Complete DataAI Function Coverage for IRIS
dataai-spark-iris depends on dataai-spark-functions, so it exposes the full portable DataAI catalog without copying algorithms into the IRIS adapter.
| Family | APIs |
|---|---|
| ETL and declarative validation | DataAiPipeline, RuleSpec |
| Automatic quality diagnostics | DataQualityFunctions |
| Statistics, pivot, variance, ranking, correlation, regression | AnalyticsFunctions |
| Time summaries and rolling analysis | TimeSeriesFunctions |
| Outliers, anomalies, drift, Pareto, cohorts, funnels, KPIs | BusinessFunctions |
| Demand, pricing, elasticity, baskets, segments, churn, risk, inventory, profit, scenarios | MarketFunctions |
| Geographic readiness | MapFunctions |
| Cross-tabs and iterative matrix balancing | MatrixFunctions |
| Dictionaries, chart recommendations, local narratives, alerts | InsightFunctions |
Pass any result DataFrame to IrisFunctionOutputs.withRunMetadata(...), then use IrisDataFrames.writer(...) and explicitly select the IRIS table and save mode. Result records containing several DataFrames, such as map readiness, should be persisted as separate named IRIS tables.
The adapter does not execute external AI, host a service, or transmit data to DataAI. Deterministic narratives remain local; a customer may separately send governed summaries to its own approved AI integration.