[BUG] DataFrame.rank() ignores axis option

[tfeher]

Describe the bug
The DataFrame.rank() method ignores the axis option and ranks always along axis=0.

Steps/Code to reproduce bug

import cudf
import numpy as np
gdf = cudf.DataFrame(data={'A': [4, 2, 4, 8, np.nan], 'B':[5,4,3,2,1]})
gdf.rank(axis=1)

This gives

A	B
2.5	5.0
1.0	4.0
2.5	3.0
4.0	2.0
NA	1.0

Expected behavior

Rank along columns when axis=1 is specified. Here is the expected output produced by pandas

import pandas as pd
df = pd.DataFrame(data={'A': [4, 2, 4, 8, np.nan], 'B':[5,4,3,2,1]})
df.rank(axis=1)

Prints:

A	B
1.0	2.0
1.0	2.0
2.0	1.0
2.0	1.0
NaN	1.0

Environment overview

Tested with 0.20 nightly docker image (rapidsai/rapidsai-core-dev-nightly:0.20-cuda11.2-devel-ubuntu18.04-py3.8).

Additional context
The ranking method is used by cheminformatics

[mlgill]

Thanks for filing this @tfeher -- this exactly the describes the bug as I initially identified it.

[kkraus14]

We should at least be throwing for now that we don't yet support axis=1.