AiConfig is enough.
>>> import atoti as tt
>>> from atoti_ai import AiConfig
>>> session_config = tt.SessionConfig(ai=AiConfig())
atoti_ai.AiConfig.connection and atoti_ai.AiConfig.chat are both not None:
>>> summary_unavailable_message = (
... "Summary generation is not available (no AI provider configured)"
... )
>>> summary_unavailable_message in _fetch_autoexplain_summary(session_without_chat)
True
>>> summary_unavailable_message in _fetch_autoexplain_summary(session_with_chat)
False
See also:
atoti.Cube.auto_explain.excluded_hierarchies | The hierarchies that are never explored. |
excluded_hierarchies_per_measure | The per-measure overrides of excluded_hierarchies, keyed by measure name. |
included_hierarchies | The hierarchies exploration is restricted to. |
included_hierarchies_per_measure | The per-measure overrides of included_hierarchies, keyed by measure name. |
max_depth | The maximum number of hierarchies the algorithm drills down through. |
max_distinct_hierarchies | The maximum number of distinct hierarchies reported across the explanation. |
max_entropy | The maximum entropy above which a level is considered too uniform to explain. |
max_members_per_level | The maximum number of members reported per level. |
min_percentage_absolute_contribution | The minimum absolute contribution (as a percentage) a member must reach to be reported. |
min_percentage_relative_contribution | The minimum relative contribution (as a percentage) a member must reach to be reported. |
min_variation_threshold | The minimum variation a contributor must reach to be reported. |
opposite_contributors_included | Whether contributors moving in the opposite direction of the overall variation are also reported. |