Skip to main content

📁 Upload

How does it work?​

It uses the endpoints of the backend to upload the files in chunks. The steps are:

  • Only show button if upload is enabled in admin area
  • Even then, the button and the dropzone stay disabled (greyed out, with a "Scan directory not configured - contact administrator" tooltip) if the logged in user has no scan directory. The backend enforces the same rule and answers complete with a 400 when the user's scan directory is unset or does not exist on disk
  • Check if file is on the server by using the exists endpoint.
  • If it is, don't upload it
  • If it isn't, upload it with the upload endpoint
  • If the upload is complete, call the complete endpoint

The frontend never triggers a scan itself. While the backend handles complete, it queues a django-q Chain for that one file: handle_new_image, caption generation, geolocation, adding the location to the album dates and face extraction. Duplicates are skipped — the endpoint answers with "Photo duplicated. No new import performed." and no chain is queued. Once the last file is done, the client only invalidates its cached queries (recently added photos, date albums, count stats, photo month count and storage stats).

note

GET /api/scanuploadedphotos still exists on the backend, but it is only an alias of the same view as /api/fullscanphotos and nothing in the web frontend calls it.

This is implemented in the ChunkedUploadButton component in the frontend. The progress is saved within the state of the component.