feat(connector): add Amazon Athena connector via Glue Data Catalog (#309)

* feat(connector): add Amazon Athena connector via Glue Data Catalog

* fix(athena): address reviewer feedback

* fix(athena): wire scope discovery, fix normalizeDriver, tighten types and tests

* fix(athena): honor databases scope, wire sql-analysis dialect, harden config resolution

- introspect() limits to the configured `databases` scope instead of scanning
  every Glue database in the account (docs promised this; connector ignored it)
- add athena -> athena to sql-analysis SQLGLOT_DIALECTS so `ktx sql` and MCP
  read-only validation parse Athena SQL under the Trino grammar, not postgres
- stringConfigValue coerces a resolved-empty `env:` reference to undefined so
  optional fields fall back to their defaults (workgroup 'primary', catalog
  'AwsDataCatalog') instead of ''
- drop trailing whitespace in dialect.test.ts

* fix(athena): integrate with main's SQL/non-SQL dialect split and add dialect notes

Rebase onto main, which introduced the KtxDialect (core) vs KtxSqlDialect
(SQL-only) split for MongoDB:
- KtxAthenaDialect implements KtxSqlDialect; the connector resolves it via
  getSqlDialectForDriver so SQL-generation methods stay in scope
- add authored athena.md SQL notes for the sql_dialect_notes MCP tool, required
  now that athena resolves to the athena sqlglot dialect (dialect-notes coverage
  is derived from the warehouse-driver registry)

---------

Co-authored-by: Andrey Avtomonov <andreybavt@gmail.com>

packages/cli/test/context/sql-analysis/dialect.test.ts

@@ -12,6 +12,7 @@ describe('sqlAnalysisDialectForDriver', () => {
     expect(sqlAnalysisDialectForDriver('duckdb')).toBe('duckdb');
     expect(sqlAnalysisDialectForDriver('clickhouse')).toBe('clickhouse');
     expect(sqlAnalysisDialectForDriver('databricks')).toBe('databricks');
+    expect(sqlAnalysisDialectForDriver('athena')).toBe('athena');
   });
 
   it('maps local connection-type spellings to sqlglot dialects', () => {