Right now, there is not an easy way to extract a specific subset of digital objects from AtoM. I've previously posted about the structure of the uploads directory in AtoM in the following thread, in case it is of use:
You might also find the following post of interest:
When you export a CSV of descriptions that include digital objects, AtoM will include a HTTP link to the digital object in the digitalObjectURI column of your export. If the source AtoM instance is publicly accessible (e.g. not behind any firewalls, descriptions are published, etc) and you don't need to store the master digital objects in the target AtoM instance, then you should just be able to import the CSV into your second AtoM instance. In that case, AtoM will treat the digital object links as remotes - it will use the URLs provided in the CSV to generate local derivatives (e.g. thumbnail for search/browse results and reference display copy for view pages) in your target AtoM instance, but instead of storing the master, it will save the source URL to point to it.
If you do need the original master digital objects to be stored in the target (e.g. second) AtoM instance as well, then you may have to do more manual work - something like this should work:
- Export your descriptions from the source system
- Copy the digital objects directory from the source system, and rename it to something like "import"
- Place the "import" digital object directory in the root AtoM directory of the target system
- Open your CSV
- Change the header on the digitalObjectURI column to digitalObjectPath
- Edit the values in the digitalObjectPath column to match those of your new "import" digital objects directory
For example, let's say you had a URL like this in exported CSV for a digital object:
To make this match our new path, we would want to remove the base URL (e.g. http://www.example.com/
), and since we renamed the directory from "uploads" to "import", in the path we will change it in our path as well - so your final digitalObjectPath value should look something like this:
Once you have done this for all rows in the CSV, you should be able to import the CSV to the target system, and then delete the whole "import" directory.
Info on the digital object columns in a CSV import:
Please note: I haven't tested out this particular workflow myself! You might have to experiment a bit. For example, it's possible that the digitalObjectPath column will require an absolute path - in that case, the final path you add might look more like this (assuming you have followed our recommended installation instructions and your root AtoM directory is located at /usr/share/nginx/atom):
Let us know how it goes!