[Model Registry] Implement ModelVersion stage state machine

[jaynomyaro]

Summary

This PR introduces a stage-based state machine for ModelVersion, providing controlled lifecycle management for machine learning model versions. The implementation enforces valid state transitions between stages, ensuring models progress through the registry in a predictable and auditable manner.

Changes Made

• Implemented a state machine for ModelVersion lifecycle management.
• Defined supported stages (e.g., Draft, Training, Validated, Staging, Production, Archived) and their allowed transitions.
• Enforced validation to prevent invalid or out-of-order stage transitions.
• Added helper methods to transition between stages through a consistent API.
• Recorded stage transition timestamps and metadata where applicable.
• Improved error handling with descriptive exceptions for invalid state changes.
• Added unit tests covering valid transitions, invalid transitions, terminal states, and edge cases.
• Updated developer documentation with the state machine design, transition diagram, and usage examples.

Benefits

• Ensures consistent and predictable model lifecycle management.
• Prevents invalid stage transitions that could compromise deployment workflows.
• Improves auditability by tracking model progression through each lifecycle stage.
• Simplifies future enhancements such as approval workflows, automated promotions, and rollback support.
• Increases reliability through comprehensive validation and automated testing.

Testing

• Verified all valid stage transitions execute successfully.
• Confirmed invalid transitions are rejected with appropriate error messages.
• Tested terminal states to ensure no unauthorized transitions occur.
• Validated transition metadata and timestamp recording.
• Ran the complete test suite to ensure no regressions.
  closed [Model Registry] Implement ModelVersion stage state machine #313

[drips-wave]

@jaynomyaro Great news! 🎉 Based on an automated assessment of this PR, the linked Wave issue(s) no longer count against your application limits.

You can now already apply to more issues while waiting for a review of this PR. Keep up the great work! 🚀

Learn more about application limits