SNOW-704114: Implement `DataFrame.summary` to show percentiles

Update: As of Mar.27 2023, @sfc-gh-stan and @sfc-gh-gfrere discussed this and concluded we should implement DataFrame.summary() to display the percentiles instead of modifying the existing DataFrame.describe().

What is the current behavior?

    Example::
        >>> df = session.create_dataframe([[1, 2], [3, 4]], schema=["a", "b"])
        >>> desc_result = df.describe().sort("SUMMARY").show()
        -------------------------------------------------------
        |"SUMMARY"  |"A"                 |"B"                 |
        -------------------------------------------------------
        |count      |2.0                 |2.0                 |
        |max        |3.0                 |4.0                 |
        |mean       |2.0                 |3.0                 |
        |min        |1.0                 |2.0                 |
        |stddev     |1.4142135623730951  |1.4142135623730951  |
        -------------------------------------------------------

Percentiles are not shown.

What is the desired behavior?

Show percentiles 25, 50, 75 like Pandas.describe(). Using updated approximate percentile
'alter session set APPROX_PERCENTILE_EXACT_IF_POSSIBLE = true;’

How would this improve snowflake-snowpark-python?

It would match Pandas behavior and is requested by customers. Given the improvements being made to APPROX_PERCENTILE% we should implement this change. Before we may have thought this was too expensive or inaccurate due to APPROX_PERCENTILE% issues on few samples.

References, Other Background

APPROX_PERCENTILE is being updated by the SQL Compiler team to allow for a larger buffer usage and if the buffer doesn't fill PERCENTILE_CONT is automatically used.
https://snowflakecomputing.atlassian.net/browse/SNOW-704037

Start by locating the DataFrame.summary() and existing DataFrame.describe() entry points in the snowpark-python codebase, then inspect how summary statistics are assembled. Verify the expected percentile behavior against the provided example and Pandas.describe(), including the requested approximate-percentile session setting. Done means DataFrame.summary() displays the 25th, 50th, and 75th percentiles without changing DataFrame.describe().