feat: Add date_partition_column_format for spark source

[joeyutong]

What this PR does / why we need it:

Add format option for date partition column of spark source. e.g.

spark_source = SparkSource(
    ...
    date_partition_column="effective_date",
    date_partition_column_format="%Y%m%d",
)

It is applied to the sql of pull_latest_from_table_or_query or get_historical_features

SELECT
...
WHERE effective_date >= '20210101' and effective_date <= '20210101'

Which issue(s) this PR fixes:

Misc