Independent synthetic query snapshot lab / 독립 합성 조회 스냅샷 실험실
Python standard library only. Run python test_query.py and python run_examples.py after extraction.
No company records, server, database, real export file, external dependencies or network.
make_snapshot receives at most1000 already-read synthetic rows and a company/status/ID-substring query. The company scope must come from a trusted server principal. This is not authentication. The immutable tuple projects id/tick/quantity/status only. A real database must obtain a consistent authorized read separately; copying a supplied array cannot repair an inconsistent upstream read.
Cursor pagination uses tick plus ID ordering and a SHA-256 fingerprint of query/revision/projected contents. A cursor from another content snapshot/query is rejected, even when a caller reused the same revision. Summary, pages and export_rows use the exact same frozen records. Exports are returned Python values, not downloaded files. Report totals are synthetic quantities, not accounting results.
Cursor is base64 JSON, not encryption, signature, permission or a tamper-proof token. Tests intentionally show that a coherent cursor can be forged to jump to another existing position within the same snapshot. Production authorization and any required cursor signing remain separate. No persistent snapshot store, expiry, distributed snapshot service, live database transaction or unbounded dataset support.
Recorded JSON is actual local runner output. Viewer shows recorded results, not Python/DB execution in a browser. No actual client data, physical-device, accessibility or production-security audit.
수정·재배포·상업적 사용을 허용하며 무보증입니다. You may use, modify and redistribute these original synthetic files, including commercially, as-is without warranty.
