From b15e5e33fd0d20e24ac52c9397374af27a5bc987 Mon Sep 17 00:00:00 2001
From: Young <afe.young@gmail.com>
Date: Mon, 12 Apr 2021 06:33:31 +0000
Subject: [PATCH] Fix the multi-processing bug

---
 qlib/model/ens/group.py       | 8 +++-----
 qlib/utils/serial.py          | 6 ++++--
 qlib/workflow/task/collect.py | 2 +-
 3 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/qlib/model/ens/group.py b/qlib/model/ens/group.py
index f5ab5d8a7..c80959b0d 100644
--- a/qlib/model/ens/group.py
+++ b/qlib/model/ens/group.py
@@ -1,10 +1,9 @@
 from qlib.model.ens.ensemble import Ensemble, RollingEnsemble
 from typing import Callable, Union
-from qlib.utils.serial import Serializable
 from joblib import Parallel, delayed
 
 
-class Group(Serializable):
+class Group:
     """Group the objects based on dict"""
 
     def __init__(self, group_func=None, ens: Ensemble = None):
@@ -45,9 +44,8 @@ class Group(Serializable):
             dict: grouped_dict like {G1: object, G2: object}
         """
 
-        # FIXME: The multiprocessing will raise the following error
-        # NotImplementedError: Please specify valid `_ens_func`.
-        # The problem maybe the state of the function is lost
+        # NOTE: The multiprocessing will raise error if you use `Serializable`
+        # Because the `Serializable` will affect the behaviours of pickle
         grouped_dict = self.group(ungrouped_dict, *args, **kwargs)
 
         key_l = []
diff --git a/qlib/utils/serial.py b/qlib/utils/serial.py
index b94be464b..1b775d99a 100644
--- a/qlib/utils/serial.py
+++ b/qlib/utils/serial.py
@@ -7,8 +7,10 @@ import pickle
 
 class Serializable:
     """
-    Serializable behaves like pickle.
-    But it only saves the state whose name **does not** start with `_`
+    Serializable will change the behaviours of pickle.
+    - It only saves the state whose name **does not** start with `_`
+    It provides a syntactic sugar for distinguish the attributes which user doesn't want.
+    - For examples, a learnable Datahandler just wants to save the parameters without data when dumping to disk
     """
 
     def __init__(self):
diff --git a/qlib/workflow/task/collect.py b/qlib/workflow/task/collect.py
index 9bd609670..f651ef8d8 100644
--- a/qlib/workflow/task/collect.py
+++ b/qlib/workflow/task/collect.py
@@ -4,7 +4,7 @@ from qlib.workflow.task.utils import list_recorders
 from qlib.utils.serial import Serializable
 
 
-class Collector(Serializable):
+class Collector:
     """The collector to collect different results"""
 
     def collect(self, *args, **kwargs):