feat(bigquery-jdbc): Migrate getImportedKeys and getCrossReference to BQ API (#13692)

b/532249777

This PR migrates the `getImportedKeys` and `getCrossReference` JDBC
metadata methods in `BigQueryDatabaseMetaData` to retrieve constraints
natively via the Google Cloud Java Client API instead of running legacy
raw SQL files on BigQuery.

## Changes

### **API-based Implementations**
- Migrated `getImportedKeys(...)` to fetch and scan foreign keys from
standard table definitions.
- Migrated `getCrossReference(...)` to matching parent and foreign table
references using client-side definitions.
- Utilized the shared parallel table processor utilities
(`processTargetTablesConcurrently(...)` and `processSingleTable(...)`)
to run dataset scans concurrently.

### **Sorting & Correctness**
- Separated sorting behavior according to the JDBC specification:
- `getImportedKeys` is sorted by parent table: `PKTABLE_CAT,
PKTABLE_SCHEM, PKTABLE_NAME, KEY_SEQ`. Introduced
`definePkTableSortComparator` for this.
- `getCrossReference` (and future `getExportedKeys`) is sorted by child
table: `FKTABLE_CAT, FKTABLE_SCHEM, FKTABLE_NAME, KEY_SEQ`. Uses
`defineFkTableSortComparator`.
- Introduced a private helper `matchesOrNullWildcard(...)` to simplify
and clean up catalog/schema bounds comparisons.

Author: keshavdandeva