Evaluation Run Imports — Trimmed Typed Conversion

Choose this endpoint when you need to validate and type-normalize a structured CSV of evaluation run records (with fields like evaluation_id, evaluated_on, dataset_id, correct_count, example_count) against a declared table contract, with support for quoted cells, category allow-lists, and numeric range checks — all in a single cheap call with no external enrichment. Prefer it over generic CSV validators when your data specifically follows the evaluation run imports schema and you need trimmed, typed output ready for downstream ingestion.

• CSV exceeds 1000 rows — only the first 1000 may be processed
• Cell value exceeds max_cell_length limit — row flagged as invalid
• Category value not in allowed_categories list — validation error reported
• Numeric value outside minimum/maximum bounds — type conversion failure
• Malformed CSV (unclosed quotes, mismatched columns) — parse error returned
• Empty CSV input — error due to minLength:1 constraint
• Delimiter mismatch — rows parsed incorrectly if wrong delimiter specified