liuyu/qlib
1
0
Fork 0
mirror of https://github.com/microsoft/qlib.git synced 2026-07-10 22:36:55 +08:00
Code Activity

bug fixed

Browse Source
This commit is contained in:
lzh222333
2021-05-14 11:31:50 +00:00
parent aef3f186c1
commit a986379deb
3 changed files with 71 additions and 23 deletions
Download Patch File Download Diff File Expand all files Collapse all files

25
qlib/workflow/online/manager.py
View File

@@ -262,12 +262,29 @@ class OnlineManager(Serializable):
Prepare all models and signals if something is waiting for preparation.
Args:
model_kwargs: the params for `prepare_online_models`
model_kwargs: the params for `end_train`
signal_kwargs: the params for `prepare_signals`
"""
last_models = {}
signals_time = D.calendar()[0]
need_prepare = False
for cur_time, strategy_models in self.history.items():
self.cur_time = cur_time
for strategy, models in strategy_models.items():
models = self.trainer.end_train(models, experiment_name=strategy.name_id)
# NOTE: Assumption: the predictions of online models need less than next cur_time, or this method will work in a wrong way.
self.prepare_signals(**signal_kwargs)
# only new online models need to prepare
if last_models.setdefault(strategy, set()) != set(models):
models = self.trainer.end_train(models, experiment_name=strategy.name_id, **model_kwargs)
strategy.tool.reset_online_tag(models)
need_prepare = True
last_models[strategy] = set(models)
if need_prepare:
# NOTE: Assumption: the predictions of online models need less than next cur_time, or this method will work in a wrong way.
self.prepare_signals(**signal_kwargs)
if signals_time > cur_time:
self.logger.warn(
f"The signals have already parpred to {signals_time} by last preparation, but current time is only {cur_time}. This may be because the online models predict more than they should, which can cause signals to be contaminated by the offline models."
)
need_prepare = False
signals_time = self.signals.index.get_level_values("datetime").max()

11
qlib/workflow/recorder.py
View File

@@ -39,6 +39,9 @@ class Recorder:
def __str__(self):
return str(self.info)
def __hash__(self) -> int:
return hash(self.info["id"])
@property
def info(self):
output = dict()
@@ -232,6 +235,14 @@ class MLflowRecorder(Recorder):
client=self.client,
)
def __hash__(self) -> int:
return hash(self.info["id"])
def __eq__(self, o: object) -> bool:
if isinstance(o, MLflowRecorder):
return self.info["id"] == o.info["id"]
return False
@property
def uri(self):
return self._uri