[REVIEW] Upgrade pandas to 1.2

[galipremsagar]

Fixes: #7367, #7446

This PR upgrades pandas to 1.2.2 in cudf. Changes include:

• Bumping up pandas version.
• Fixing isin behavior which now takes in types into accout: DOC: Undocumented change in .isin behavior from 1.1.5 to 1.2.0 pandas-dev/pandas#38781
• CategoricalColumn.__setitem__ will now not allow setting of values that are not in existing categories.
• Introduced cudf.core._compat.PANDAS_GE_120 variable to create back-ward compatibility.
• Updated usages of pd.core.tools.datetimes._guess_datetime_format to pd.core.tools.datetimes.guess_datetime_format
• Introduced std & median in DateTimeColumn.
• Fixed incorrect handling of passing StringMethods as an input to methods in string APIs.
• Fixed a typo in calling is_valid of Scalar.
• Removed unnecessary special handling in TimeDeltaColumn.sum logic for empty inputs.
• Introduced passing dtype='float64' wherever there is an empty series being created since pandas will soon be defaulting to object dtype if no type is passed and we don't have a perfectly resembling object dtype as that of pandas.
• Fixed deprecation warnings of Index.__or__ and Index.__xor__ by replacing with union & symmetric_difference APIs.
• Introduced mapping of our float32 & float64 dtypes to pandas Nullable dtypes FLoat32Dtype & Float64Dtype when nullable=True in to_pandas.
• With introduction of nullable float dtypes, there is an issue in creating MultiIndex from dataframe: BUG: Unable to create a MultiIndex with nan values in nullable Float dtypes pandas-dev/pandas#39984, so introduced a workaround in our MultiIndex.__repr__ code.
• Removed usages of check_less_precise in our code-base as this is deprecated and is replaced with rtol & atol. Retained its usages in our testing APIs for back-ward compatibility.
• Removed good number xfail cases which are actually passing right now because of resolved issues in both pandas & cudf.
• Did some miscellaneous code-cleanup in pytests.
• Fixed pytests that will fail when run in parallel due to access to shared pytest params being manipulated inplace.
• Follow a standard import pattern across pytest files, some files do from pandas import Series and some do from cudf.core import Series. So removed both patterns and doing only simple import cudf & import pandas as pd to avoid confusion while debugging test failures across multiple files. (Made this change in all pytest files which I had to touch as part of pandas upgrade, we can make similar changes in future for the files which we touch).
• Fix issue with assigning np.nan values to a CategoricalColumn and fix related __repr__ code: [BUG] Unable to assign null/nan in categorical column #7446

[galipremsagar]

rerun tests

[kkraus14]

@gpucibot merge

[kkraus14]

Does this need to be added to the integration repo as well?

[galipremsagar]

Would defer to @ajschmidt8 here, there was a build job-related failure at this line https://github.com/rapidsai/cudf/blob/branch-0.19/ci/cpu/build.sh#L78-L79 so AJ suggested we add it here.

[ajschmidt8]

yup, i will open a PR.

[ajschmidt8]

rapidsai/integration#225

Feel free to review. I will wait to merge until this PR is confirmed passing also.

[codecov]

Codecov Report

Merging #7375 (416bc92) into branch-0.19 (43b44e1) will increase coverage by 0.52%.
The diff coverage is 90.33%.

@@               Coverage Diff               @@
##           branch-0.19    #7375      +/-   ##
===============================================
+ Coverage        81.80%   82.33%   +0.52%     
===============================================
  Files              101      101              
  Lines            16695    17198     +503     
===============================================
+ Hits             13658    14160     +502     
- Misses            3037     3038       +1     

Impacted Files	Coverage Δ
python/cudf/cudf/utils/docutils.py	97.36% <50.00%> (-2.64%)	⬇️
python/cudf/cudf/testing/testing.py	80.00% <57.14%> (-1.04%)	⬇️
python/cudf/cudf/core/column/timedelta.py	88.57% <75.00%> (+0.33%)	⬆️
python/cudf/cudf/core/column/categorical.py	91.74% <77.14%> (-0.49%)	⬇️
python/cudf/cudf/core/column/datetime.py	89.63% <78.57%> (+0.58%)	⬆️
python/cudf/cudf/core/multiindex.py	82.90% <90.90%> (+0.73%)	⬆️
python/cudf/cudf/core/tools/datetimes.py	84.53% <93.33%> (+0.89%)	⬆️
python/cudf/cudf/core/column_accessor.py	95.47% <95.65%> (+2.53%)	⬆️
python/cudf/cudf/core/column/column.py	87.86% <96.42%> (+0.43%)	⬆️
python/cudf/cudf/core/_compat.py	100.00% <100.00%> (ø)
... and 62 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 7526be7...416bc92. Read the comment docs.