liuyu/qlib
1
0
Fork 0
mirror of https://github.com/microsoft/qlib.git synced 2026-07-12 07:16:54 +08:00
Code Activity

Add a few comments

Browse Source
This commit is contained in:
Yuge Zhang
2021-06-02 16:49:52 +08:00
parent d515efb46e
commit cc8339acd9
2 changed files with 13 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

BIN
rl_orders Normal file
View File

Binary file not shown.

16
rl_playground.py
View File

@@ -16,7 +16,7 @@ from qlib.tests.data import GetData
from qlib.utils import init_instance_by_config, exists_qlib_data from qlib.utils import init_instance_by_config, exists_qlib_data
from torch.utils.data import Dataset, DataLoader from torch.utils.data import Dataset, DataLoader
from tianshou.data import Batch, Collector from tianshou.data import Batch, Collector
from tianshou.env import DummyVectorEnv from tianshou.env import DummyVectorEnv, SubprocVectorEnv
from tianshou.policy import BasePolicy from tianshou.policy import BasePolicy
@@ -51,7 +51,8 @@ def price_advantage(exec_price: float, baseline_price: float, direction: int) ->
@dataclass @dataclass
class EpisodicState: class EpisodicState:
""" """
A simplified data structure for RL-related components to process observations and rewards A simplified data structure as the input of RL-related components to calculate observations and rewards.
Some of the metrics info are calculated on-the-fly in this class.
""" """
# requirements # requirements
stock_id: int stock_id: int
@@ -181,6 +182,7 @@ class SingleOrderEnv(gym.Env):
return self.observation.observation_space return self.observation.observation_space
def retrieve_backtest_data(self, field: str): def retrieve_backtest_data(self, field: str):
# Retrieve backtest data for RL-specific use (including reward calculation)
return D.features( return D.features(
[self.cur_order.stock_id], [self.cur_order.stock_id],
['$open', '$close', '$high', '$low', '$volume'], ['$open', '$close', '$high', '$low', '$volume'],
@@ -190,6 +192,7 @@ class SingleOrderEnv(gym.Env):
)[field].to_numpy() )[field].to_numpy()
def initialize_state(self): def initialize_state(self):
# Synchronous state for executor to EpisodicState
self.executor.reset(start_time=self.cur_order.start_time, end_time=self.cur_order.end_time) self.executor.reset(start_time=self.cur_order.start_time, end_time=self.cur_order.end_time)
state = EpisodicState( state = EpisodicState(
stock_id=self.cur_order.stock_id, stock_id=self.cur_order.stock_id,
@@ -207,6 +210,7 @@ class SingleOrderEnv(gym.Env):
return state return state
def update_state(self, exec_vol): def update_state(self, exec_vol):
# Synchronous exec_vol to executor and synchronous back to EpisodicState
calendar = self.executor.trade_calendar calendar = self.executor.trade_calendar
state = self.ep_state state = self.ep_state
@@ -273,6 +277,7 @@ class SingleOrderEnv(gym.Env):
'ins': self.ep_state.stock_id, 'ins': self.ep_state.stock_id,
'date': self.ep_state.start_time, 'date': self.ep_state.start_time,
} }
# TODO: collect logs
pprint(info) pprint(info)
return self.observation(self.ep_state), reward, self.ep_state.done, info return self.observation(self.ep_state), reward, self.ep_state.done, info
@@ -327,13 +332,18 @@ def _main():
) )
return SingleOrderEnv( return SingleOrderEnv(
observation, action, reward_fn, observation, action, reward_fn,
iter(DataLoader(QlibOrderDataset('rl.pkl'), batch_size=None, shuffle=True)), executor) iter(DataLoader(QlibOrderDataset('rl_orders'), batch_size=None, shuffle=True)), executor)
policy = DummyPolicy() policy = DummyPolicy()
# This can not be replaced with SubprocVectorEnv
# File "/xxx/qlib/qlib/data/data.py", line 462, in dataset_processor
# p = Pool(processes=workers)
# AssertionError: daemonic processes are not allowed to have children
envs = DummyVectorEnv([dummy_env for _ in range(4)]) envs = DummyVectorEnv([dummy_env for _ in range(4)])
test_collector = Collector(policy, envs) test_collector = Collector(policy, envs)
policy.eval() policy.eval()
# TODO: create a queue for all orders and make it auto-complete when all the orders are processed
test_collector.collect(n_episode=10) test_collector.collect(n_episode=10)