[BUG] Fix issue with EncodeNormalizer(method='standard', center=False) for scale value

[fnhirwa]

Fixes #1901
This PR fixes the issue with

encoder = EncoderNormalizer(
    method='standard',
    center=False,
    max_length=None,
    transformation=None,
    method_kwargs={}
)

which was returning scale_ as mean of x instead its std

Temporarily backported lazywhere implementation from scipy._lib._util. to avoid import errors in the statsmodels with the latest scipy version.

[fnhirwa]

The variance distorts when the data is not centered, as the log1p transformation is log(1+x)

[codecov]

Codecov Report

Attention: Patch coverage is 27.27273% with 8 lines in your changes missing coverage. Please review.

Please upload report for BASE (main@b2cfc14). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
...sting/models/temporal_fusion_transformer/tuning.py	27.27%	8 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1902   +/-   ##
=======================================
  Coverage        ?   86.25%           
=======================================
  Files           ?       96           
  Lines           ?     7772           
  Branches        ?        0           
=======================================
  Hits            ?     6704           
  Misses          ?     1068           
  Partials        ?        0           

Flag	Coverage Δ
cpu	86.25% <27.27%> (?)
pytest	86.25% <27.27%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[cngmid]

Linked to this issue, the method

    @property
    def min_length(self):
        if self.method == "identity":
            return 0  # no timeseries properties used
        elif self.center:
            return 1  # only calculation of mean required
        else:
            return 2  # requires std, i.e. at least 2 entries

of class EncoderNormalizer might also need to be reviewed.
If the method is not "identity", one always needs at least 2 entries, independently whether self.center is True or False.

[fkiraly]

looks good to me in terms of the location directly fixed, though see above comment by @cngmid

[fkiraly]

FYI, if you have time to look at the __array_wrap__ issue, that would be appreciated - I tried here, but apparently I am not catching all cases that the function input may have: #1855

[fnhirwa]

FYI, if you have time to look at the __array_wrap__ issue, that would be appreciated - I tried here, but apparently I am not catching all cases that the function input may have: #1855

sure will look into this🤞