Dataset QA

Dataset QA

Dataset QA is the release-control process that determines whether records are ready for public exposure.

What Dataset QA Checks

• Source URL presence and source fit.
• Current hash integrity.
• Snapshot and version record alignment.
• Check-log coverage.
• Retrieval status and last successful check.
• Public evidence eligibility.
• Archive timestamp evidence.
• KPI assessment coverage.
• Regional impact coverage.
• Subscriber hygiene.
• Admin access and review log consistency.

Public QA Rule

When anomalies are detected in the latest fetching or update cycle, the affected source is temporarily suspended. The public interface may show minimal metadata and a suspension notice, but it should not publish analysis based on uncertain evidence.

Issue Decisions

Dataset QA issues can be:

• open;
• reviewed;
• ignored with reason;
• reopened.

Every decision should write an append-only review-log event.

Typical Remediation Workflow

flowchart LR
    issue["QA issue"] --> inspect["Inspect source, log and final URL"]
    inspect --> remediate["Update URL or scope"]
    inspect --> suspend["Keep source suspended"]
    remediate --> scan["Run targeted scan"]
    scan --> pass["Available / Reviewed"]
    scan --> fail["Needs Review / Unavailable"]
    fail --> issue

Loading

CLI Support

Useful commands:

npm run qa:dataset
npm run db:repair
npm run db:backfill-check-logs

For Hostinger SSH environments where npm or npx may not be available, use the dedicated Hostinger helper scripts documented in the deployment guide.