DreamLake

Visualization .dreamrc

Visualization configuration remains owned by @dreamlake/viz/dataset-viz. The execution and registration loaders do not reinterpret these fields.

FieldMeaning
versionSchema version, currently 1.
nameOptional dataset display name.
storageOptional registered driver config; inherited from host storage when omitted. Explicit storage wins.
storage.driverRegistered storage driver; additional fields follow that driver's schema. No credentials.
dataset.formatRegistered dataset format.
dataset.episodesauto, supported glob, or discovery options.
dataset.episodes.globEpisode path glob; * supported, ** unsupported.
dataset.episodes.sortname, name-desc, or none.
dataset.episodes.limitEnumeration limit.
dataset.annotationsNamed annotation paths or declaration objects.
dataset.labelsField/dimension glob to display label mapping.
dataset.*Additional format-specific fields validated by that adapter.
views[]Explicit view/layout tree; no guessed automatic layout.
views[].viewRegistered view component; its bindings/options follow its schema.
views[].splitLayout direction: row, column, grid.
views[].childrenNested view nodes.

Complete existing documentation lives in the visualization project:

These are owning reference pages, not a second schema maintained by the CLI.

The schema is extensible: storage drivers, dataset formats, annotation codecs, field payload kinds and component options have their own contracts. The full typed field definitions and field validation are the authoritative reference for nested fields, bounds and component-specific options. This index does not impose a new wrapper or migrate existing files.