Evaluation Run Imports — Calendar Date Check

Choose this endpoint when you need a fast, low-cost ($0.01 USDC) check specifically on calendar date columns within evaluation run import CSVs, especially when you want to catch invalid or out-of-range dates before ingesting data into a pipeline. It is purpose-built for the evaluation run import schema (evaluation_id, evaluated_on, dataset_id, correct_count, example_count) and supports quoted cells, configurable delimiters, date range bounds, and category allowlists — making it more targeted than a general-purpose CSV validator.

• CSV exceeds 100,000 character limit — request rejected
• CSV has zero rows or is empty — validation cannot proceed
• Delimiter mismatch causes rows to parse incorrectly, producing false positives
• Date column not present or misnamed in headers — no dates to validate
• Rows exceed 1000 limit — only first 1000 rows processed
• Cell length exceeds max_cell_length setting — cell rejected
• Malformed CSV quoting causes parse errors