[SPARK-44734][PYTHON][DOCS] Expand PySpark type conversion guide

[RaghunandanKumar]

What changes were proposed in this pull request?

This updates the PySpark type conversion guide to make it more useful for day-to-day users.

Changes:

• remove unfinished TODO placeholders from the page
• fix a small formatting typo in the DecimalType row
• add practical notes about runtime conversion behavior, including:
  • null handling
  • fixed-width numeric limits
  • schema inference pitfalls for local Python data
  • nested dict inference as MapType vs StructType
  • BinaryType behavior under spark.sql.execution.pyspark.binaryAsBytes
• add a short section describing Arrow-enabled Python UDF conversion caveats
• add config-driven examples for:
  • spark.sql.pyspark.inferNestedDictAsStruct.enabled
  • spark.sql.timestampType=TIMESTAMP_NTZ
  • spark.sql.execution.pyspark.binaryAsBytes

Why are the changes needed?

The page had visible TODO markers and was missing several practical conversion cases that commonly confuse PySpark users, especially around schema inference and configuration-dependent behavior.

Does this PR introduce any user-facing change?

Yes, documentation only.

How was this patch tested?

Reviewed the rendered diff locally.

Full Spark docs/test validation was not run in this environment because Java is not installed.