Some Shrine plugins do not seem to run with shrine-tus about shrine-tus

[janko]

When you do any kind of direct uploads (custom endpoint, S3, tus), Shrine won't extract the metadata automatically, and will instead just use metadata that was set on the client side. This is because, in order to extract metadata, Shrine needs to retrieve the file contents (though determine_mime_type and store_dimensions plugins usually need just first few kilobytes). And that can potentially be expensive if the file is uploaded to remote storage like S3.

You can enable metadata extraction on attachment assignment by loading the restore_cached_data plugin. This is useful if you want to validate the extracted metadata, but it means metadata extraction will happen in the foreground.

Shrine.plugin :restore_cached_data

Alternatively you can use the refresh_metadata plugin (which restore_cached_data uses internally) to trigger the extraction yourself. Then you could e.g. call it in a background job if you're using it, by putting it in the process step when cached file is uploaded to permanent storage:

Shrine.plugin :refresh_metadata

process(:store) do |io, context|
  io.tap(&:refresh_metadata!) # return original cached file with extracted metadata
end

[codezomb]

Thanks for the quick reply! Does this also apply when using shrine-this with option C? The upload location and store location are both mounted on the host and accessible.

…

On Sep 29, 2018, 6:48 PM -0700, Janko Marohnić ***@***.***>, wrote: When you do any kind of direct uploads (custom endpoint, S3, tus), Shrine won't extract the metadata automatically, and will instead just use metadata that was set on the client side. This is because, in order to extract metadata, Shrine needs to retrieve the file contents (though determine_mime_type and store_dimensions plugins usually need just first few kilobytes). And that can potentially be expensive if the file is uploaded to remote storage like S3. You can enable metadata extraction on attachment assignment by loading the restore_cached_data plugin. This is useful if you want to validate the extracted metadata, but it means metadata extraction will happen in the foreground. Shrine.plugin :restore_cached_data Alternatively you can use the refresh_metadata plugin (which restore_cached_data uses internally) to trigger the extraction yourself. Then you could e.g. call it in a background job if you're using it, by putting it in the process step when cached file is uploaded to permanent storage: Shrine.plugin :refresh_metadata process(:store) do |io, context| io.tap(&:refresh_metadata!) # return original cached file with extracted metadata end — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

[codezomb]

Worked like a charm, thanks!

[janko]

Yes, it's also the case with option C, as Shrine doesn't differentiate between storages, as it avoids storage-specific logic in order for the behaviour to stay the same when switching between different storages.

I will add the note about extracting metadata to the README.