Skip to content

6.9.2.9 ScoringBatch

6.9.2.9.1 Purpose

One batch scoring: a deployed model version reads the farm data of a scope at a write number, records a run and appends its predictions to farm as computed records, one write per farm. The Console starts batches and follows them.

6.9.2.9.2 Questions it answers

  • Which farms and period did a batch score, at which write number?
  • Did it finish, and how many predictions did it write?
  • Which run makes its outputs reproducible?

6.9.2.9.3 Fields

Field Type Cardinality Description
id string (uuid) 1..1 Key.
model_version ModelVersion 1..1
scope string 1..1 Farms, animals and period covered; same meaning as 62.7.
data_seq WriteCommit 1..1
status enum 1..1 Values: queued, running, done, failed. Default: queued.
model_run ModelRun 0..1 Set when the batch starts reading.
predictions_written integer 1..1 Default: 0.
started_at datetime 0..1
finished_at datetime 0..1
last_error string 0..1
created_by MoonitoringUser 1..1
created_at datetime 1..1
updated_at datetime 1..1

6.9.2.9.4 Relationships

References

6.9.2.9.5 Rules & constraints

  • One batch scoring: reads farm data at a write number with farm.canonical_data.read_all (farm.canonical_data.read_all), records a ModelRun and appends the predictions to farm as computed records, one WriteCommit per farm.