[Feature]: Compliance: SCHEMA_DRIFT check function

Parent PRD

#125

What to build

Implement the SCHEMA_DRIFT(physical_name) DSL check function end-to-end:

• Uses EvaluationContext.workspace_client to run DESCRIBE TABLE on the physical UC table
• Compares UC columns/types against contract schema properties from the enriched entity dict (requires schema facet)
• Detects: missing columns (in contract but not in table), extra columns (in table but not in contract), type-changed columns
• Returns CheckResult with pass/fail, human-readable message, and details_json containing the drift specifics
• Reuses logic from the existing data_contract_validation.py check_schema_drift function
• Register in the check function registry
• Tests with mocked WorkspaceClient covering: no drift, missing columns, extra columns, type changes, empty schema

Acceptance criteria

• SCHEMA_DRIFT function registered in check function registry
• Compares contract schema properties against UC DESCRIBE TABLE output
• Correctly identifies missing, extra, and type-changed columns
• Returns structured details_json with drift specifics
• Tests with mocked WorkspaceClient cover all drift scenarios

Blocked by

• Blocked by #127 (EvaluationContext + check function registry)
• Blocked by #128 (facet-based entity enrichment for schema facet)

User stories addressed

• User story 9

This work depends on #127 and #128; start with the EvaluationContext and check function registry entry points, then review the existing check_schema_drift function in data_contract_validation.py. Use mocked WorkspaceClient tests for no drift, missing columns, extra columns, type changes, and an empty schema; done means the registered SCHEMA_DRIFT function returns structured drift details and correct pass/fail results.