Persist CDN invalidation requests and virus scan results

[SYLVIANNORUKA]

Description: CdnService.invalidations and VirusScanService.results are both process-local Maps; GET /cdn/invalidations/:id and any future virus-scan-status endpoint return nothing after a restart, and in a horizontally-scaled deployment a request handled by one instance can never see the invalidation/scan state recorded by another, making CdnController.getInvalidation/listInvalidations unreliable in production.

Requirements:

• Add a CdnInvalidationRequest Prisma model (id, paths JSON, status, requestedAt, completedAt, error) and a FileScanResult Prisma model (fileId, status, startedAt, completedAt, details, attempts).
• Replace the in-memory invalidations Map in CdnService and results Map in VirusScanService with Prisma persistence.
• Add GET /uploads/scans/:fileId (or fold into the new UploadsController) exposing VirusScanService.getScanStatus, which currently has no endpoint at all.
• Add migrations and tests verifying both stores survive across service instantiations.

Acceptance Criteria:

• CDN invalidation requests remain retrievable via GET /cdn/invalidations/:id after an application restart.
• Virus scan results are queryable via a new endpoint and persist across restarts.

Technical Details:
Files: src/modules/uploads/cdn.service.ts, src/modules/uploads/virus-scan.service.ts, src/modules/uploads/cdn.controller.ts, prisma/schema.prisma
Libraries: prisma, @prisma/client

[BigGoodness]

I’ve reviewed the issue and I’m interested in working on it. I’ll first inspect the existing CDN invalidation and virus scanning flow, identify where the results are currently handled, and implement persistent storage following the project’s existing patterns. I’ll also add the necessary tests and make sure the changes are properly integrated without breaking the current flow. I can submit a PR with the implementation and tests within 3 hours.