"Not enough data" quit in the multistate script crashes on keep_this #8

In discrete-multistate-intersection/01_prediction_multistate.R, the "Not enough data" quit at line 126 builds its message with length(keep_this), but keep_this is only created at line 134. The deliberate quit becomes Error: object 'keep_this' not found and "Execution halted". In the v1.0.6 run (#7) this happened for COGEM pathogenicity rating at strain rank, and a wrapper that checks the logs for R errors could not tell it from a crash. The binary script creates keep_this before its checks and is not affected.

Fix: count the attributes from fdat in that message, for example length(unique(fdat$Attribute)).

Drafted with Claude Code.