Using CSVImport with AWS S3 without downloading file


#1

Is there a way to direct CSVImport to map to files by url without downloading the file?

We have our files intended for import stored in an AWS S3 bucket and have Omeka pointing at the same bucket for storage of Omeka objects. We would like to preserve the file structure and naming of the files while also bulk uploading into Omeka Classic. But CSVImport seems to not have an option to just map the objects to a file location, instead downloading the file and saving it under a random file name and storing it again in the S3 bucket. Thus, as I understand it, if I bulk upload all of our files in our S3 bucket, we will end up with copies of every file and lose the naming convention we created. Does anyone have a work around to make sure it maps the url and creates a thumbnail without downloading the object and creating a new url? Thanks!


#2

Creating the thumbnail requires downloading the file so we can operate on it.

As for the changing of the filenames, there are some options for preserving original names, though none built in to the core Omeka software. The main option that comes to mind is @Daniel_KM’s Archive Repertory plugin, though I’m not 100% certain on how it interacts with S3 storage.