BUG: rolling with uneven nanosecond window have wrong results

[rosagold]

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

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

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

Reproducible Example

import pandas as pd

index = pd.date_range("2020", periods=4, freq="1ns")
df = pd.Series([1, 1, 1, 1], index=index, dtype=float)
result = df.rolling('3ns', closed='neither', center=True).sum()

print(result)
# 2020-01-01 00:00:00.000000000    1.0
# 2020-01-01 00:00:00.000000001    1.0
# 2020-01-01 00:00:00.000000002    1.0
# 2020-01-01 00:00:00.000000003    1.0
# Freq: N, dtype: float64

Issue Description

Rolling with smallest possible resolution (currently nanoseconds) centered windows and uneven nanoseconds produce wrong results.

Eg. rolling with a centered window='3ns', regardless of closed, should result in a forward facing and a backward facing window of 1.5ns, which is effectively a fwd/bwd facing window of 1ns with closed='both'.

btw. thats exactly the code change in the followup PR will introduce ;)

Expected Behavior

>>> result
2020-01-01 00:00:00.000000000    2.0
2020-01-01 00:00:00.000000001    3.0
2020-01-01 00:00:00.000000002    3.0
2020-01-01 00:00:00.000000003    2.0
Freq: N, dtype: float64

Installed Versions

INSTALLED VERSIONS

commit : 075ec87
python : 3.8.12.final.0
python-bits : 64
OS : Linux
OS-release : 5.4.0-88-generic
Version : #99-Ubuntu SMP Thu Sep 23 17:29:00 UTC 2021
machine : x86_64
processor : x86_64
byteorder : little
LC_ALL : None
LANG : en_US.UTF-8
LOCALE : en_US.UTF-8
pandas : 1.4.0.dev0+869.g075ec87643
numpy : 1.21.2
pytz : 2021.1
dateutil : 2.8.2
pip : 21.2.4
setuptools : 58.2.0
Cython : 0.29.24
pytest : 6.2.5
hypothesis : 6.23.1
sphinx : 4.2.0
blosc : None
feather : None
xlsxwriter : 3.0.1
lxml.etree : 4.6.3
html5lib : 1.1
pymysql : None
psycopg2 : None
jinja2 : 3.0.1
IPython : 7.28.0
pandas_datareader: None
bs4 : 4.10.0
bottleneck : 1.3.2
fsspec : 2021.10.0
fastparquet : 0.7.1
gcsfs : 2021.10.0
matplotlib : 3.4.3
numexpr : 2.7.3
odfpy : None
openpyxl : 3.0.9
pandas_gbq : None
pyarrow : 5.0.0
pyxlsb : None
s3fs : 2021.10.0
scipy : 1.7.1
sqlalchemy : 1.4.25
tables : 3.6.1
tabulate : 0.8.9
xarray : 0.18.2
xlrd : 2.0.1
xlwt : 1.3.0
numba : 0.53.1