig(filename=base_log_dir + '/programlog', level=logfile_level)
console = logging.StreamHandler()
console.setLevel(console_level)
logging.getLogger().addHandler(console)
print("training at {} started".format(now_str))
tt = TraderTrainer(load_config(), save_path=base_log_dir + '/netfile', device=device) # load_config() : load json file depending on Python version.
train_retval = tt.train_net(log_file_dir=base_log_dir + '/tensorboard', index=now_str)
print("training at {} finished".format(now_str))
return train_retval
def train_one(save_path, config, log_file_dir, index, logfile_level,
console_level, device):
"""
train an agent
:param save_path: the path to save the tensorflow model (.ckpt), could be None
:param config: the json configuration file
:param log_file_dir: the directory to save the tensorboard logging file, could be None
:param index: identifier of this train, which is also the sub directory in the train_package,
if it is 0. nothing would be saved into the summary file.
:param logfile_level: logging level of the file
:param console_level: logging level of the console
:param device: 0 or 1 to show which gpu to use, if 0, means use cpu instead of gpu
:return : the Result namedtuple
"""
if log_file_dir:
logging.basicConfig(filename=log_file_dir.replace(
"tensorboard", "programlog"),
level=logfile_level)
console = logging.StreamHandler()
console.setLevel(console_level)
logging.getLogger().addHandler(console)
else:
print('log_file_dir is not existent')
print("training at %s started" % index)
pdb.set_trace()
trainer = TraderTrainer(config, save_path=save_path,
device=device) #初始化训练器
return trainer.train_net(log_file_dir=log_file_dir, index=index) #训练网络
def __init__(self,
config,
restore_dir=None,
save_path=None,
agent=None,
device="cpu"):
config["training"]["buffer_biased"] = config["trading"][
"buffer_biased"]
config["training"]["learning_rate"] = config["trading"][
"learning_rate"]
TraderTrainer.__init__(self,
config,
restore_dir=restore_dir,
save_path=save_path,
agent=agent,
device=device)
def train_one(save_path, config, log_file_dir, index, logfile_level,
console_level, device):
"""
aim: train an agent
:param save_path: the path to save the tensorflow model (.ckpt), could be None 存储模型的checkpoints文件
:param config: the json configuration file 存储config的json文件
:param log_file_dir: the directory to save the tensorboard logging file, could be None 存储tensorboard的文件夹
:param index: identifier of this train, which is also the sub directory in the train_package, 训练的子文件夹是第几个？
if it is 0. nothing would be saved into the summary file.
:param logfile_level: logging level of the file
:param console_level: logging level of the console
:param device: 0 or 1 to show which gpu to use, if 0, means use cpu instead of gpu
:return : the Result namedtuple
"""
if log_file_dir:
logging.basicConfig(
filename=log_file_dir.replace("tensorboard", "programlog"),
level=logfile_level) #initialize the logging console
console = logging.StreamHandler(
) # output the logging information to sys.stdout, sys.stderr
console.setLevel(console_level)
logging.getLogger().addHandler(console)
print("training at %s started" % index)
# train the network and execute backtest
return TraderTrainer(config, save_path=save_path,
device=device).train_net(log_file_dir=log_file_dir,
index=index)
def train_one(save_path, config, log_file_dir, index, logfile_level,
console_level, device):
"""
エージェントをトレーニングします。
:param save_path: TensorFlowのモデル(.ckpt)を保存するパス。None可。
:param config: jsonのコンフィグファイル
:param log_file_dir: TensorBoardのログファイルを保存するディレクトリ。None可。
:param index: トレーニングを一意に定めるインデックス。train_packageのサブディレクトリ名。
:param logfile_level: ファイルのログレベル
:param console_level: コンソールのログレベル
:param device: 0ならCPU、1ならGPU。
:return: 結果のnamedtuple
"""
if log_file_dir:
logging.basicConfig(filename=log_file_dir.replace(
'tensorboard', 'programlog'),
lovel=logfile_level)
console = logging.StreamHandler()
console.setLevel(console_level)
logging.getLogger().addHandler(console)
print('training at %s started' % index)
trainer = TraderTrainer(config, save_path=save_path, device=device)
return trainer.train_net(log_file_dir=log_file_dir, index=index)
@author: Joukey
"""
import os
os.chdir('D:/EIIE_keras/')
