Start an import (parse + suggested mapping)

Starts an import from a spreadsheet. Two intake shapes: a multipart request carrying the file (CSV/XLS/XLSX, ≤10MB — larger files go through POST /business/imports/presign first) or a JSON body naming a presigned upload's importId. The call only parses: it answers 200 with the mapping-confirmation payload — rowCount, the detected headers, up to five sampleRows, and suggestedMapping (source header → standard field label such as First Name, Email, Company; null when nothing was suggested) — and stores that suggestion on the import. Nothing is written to any list until POST /business/imports/{importId}/commit, which is the only call that takes a destination. Parsing spends no credits.

Errors:

• 400 IMPORT_FILE_MISSING — Neither a multipart file part nor an importId was sent.
• 404 IMPORT_NOT_FOUND — importId does not decode to an upload of the token owner, or the uploaded object is gone.
• 409 IMPORT_ALREADY_COMMITTED — The import was already committed; a committed file is never re-parsed. Upload it again for a new import.
• 422 IMPORT_FILE_UNSUPPORTED — The multipart part carries no Content-Type or one that is not CSV/XLS/XLSX, or the file could not be parsed.
• 422 IMPORT_FILE_TOO_LARGE — Multipart file over 10MB, or a presigned object over 50MB.
• 422 IMPORT_FILE_EMPTY — The file has no data rows.
• 422 VALIDATION_FAILED — A body key other than importId (destination fields belong to the commit call).
• 429 RATE_LIMITED — The lists rate bucket (120 requests/minute, 10,000/day per token owner) is exhausted.

Requires one of the following token scopes: lists.