Invalid namespace URI in Tufts dc_detail datastream

[mark-dce]

In finalizing mira v3.1.0, we've identified a few issues that will require a data migration to properly fix that is outside the scope of any of the DCE-Tufts SOWs to-date.

dcterms URI

In the original Tufts Digital Library code, there was a typo that neither Tufts nor DCE identified which references the incorrect URI for Dublin Core terms:
in https://github.com/mkorcy/tdl_hydra_head/blame/master/app/models/tufts_dca_meta.rb#L20

"xmlns:dcterms" => "http://purl.org/d/terms/"
should have been
"xmlns:dcterms" => "http://purl.org/dc/terms/"

This typo was carried forward into the MIRA codebase and any objects created by MIRA will have a namespace reference to '...d/terms' instead of the proper reference to '...dc/terms'.

Correcting the error in MIRA at this point will make any terms from dcterms unreadable from any previously created or edited records.

These two references need to be changed to correct the issue from a code standpoint:
https://github.com/curationexperts/tufts_models/blob/master/app/models/datastreams/tufts_dc_detailed.rb#L11
https://github.com/curationexperts/tufts_models/blob/master/app/models/datastreams/tufts_dc_detailed.rb#L84
However, when that change is made, existing content in the Tufts repository will need to be scanned for any data streams that still contain references to http://purl.org/d/terms/. Those references will need to be replaced with http://purl.org/dc/terms/ in order for the updated MIRA to read and write any related metadata terms correctly.

extent Namespace

Similar to the above, Tufts data model lists extent as part of the dc-elements namespace; however, it is actually defined in the terms vocabulary, not the elements vocabulary.

The initial issue can initially be seen here:
https://github.com/mkorcy/tdl_hydra_head/blob/master/app/models/tufts_dca_meta.rb#L35
and later here:
https://github.com/TuftsUniversity/tuftsification-hydra/blob/master/app/models/datastreams/tufts_dc_detailed.rb#L34

In both cases, the namespace should be in dcterms: http://dublincore.org/documents/dcmi-terms/#terms-extent, not dces: http://dublincore.org/documents/dces/ (searching for extent on this page yields no valid results)

For data compatibility MIRA perpetuates this issue.

To fix this issue, the following line in tufts_models would need to be changed:
https://github.com/curationexperts/tufts_models/blob/master/app/models/datastreams/tufts_dca_meta.rb#L29
Once that code change is made, however, any existing content in the Tufts repo will need to be updated in order to be read and managed appropriately by MIRA.

[mark-dce]

@mkorcy - did you know about these two issues? Are they something you want to take care of in this round? They would require both model updates and data migrations on your side.

[mkorcy]

@mark-dce I don't think I knew about this. I'll take a look and take care of this one.