# `MimirAnalytics.Mappers.EvalReport`
[🔗](https://github.com/cash-mckeeman/mimir/blob/mimir_analytics-v0.8.1/mimir_analytics/lib/mimir_analytics/mappers/eval_report.ex#L1)

`eval_runs/*.json` report → `eval_outcomes` rows (the original eval-ingest
long format, widened with agent/suite/runtime/run_id). Idempotent by
file contents.

Eval reports are left in place, and `eval_outcomes` has no row key, so a
report rewritten after it was ingested would add its rows a second time.
New contents under a report name already ingested are therefore rejected
with `{:error, {:changed_after_ingest, name}}`, and an empty report with
`{:error, :empty_file}`. Reports have other readers, so a rejected report
stays where it is: the rejection is recorded in `ingest_rejections` by
content digest, and the same contents answer `:already_rejected` from
then on. New contents are judged afresh.

---

*Consult [api-reference.md](api-reference.md) for complete listing*