placement of new methods — Added generate_description, generate_title, get_trending_objects to ComputeService in compute/service.py

• Alternatives: Create a new CatalogAiService, Add to CatalogWorkspaceContentService
• Why: All existing AI methods (search_ai, ai_chat, sync_metadata) live in ComputeService; keeping AI catalog actions together in the same class is consistent with the established pattern.

return types for new methods — Return raw generated API types (GenerateDescriptionResponse, GenerateTitleResponse, TrendingObjectsResult)

• Alternatives: Create SDK wrapper classes, Return dict
• Why: Existing AI methods (ai_chat → ChatResult, search_ai → SearchResult) also return raw generated types; no SDK wrapper classes are used for action responses in this service.

enable_hybrid_search parameter position — Inserted enable_hybrid_search between deep_search and limit, alphabetically adjacent to deep_search

• Alternatives: Append at end
• Why: Maintains alphabetical ordering of optional boolean flags, consistent with how the existing optional params are ordered.

• The actions_api.generate_description / generate_title / trending_objects methods are available in the ActionsApi (confirmed in gooddata-api-client/gooddata_api_client/api/actions_api.py)
• GenerateDescriptionRequest requires object_id and object_type as positional args (confirmed in the generated model)
• The existing pytest.skip at module level in test_compute_service.py already covers the new tests, so no additional skip is needed

• test_generate_description and test_generate_title assume an object with id 'revenue' and type 'Metric' exists in the test workspace loaded from load/ai; if no such object exists the API may return an error response

• public_api — New methods on ComputeService; no new public exports needed since ComputeService was already exported
  • packages/gooddata-sdk/src/gooddata_sdk/compute/service.py
• tests — Integration tests for the four new/updated methods
  • packages/gooddata-sdk/tests/compute/test_compute_service.py