update fix CI tests bugs

qlib/workflow/record_temp.py

@@ -16,7 +16,7 @@ from ..data.dataset.handler import DataHandlerLP
 from ..backtest import backtest as normal_backtest
 from ..utils import init_instance_by_config, get_module_by_module_path
 from ..log import get_module_logger
-from ..utils import flatten_dict
+from ..utils import flatten_dict, class_casting
 from ..utils.time import Freq
 from ..strategy.base import BaseStrategy
 from ..contrib.eva.alpha import calc_ic, calc_long_short_return, calc_long_short_prec
@@ -32,6 +32,7 @@ class RecordTemp:
     """
 
     artifact_path = None
+    depend_cls = None  # the depend class of the record; the record will depend on the results generated by `depend_cls`
 
     @classmethod
     def get_path(cls, path=None):
@@ -98,21 +99,30 @@ class RecordTemp:
         """
         return []
 
-    def check(self, cls="self"):
+    def check(self, include_self: bool = False):
         """
         Check if the records is properly generated and saved.
+        It is useful in fololwing examples
+        - checking if the depended files complete before genrating new things.
+        - checking if the final files is completed
+
+        Parameters
+        ----------
+        include_self : bool
+            is the file generated by self included
 
         Raise
         ------
         FileExistsError: whether the records are stored properly.
         """
         artifacts = set(self.recorder.list_artifacts())
-        if cls == "self":
-            cls = self
-        flist = cls.list()
-        for item in flist:
-            if item not in artifacts:
-                raise FileExistsError(item)
+        if include_self:
+            for item in self.list():
+                if item not in artifacts:
+                    raise FileExistsError(item)
+        if self.depend_cls is not None:
+            with class_casting(self, self.depend_cls):
+                self.check(include_self=True)
 
 
 class SignalRecord(RecordTemp):
@@ -127,26 +137,20 @@ class SignalRecord(RecordTemp):
 
     @staticmethod
     def generate_label(dataset):
-        # NOTE:
-        # Python doesn't provide the downcasting mechanism.
-        # We use the trick here to downcast the class
-        orig_cls = dataset.__class__
-        dataset.__class__ = DatasetH
-
-        params = dict(segments="test", col_set="label", data_key=DataHandlerLP.DK_R)
-        try:
-            # Assume the backend handler is DataHandlerLP
-            raw_label = dataset.prepare(**params)
-        except TypeError:
-            # The argument number is not right
-            del params["data_key"]
-            # The backend handler should be DataHandler
-            raw_label = dataset.prepare(**params)
-        except AttributeError:
-            # The data handler is initialize with `drop_raw=True`...
-            # So raw_label is not available
-            raw_label = None
-        dataset.__class__ = orig_cls
+        with class_casting(dataset, DatasetH):
+            params = dict(segments="test", col_set="label", data_key=DataHandlerLP.DK_R)
+            try:
+                # Assume the backend handler is DataHandlerLP
+                raw_label = dataset.prepare(**params)
+            except TypeError:
+                # The argument number is not right
+                del params["data_key"]
+                # The backend handler should be DataHandler
+                raw_label = dataset.prepare(**params)
+            except AttributeError:
+                # The data handler is initialize with `drop_raw=True`...
+                # So raw_label is not available
+                raw_label = None
         return raw_label
 
     def generate(self, **kwargs):
@@ -235,7 +239,7 @@ class SigAnaRecord(RecordTemp):
     """
 
     artifact_path = "sig_analysis"
-    pre_class = SignalRecord
+    depend_cls = SignalRecord
 
     def __init__(self, recorder, ana_long_short=False, ann_scaler=252, label_col=0):
         super().__init__(recorder=recorder)
@@ -244,7 +248,7 @@ class SigAnaRecord(RecordTemp):
         self.label_col = label_col
 
     def generate(self, **kwargs):
-        self.check(self.pre_class)
+        self.check()
 
         pred = self.load("pred.pkl")
         label = self.load("label.pkl")