from pgportfolio.learn.tradertrainer import TraderTrainer
from pgportfolio.tools.configprocess import load_config
config_2 = load_config(2)
train_dir = "train_package"
s_path = "./" + train_dir + "/" + '2' + "/netfile"
l_path = "./" + train_dir + "/" + '2' + "/tensorboard"
t_trainer = TraderTrainer(config_2, save_path=s_path, device='gpu')
#%%
import logging
console_level = logging.INFO
logfile_level = logging.DEBUG
logging.basicConfig(filename=l_path.replace("tensorboard", "programlog"),
level=console_level)
console = logging.StreamHandler()
console.setLevel(console_level)
logging.getLogger().addHandler(console)
t_trainer.train_net(log_file_dir=l_path, index='2')
def __init__(self, config, restore_dir=None, save_path=None, agent=None, device="cpu"):
config["training"]["buffer_biased"] = config["trading"]["buffer_biased"]
config["training"]["learning_rate"] = config["trading"]["learning_rate"]
TraderTrainer.__init__(self, config, restore_dir=restore_dir, save_path=save_path,
agent=agent, device=device)
def main(logPath, device):
parser = build_parser()
options = parser.parse_args()
'''
options.folder = oneKey
options.stockList = stockList
options.featureList = featureList
options.start_date = startDate
options.end_date = endDate
'''
if not os.path.exists("./" + "database"):
os.makedirs("./" + "database")
#options.repeat = 1
if options.mode == "train": #训练数据
if not options.algo:
save_path = logPath + str(options.folder) + "/netfile"
# 读取配置文件
with open(logPath + str(options.folder) +
"\\net_config.json") as file:
config_json = None
config_json = json.load(file)
config = preprocess_config(config_json)
log_file_dir = logPath + str(options.folder) + "/tensorboard"
# 定义错误等级
logfile_level = logging.DEBUG
console_level = logging.INFO
logging.basicConfig(filename=log_file_dir.replace(
"tensorboard", "programlog"),
level=logfile_level)
console = logging.StreamHandler()
console.setLevel(console_level)
logging.getLogger().addHandler(console)
trainer = TraderTrainer(config,
options.stockList,
options.featureList,
options.start_date,
options.end_date,
save_path=save_path,
device=device) #初始化训练器
trainer.train_net(log_file_dir=log_file_dir,
index=str(options.folder)) #训练网络
else:
for folder in options.folder:
raise NotImplementedError()
# 生成配置文件到路径中，要想修改配置，直接修改PGPortfolio\pgportfolio\net_config.json
elif options.mode == "generate":
import pgportfolio.autotrain.generate as generate
logging.basicConfig(level=logging.INFO)
config_ = load_config()
train_dir = logPath
generate.add_packages(train_dir, load_config(), int(options.repeat))
elif options.mode == "download_data":
from pgportfolio.marketdata.datamatrices import DataMatrices
with open("./pgportfolio/net_config.json") as file:
config = json.load(file)
config = preprocess_config(config)
start = time.mktime(
datetime.strptime(options.start_date, "%Y/%m/%d").timetuple())
end = time.mktime(
datetime.strptime(options.end_date, "%Y/%m/%d").timetuple())
DataMatrices(
start=start,
end=end,
feature_number=len(options.featureList),
window_size=config["input"]["window_size"],
online=True,
period=config["input"]["global_period"],
volume_average_days=config["input"]["volume_average_days"],
coin_filter=len(options["stockList"]),
is_permed=config["input"]["is_permed"],
test_portion=config["input"]["test_portion"],
portion_reversed=config["input"]["portion_reversed"])
elif options.mode == "backtest":
config = _config_by_algo(options.algo) #读取配置文件
_set_logging_by_algo(logging.DEBUG, logging.DEBUG, options.algo,
"backtestlog") #设置log的路径
values = execute_backtest(options.algo, config) #执行回测的步数为训练集的长度
elif options.mode == "save_test_data":
# This is used to export the test data
save_test_data(load_config(options.folder)) #保存测试集数据
elif options.mode == "plot":
logging.basicConfig(level=logging.INFO)
algos = options.algos.split(",")
if options.labels:
labels = options.labels.replace("_", " ")
labels = labels.split(",")
else:
labels = algos
plot.plot_backtest(load_config(), algos, labels)
elif options.mode == "table":
algos = options.algos.split(",")
if options.labels:
labels = options.labels.replace("_", " ")
labels = labels.split(",")
else:
labels = algos
plot.table_backtest(load_config(),
algos,
labels,
format=options.format)