FEAT/FIX support methods other than nearest for RangeIndex + end inclusiveness

[ianhi]

• Fixed the behavior of RangeIndex not being inclusive of the end of the slice (the pandas behavior)
• Allow method=None for RangeIndex slices, and for TransformIndex single point selection.

For the latter I had to play some small tricks using eps to account for floating point error. I think this is the least invasive way to account for the case mentioned in the comments (0.7 - 0.1) / 0.1 = 5.999999999999999 instead of 6.0

Examples:

import numpy as np
import xarray as xr
from xarray.indexes import RangeIndex

# Create a dataset with RangeIndex
index = RangeIndex.arange(0.0, 1.0, 0.1, dim='x')
ds = xr.Dataset(coords=xr.Coordinates.from_xindex(index))

# Slice selection with method="nearest" (rounds to nearest positions)
print(ds.sel(x=slice(0.12, 0.28), method="nearest"))
# Returns x=[0.1, 0.2, 0.3] - inclusive on both ends
print('\n\n----\n\n')
# Slice selection with method=None (exact matching with ceil/floor)
print(ds.sel(x=slice(0.15, 0.65)))
# Returns x=[0.2, 0.3, 0.4, 0.5, 0.6] 

print('\n\n----\n\n')
# Point selection now works without method for exact values
print(ds.sel(x=0.2))
# Returns x=0.2

<xarray.Dataset> Size: 24B
Dimensions:  (x: 3)
Coordinates:
  * x        (x) float64 24B 0.1 0.2 0.3
Data variables:
    *empty*
Indexes:
    x        RangeIndex (start=0.1, stop=0.4, step=0.1)


----


<xarray.Dataset> Size: 40B
Dimensions:  (x: 5)
Coordinates:
  * x        (x) float64 40B 0.2 0.3 0.4 0.5 0.6
Data variables:
    *empty*
Indexes:
    x        RangeIndex (start=0.2, stop=0.7, step=0.1)


----


<xarray.Dataset> Size: 8B
Dimensions:  ()
Coordinates:
    x        float64 8B 0.2
Data variables:
    *empty*

also passes @dcherian request in #10533 (comment) of equivalency to PandasIndex slicing

import numpy as np
import pandas as pd
from xarray.indexes import RangeIndex, PandasIndex

range_idx = RangeIndex.arange(20, dim='x')
pandas_idx = PandasIndex(pd.Index(np.arange(20)), dim='x')

range_result = range_idx.sel({'x': slice(4, 8)}).dim_indexers['x']
pandas_result = pandas_idx.sel({'x': slice(4, 8)}).dim_indexers['x']

assert range_result == pandas_result  # Both return slice(4, 9, None)

• Closes Support label-based slicing in RangeIndex #10533
• Tests added
• User visible changes (including notable bug fixes) are documented in whats-new.rst
• [NA] New functions/methods are listed in api.rst