Implement native function of `radians(68)`, `degrees(69)`

[slfan1989]

Summary

Add radians and degrees with Spark-compatible semantics, implemented natively on Arrow/DataFusion with vectorization. Cover common numeric types to improve performance on large columns and close gaps with Spark SQL function compatibility.

Background & Motivation

• Align with Apache Spark SQL’s function set to reduce migration/compatibility costs.
• Replace/augment non-vectorized UDF paths with native, vectorized implementations to lower overhead and increase throughput.
• Lay the groundwork for completing trigonometric and inverse-trigonometric function coverage.

Semantic Definition (Spark-compatible)

• radians(x): converts degrees x to radians, formula x * π / 180.
• degrees(x): converts radians x to degrees, formula x * 180 / π.
• Numeric widening: input may be any numeric type; output is DOUBLE (Float64).
• Null semantics: NULL input → NULL result.
• Special values: NaN and +/-Infinity must propagate per IEEE 754 without throwing exceptions.