Use names from cell type ontology (CL) instead of blueprint names in consensus reference

[allyhawkins]

So I guess this is fine, but I guess I'm not sure what the value of it is? If we are just getting an alternative to the ontology label with a different naming convention, do we need it at all?

Perhaps the better approach here is to directly map the ontology ids to the name associated with them ourselves and remove the blueprint_annotation_fine column all together. I think I'll go ahead and do this but in the script that creates the consensus reference to begin with and not here. I'll file a new issue and new PR to address this before we actually run this on all ScPCA samples.

Originally posted by @allyhawkins in #977 (comment)