v0.1.84 - Update Platform API contract (#113)

🌟 Summary

Release v0.1.84 updates the Platform API and Python SDK, adds a new image-prediction model option, and improves cloud CLI reliability.

📊 Key Changes

• New image prediction option: Added muse-wonder-1.0 to the Platform API contract and both synchronous and asynchronous Python SDK methods.
• Versioned API and SDK: Updated the OpenAPI contract and ultralytics-platform package from 0.1.83 to 0.1.84, with generated SDK changes aligned to the API version.
• Cloud training: ul cloud train now accepts capture_dataset_version to save an immutable dataset version for a training run.
• Clearer cloud job outcomes: Failed or cancelled cloud jobs now exit with code 1, distinguishing them from local input errors, which continue to exit with code 2.
• Earlier URI validation: Malformed ul:// model and data addresses are rejected with a clearer error before uploads or Platform requests.
• Fewer duplicate alerts: Workflow reruns no longer send duplicate Slack notifications.

🎯 Purpose & Impact

• Python SDK users can select muse-wonder-1.0 for image predictions with up-to-date API types.
• Capturing dataset versions makes it easier to preserve and identify the exact dataset used for a training run.
• More consistent exit codes help scripts and automation respond correctly to cloud job failures.
• Earlier URI checks make CLI errors easier to understand and avoid unnecessary requests.
• Suppressing rerun notifications keeps Slack alerts focused on the original workflow attempt.

What's Changed

• Suppress Slack notifications on workflow reruns by @glenn-jocher in #109
• Exit 1 when a cloud job fails or is cancelled by @JaviChulvi in #110
• Accept capture_dataset_version in ul cloud train by @JaviChulvi in #111
• Validate ul:// model and data URIs before cloud workflows run by @JaviChulvi in #112
• Update Platform API contract by @UltralyticsAssistant in #113

Full Changelog: v0.1.83...v0.1.84