Import Metadata CSV (Enhanced, with Preview of Changes, as in v6.x)

[heathergreerklein]

From release plan spreadsheet

Estimate from release plan: 14 hours
Expressing interest: none

Note this assumes the export CSV functionality is part of a different task. It also assumes the backend has separate independent endpoints for the preview and the actual import

• Create an Import metadata page
  • Reuse the dropzone from the submission to upload the file
  • Show the pending changes on the same page after the upload succeeds
  • Also reuse that page for the overview after the import is executed
• Create a service to communicate with the import csv endpoint. With methods to:
  • POST a CSV file to a preview endpoint, and get a response of what the changes would be
  • POST a CSV file to an import endpoint and get a response of what was executed.

Depends on REST ticket DSpace/DSpace#2849

[abollini]

Please check also the comment on the REST side DSpace/DSpace#2849 (comment)

In regards to the angular side we propose to raise the estimation to 24 hours and to exclude the use of a custom endpoint with the subsequent overview page of changes done to avoid the need to perform the actual import in real-time.

If a such option should be provided, with an upper limit that will force the switch to an async import the estimation should be raised by other 12 hours.

[tdonohue]

@abollini : Thanks for your feedback/analysis here.

I've verified your analysis (in DSpace/DSpace#2849 (comment)) that this feature already exists to some extent in the DSpace 7.1 & 7.2 user interface. For others watching this ticket, here's how to see a preview of Import Metadata changes in v7.2:

1. Login as an Administrative user (e.g. dspacedemo+admin@gmail.com on demo site)
2. Export Metadata from a Community or Collection (if you haven't already). Use the "Export" sidebar menu.
3. Modify the exported CSV in some way. For instance, just add an author or an abstract to an item.
4. Go to the "Processes" tab, click "+ New", and create a new "metadata-import" process
   • Select the modified CSV as the file
   • Add an additional parameter & select --validate-only.
5. Save/Start the process. The results will be that no changes are made in the system, but the Process Output/log will show the changes that would have been made for this CSV.
6. If they are correct, you can then go back and run "metadata-import" again, this time without the --validate-only flag.

So, in my opinion, I agree that process provides the basic "Preview of Changes" that we are looking for.

However, I see one minor usability issue / bug

• BUG: That same --validate-only flag CANNOT be triggered/selected if you instead use the "Import -> Metadata" menu and drag & drop your CSV onto that page.

So, my recommendation is to leave the backend as-is for now.

However, I'd like to see us add a checkbox to the "Import -> Metadata" page which says (something like) "Validate Only", with a hint/description under it that says "When selected, the CSV changes will not be executed. Instead, you will receive a report of which changes were detected in the CSV. When deselected, the CSV changes will be applied immediately."

This small change would allow for better usability of the existing feature & make it clearer to users how to first validate their CSV, then import the CSV.

Would your team be willing to add this simple checkbox to trigger the --validate-only flag? I think that's all that is necessary for 7.3.

[abollini]

perfect @tdonohue we will take care of the angular improvements.