This PR migrates the BigQuery JDBC read-path implementations (Arrow and Json) to utilize the newly introduced BigQueryTypeRegistry for unified type lookups and high-performance coercion.

Also made the following fixes in implementation and tests:

Type Registry Consolidation & Read-Path Migration

• Unified Coercion: Replaced the legacy BigQueryTypeCoercer with BigQueryTypeRegistry as the single source of truth for all type casting and format mapping.
• Result Set Integration: Refactored BigQueryBaseResultSet, BigQueryArrowResultSet, and BigQueryJsonResultSet (along with their nested Struct and Array counterparts) to funnel all getObject() and primitive getter calls through the registry.

Temporal Type & JDBC 4.2 Spec Compliance

• Strict UTC Anchoring: Anchored java.sql.Date to strictly format at UTC midnight, and java.sql.Time to 1970-01-01 in the local JVM, ensuring correct String serialization and civil time preservation.
• Illegal Coercion Blocking: Added explicit guards to throw BigQueryJdbcException for invalid cross-temporal conversions (e.g., blocking java.sql.Date -> java.sql.Time and java.sql.Time -> java.sql.Date), preserving backward compatibility with legacy strictness.
• Formatter Optimization: Extracted DateTimeFormatter compilations into static final constants to guarantee fast-path string serialization for TIMESTAMP (6 decimal precision) and TIME (3 decimal precision) without per-row object allocation penalties.

Complex Types & Boundary Protection

• Arrow Primitives: Handled raw Arrow Integer (days since epoch) and Long (microseconds since epoch) unboxing explicitly inside Arrow structures to bypass global registry mutations, fixing nested RANGE and ARRAY rendering.
• Strict Bounds Checking: Implemented explicit overflow/underflow protections (MAX_VALUE / MIN_VALUE guards) when coercing BigDecimal and Long values down to smaller primitives (Integer, Short, Byte) to prevent silent data corruption.