Update the Sign operation in ops.py

[YifanDengWHU]

Fix the bug when Sign followed by True/False value.

Description

During the alpha generation exps, I found that when the Sign operation meets True/False value, it will raise a MaybeEncodingError. Here I transform the True/False value to np.float32.

Motivation and Context

How Has This Been Tested?

• Pass the test by running: pytest qlib/tests/test_all_pipeline.py under upper directory of qlib.
• [x ] If you are adding a new feature, test on your own test scripts.

Screenshots of Test Results (if appropriate):

1. Pipeline test:
2. Your own tests:

Types of changes

• [x ] Fix bugs
• Add new feature
• Update documentation

[you-n-g]

Hi,
Many thanks for sending the pull request.
But I can't understand why the data type conversion is required. Can you give us more details? Thanks

[YifanDengWHU]

I mean when the sign function meets bool value, it will raise error. Here is the expression I test yesterday:
Sign(Gt($close, Sum(WMA($volume, 14),17)))
The $close value will be smaller than Sum(WMA($volume, 14),17)). So it is Sign(series([False, False, .......])) and it will raise error.
Though this expression is meaningless, but I think we should make it work?

[you-n-g]

@YifanDengWHU Thanks. I see now.

Please add some comments to describe this case.
In addition, please help us to add this comments above the np.float32

# TODO:  The the precision should be configurable

Thanks for your contribution!