Fix unnecessary ValueError in PairPlot: Caused by unrelated duplicated columns not used in vars

[leakyH]

A simple demo:

import seaborn as sns
import pandas as pd
df = pd.DataFrame(dict(x=rs.normal(size=60),
                           y=rs.randint(0, 4, size=(60)),
                           z=rs.gamma(3, size=60),
                           z2=rs.gamma(6, size=60),
                           }
df_with_dupe = df.copy()
df_with_dupe.columns = ["x", "y", "z", "z"] #sometimes by mistake, or the z/z2 are not important
sns.pairplot(df_with_dupe, vars=['x', 'y'])  # raise ValueError

The Traceback:

> Traceback (most recent call last):
>   File "/data1/home/----/plotReferenceMap.py", line 153, in <module>
>     sns.pairplot(df_merge,
>   File "/data1/home/--/lib/python3.9/site-packages/seaborn/axisgrid.py", line 2119, in pairplot
>     grid = PairGrid(data, vars=vars, x_vars=x_vars, y_vars=y_vars, hue=hue,
>   File "/data1/home/--/lib/python3.9/site-packages/seaborn/axisgrid.py", line 1251, in __init__
>     numeric_cols = self._find_numeric_cols(data)
>   File "/data1/home/--/lib/python3.9/site-packages/seaborn/axisgrid.py", line 1674, in _find_numeric_cols
>     if variable_type(data[col]) == "numeric":
>   File "/data1/home/--/lib/python3.9/site-packages/seaborn/_base.py", line 1498, in variable_type
>     vector = pd.Series(vector)
>   File "/data1/home/--/lib/python3.9/site-packages/pandas/core/series.py", line 367, in __init__
>     if is_empty_data(data) and dtype is None:
>   File "/data1/home/--/lib/python3.9/site-packages/pandas/core/construction.py", line 818, in is_empty_data
>     is_simple_empty = is_list_like_without_dtype and not data
>   File "/data1/home/--/lib/python3.9/site-packages/pandas/core/generic.py", line 1527, in __nonzero__
>     raise ValueError(
> ValueError: The truth value of a DataFrame is ambiguous. Use a.empty, a.bool(), a.item(), a.any() or a.all().

The error happens inside self._find_numeric_cols(data), which is unnecessary when vars is provided. So I skip it and extend to some other similar scenarios:

1. no duplication in df.columns, but duplications in vars: Gives a simple warning. It just generates unexpected figures but does not crash.
2. duplication in df.columns, and one of the duplicants is included in vars: raise ValueError in PairGrid Class, specify the related duplicants.
3. duplication in df.columns, and vars is not provided: raise ValueError in PairGrid Class, specify the all duplicants.

These tests are all included in the test_axisgrid.py

Please let me know if any other modifications are needed.