Section 1Validation run
05_validate_rfi.py read data/processed/rfi_clean.csv and data/processed/rfi_excluded.csv. It did not change either file. The results are saved in data/processed/validation/validation_results.csv.
All 9 checks passed. The clean table still has 107 RFIs, numbered RFI-001 through RFI-107, and the five excluded rows are still outside it.
Section 2Check results
| Check | Result | Detail |
|---|---|---|
| columns | PASS | rfi_id, from_party, tasked_to, date_received, question, response, notes, status, response_available |
| row_count | PASS | 107 rows |
| rfi_id_sequence | PASS | RFI-001 to RFI-107 |
| question_present | PASS | 0 blank |
| from_party_present | PASS | 0 blank |
| tasked_to_present | PASS | 0 blank |
| date_received_iso | PASS | 0 invalid |
| status_matches_response | PASS | 0 mismatches |
| excluded_rows_absent | PASS | 5 excluded, 0 still in the clean table |
Section 3What each check covers
The column check requires those nine fields, in that order. The identifier check requires the rows to stay in order from RFI-001 to RFI-107, with no gaps and no duplicates.
Every kept row must have a question, a submitting party, a tasked party, and a real calendar date. Answered is allowed only when a response is present and response_available is Yes. Open is allowed only when the response is blank and response_available is No.
The last check reads the five excluded identifiers — the section label and numbers 108 through 111 — and confirms none of them appear in the clean table.