[FEATURE]: Unify parsing semantics for geospatial checks #1513

Is there an existing issue for this?

• I have searched the existing issues

Problem statement

Spark 4.2 introduced native geometry and geography types for DataFrames. DQX offers a module for validating properties of geospatial data. However, this module does not:

• Consistently accept and handle geometry and geography columns as check function inputs
• Consistently generate metadata when checks are skipped due to invalid WKB/WKT/GeoJSON parsing; These currently appear in results as NULL when the input values are invalid

Proposed Solution

• Update the geospatial check functions to allow geometry and geography columns
• Create metadata messages when checks are skipped because the input WKB/WKT/GeoJSON values are invalid

Additional Context

These changes were introduced for is_geo_within_distance in #1510 . This issue expands that work to other geospatial check functions.