[Feature]: Compliance: per-entity scheduling + Databricks job refactor

Parent PRD

#125

What to build

Implement per-entity schedule overrides and refactor the Databricks compliance job:

• Per-entity scheduling: The schedule field on entity_compliance_checks supports: daily, weekly, on_change, manual.
• on_change trigger: Wire into entity update lifecycle hooks (on_update for contracts/products). When an entity with on_change bindings is updated, trigger run_entity_check for those bindings immediately.
• ComplianceManager.run_all_entity_checks(db, workspace_client, schedule_filter): Iterate all active bindings matching the schedule filter, run each. Respects last_run_at to avoid re-running within the schedule window.
• Job refactor: Add --mode parameter to compliance_checks.py: policies (existing sweep behavior) or entity_checks (calls run_all_entity_checks). Job can run both modes in sequence.
• Deprecate: Mark data_contract_validation.py job as deprecated (its logic is now in DSL check functions). Remove or keep as no-op with deprecation warning.

Acceptance criteria

• Schedule field respected: daily/weekly/on_change/manual
• on_change bindings triggered on entity update via lifecycle hooks
• run_all_entity_checks iterates active bindings for the given schedule
• Databricks job supports --mode=policies and --mode=entity_checks
• data_contract_validation.py deprecated
• last_run_at checked to avoid redundant runs within schedule window

Blocked by

• Blocked by #126 (entity_compliance_checks data layer)
• Blocked by #134 (lifecycle hooks)
• Blocked by #138 (run_entity_check execution engine)

User stories addressed

• User story 6
• User story 24
• User story 25

Start by reviewing blockers #126, #134, and #138, then inspect compliance_checks.py and data_contract_validation.py. Trace the entity update lifecycle and ComplianceManager entry points mentioned in the issue. Done means all listed schedule behaviors, mode handling, last_run_at checks, and deprecation behavior satisfy the acceptance criteria.