Query completion should use the feature-model schema as its source of truth when schema metadata is available.
The current completion path derives field candidates from the runtime model and enum-like constants from uppercase strings in the string pool. For MapViewer search this makes completion dependent on the loaded sample data and can make completion expensive on large feature models.
Use the schema to derive field and enum-symbol candidates deterministically, so users get stable completion results for the same feature model and completion does not need to inspect large loaded model data.
Query completion should use the feature-model schema as its source of truth when schema metadata is available.
The current completion path derives field candidates from the runtime model and enum-like constants from uppercase strings in the string pool. For MapViewer search this makes completion dependent on the loaded sample data and can make completion expensive on large feature models.
Use the schema to derive field and enum-symbol candidates deterministically, so users get stable completion results for the same feature model and completion does not need to inspect large loaded model data.