Mirrors df into a queue RDS and adds status columns: status, claimed_by, started_at, finished_at.

Adds metadata columns used by workers:

  • status: PENDING | RUNNING | DONE | FAILED

  • claimed_by: tmux pane id that claimed the row

  • started_at: "YYYY-MM-DD HH:MM:SS"

  • finished_at: "YYYY-MM-DD HH:MM:SS"

  • DEoptimElapsedTime: numeric seconds (sum(diff(allIterations[allIterations < 20 minutes])))

  • machine_name: Sys.info()[["nodename"]]

  • process_id: Sys.getpid()

  • heartbeat_at: latest timestamp (as character) detected by heartbeat

  • heartbeat_iter: latest iteration number (integer) detected by heartbeat

tmuxPrepareQueueFromDF(df, queue_path)

tmuxPrepareQueueFromDF(df, queue_path)

Arguments

df

data.frame; experiment rows (columns become objects in workers)

queue_path

character; path to the queue .rds (absolute recommended)

Value

Invisibly returns queue_path.

Invisibly returns queue_path.