FEAT: Implement DB schema tracking with alembic

[behnam-o]

Summary

Adds Alembic-based, versioned memory schema migrations so DB schema changes are safe and repeatable across releases.

• Replaces ad-hoc table lifecycle logic with migration-driven lifecycle (upgrade/reset via revisions).
• Introduces the main operator/dev entrypoint: memory_migrations.py with:
  • generate: create a migration when memory models change (commit migration with model changes).
  • check: validate migrations resolve to current model schema.
• Enforces consistency in CI/dev flow via pre-commit check (fails when migrations and models diverge).
• Automatically applies migrations during memory DB initialization.
• Handles legacy unversioned DBs cautiously:
  • If existing PyRIT memory schema matches expected schema, it is stamped and continues.
  • If schema drift exists, startup migration fails (manual alignment or table reset required).
• Uses a namespaced version table (pyrit_memory_alembic_version) to avoid clashing with other Alembic-managed schemas.

[hannahwestra25]

just a few small comments!