# CSV structural preflight

Original local Python 3 checker. No installation or third-party dependencies. Reads JSON from standard input and prints structural findings; makes no network requests and writes no input data.

Run:

    python3 check.py < example.json

Review:

    python3 -m unittest discover -s .

Input is exactly {"csv_text": "..."}, up to 8,000 characters and 16,000 UTF-8 bytes. First logical record is the header. Up to 200 data records and ten fields in every record. Values, whitespace and leading zeros are preserved. Quoted embedded newlines belong to their existing logical record. Blank records are included.

Reports duplicate header names and positions, uneven record numbers and field counts, and exact duplicate complete record groups. Uses Python standard-library csv.reader(strict=True), rather than promising universal CSV dialect validation. Malformed quotes return parse_error without claiming complete counts; oversized record inventories return out_of_scope. Wrong transport inputs exit with status 2.

This prototype outputs detailed structured JSON, not a marketplace's status/report-only contract. It is not an accepted buyer submission, a paid marketplace contract, an import approval, a native Dots integration or evidence of payment. A Dot using this file needs an authorized local Python environment. No source data is corrected automatically.

The original example and tests are synthetic. The historical public CSV job informed the problem selection, but its supplied data and worker delivery are not included.