BUG: Using kwargs in Rolling.apply()

[echasnovski]

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas.

• (optional) I have confirmed this bug exists on the master branch of pandas.

---

Code Sample, a copy-pastable example

from functools import partial

import numpy as np
import pandas as pd


def foo(x, par):
    return np.sum(x + par)


df = pd.DataFrame({"gr": [1, 1], "a": [1, 2]})

# With `kwargs`
try:
    print(df.groupby("gr")["a"].rolling(1).apply(foo, kwargs={"par": 10}))
except Exception as e:
    print(e)

# Using `partial`
bar = partial(foo, par=10)
print(df.groupby("gr")["a"].rolling(1).apply(bar))

Problem description

Expected: using appropriate kwargs argument for pandas.core.window.rolling.Rolling.apply() should be equivalent to premodifying supplied function with functools.partial() (because it is stated in documentation).
Current: kwargs doesn't seem to be used at all, because it results into "missing 1 required positional argument" exception.
Note: This works as expected in pandas 0.25.3, but gives error as early as 1.0.0.

Expected Output

gr   
1   0    11.0
    1    12.0
Name: a, dtype: float64
gr   
1   0    11.0
    1    12.0
Name: a, dtype: float64

Output of pd.show_versions()

INSTALLED VERSIONS

commit : None
python : 3.8.2.final.0
python-bits : 64
OS : Linux
OS-release : 4.15.0-29-generic
machine : x86_64
processor : x86_64
byteorder : little
LC_ALL : None
LANG : ru_UA.UTF-8
LOCALE : ru_UA.UTF-8

pandas : 1.0.3
numpy : 1.18.1
pytz : 2019.3
dateutil : 2.8.1
pip : 20.0.2
setuptools : 46.1.3.post20200325
Cython : None
pytest : None
hypothesis : None
sphinx : None
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : None
pymysql : None
psycopg2 : None
jinja2 : None
IPython : None
pandas_datareader: None
bs4 : None
bottleneck : None
fastparquet : None
gcsfs : None
lxml.etree : None
matplotlib : None
numexpr : None
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : None
pytables : None
pytest : None
pyxlsb : None
s3fs : None
scipy : None
sqlalchemy : None
tables : None
tabulate : None
xarray : None
xlrd : None
xlwt : None
xlsxwriter : None
numba : None

[simonjayhawkins]

Note: This works as expected in pandas 0.25.3, but gives error as early as 1.0.0.

regression in #29428 (i.e. 1.0.0)

6e5d148 is the first bad commit
commit 6e5d148
Author: Matthew Roeschke [email protected]
Date: Thu Nov 21 04:59:30 2019 -0800

REF: Separate window bounds calculation from aggregation functions (#29428)