Ids of every dataset this analysis is attached to. The per-dataset lookup key.
Stable handle minted by makeAnalysisId (name + sorted source ids).
OptionaldescriptionOptional human-readable description of what the analysis computes.
Bounded result timeline, oldest first (see MAX_HISTORY_PER_ANALYSIS).
The analysis name. With its sorted source ids it forms analysisId.
How each result's data should be interpreted — plain JSON or a Chart.js configuration.
One analysis tracked by the session-level Analyses registry, the single home for every
analyseDatasetsanalysis, one-shot or monitored, across every dataset kind. It is pure sink state, metadata plus the bounded resulthistory. The recipe (code/inputs) and recurrence are NOT here. They live on the analysis's Job (see_jobs).affectedDatasetIdsis the per-dataset lookup key (Analyses.getAnalysesForDataset).historyis the result timeline (oldest first), whose latest is the current result, threaded back into the sandbox code asprevious.