Implement Model Configuration REST API Endpoints

[dominiquekleeven]

Implement the full CRUD API for managing ML model configurations.
This feature depends on Issue #8, which implements the model configuration structure and provides a storage service for it.

Deleting a model configuration must also delete any associated trained model.

---

API Endpoints

• Create model configuration
• Read all model configurations
• Read single model configuration
• Update model configuration
• Delete model configuration and its trained model

---

Validation & Constraints

Use Pydantic for validation (already part of the project). Ensure the following:

• ISO8601 datetime string validation
• Epoch millis validation
• Asset ID validation (base22 format)
• Realm consistency constraint:
  • The target and all regressors must be from the same realm
• Forecast validity:
  • Regressors must include predicted values that cover the forecast period

Some of these may require extending the existing MLModelConfigService to support:

• Asset realm validation
• Regressor realm consistency and forecast coverage

---

Notes

• Security and permission handling for these endpoints is covered in Issue Keycloak Integration #3
• Ensure meaningful error messages are returned when the API operation fails e.g. validation error, resource not found etc.