ml-app-deployer
ml-app-deployer copied to clipboard
mlPrintForestPlan doesn't look at payload-driven forest creation
If you set up forests under ml-config/forests/my-db/my-db-forests.json
and then run mlPrintForestPlan
, the output will not be based on the content of my-db-forests.json
.
This is by design - the preview feature is to show what forests will be produced based on the properties. Whereas the "custom forest" approach - i.e. ml-config/forests/(database)/file.json - explicitly defines the forests, and thus the preview feature isn't needed.
This may be a scenario not designed for yet - using custom forests, but then also use properties to generate replicas. Is that correct?
Correct. It would be useful to get a preview showing replicas if they are property-driven, even if the forests are payload-driven.
Prioritizing this a bit lower for now