Document Model Lifecycle Workflow

Design a small lifecycle manager that creates, combines, inspects, and cleans up document models backed by the provided cloud service. The workflow should be straightforward for anyone familiar with the SDK that manages document models.

Capabilities

Build a custom model

• Given an endpoint, API key, training container URL, and a model identifier, build a new model in the requested mode and return a summary with keys model_id, description, source, created_at, and operation_id. @test

Compose models for routing

• Given a target composed ID and an ordered list of component model IDs, create a composed model and return a summary with keys model_id, components (ordered), tags, and description. @test

Inspect catalog and quotas

• Listing models returns dictionaries containing model_id, created_at, and optional description, including those just built and composed in this workflow. @test
• Retrieving resource status exposes the current custom model count and limit, with the count never exceeding the limit. @test

Cleanup and operation tracking

• Deleting a model ID removes it so that subsequent listings do not contain that ID. @test
• Fetching an operation snapshot for a supplied operation ID returns a dictionary with status, last_updated, and percent_completed when provided by the service. @test

Implementation

@generates

API

from typing import Any, Dict, List, Optional

class ModelLifecycleManager:
    def __init__(self, endpoint: str, api_key: str) -> None: ...

    def build_custom_model(
        self,
        model_id: str,
        training_container_url: str,
        *,
        description: Optional[str] = None,
        mode: str = "neural",
        tags: Optional[Dict[str, str]] = None
    ) -> Dict[str, Any]: ...

    def compose_models(
        self,
        target_model_id: str,
        component_model_ids: List[str],
        *,
        description: Optional[str] = None,
        tags: Optional[Dict[str, str]] = None
    ) -> Dict[str, Any]: ...

    def list_models_summary(self) -> List[Dict[str, Any]]: ...

    def get_resource_quota(self) -> Dict[str, Any]: ...

    def delete_model(self, model_id: str) -> None: ...

    def get_operation_snapshot(self, operation_id: str) -> Dict[str, Any]: ...

Dependencies { .dependencies }

azure-ai-documentintelligence { .dependency }

Python SDK for managing and analyzing document intelligence models.