bibo:peerReviewed doesn't exist, look again at BIBO ontology

[wickr]

http://purl.org/ontology/bibo/peerReviewed which is mapped in default metadata doesn't actually exist.

There is a http://purl.org/ontology/bibo/status/peerReviewed and it's counterpart http://purl.org/ontology/bibo/status/nonPeerReviewed but they're term values for I think the bibo:DocumentStatus property.

The ontology is here: https://github.com/structureddynamics/Bibliographic-Ontology-BIBO/blob/master/bibo.owl

[wickr]

@vantuyls @simholt

Option A: Change predicate to http://dbpedia.org/ontology/isPeerReviewed in DefaultMetadata and all items with this predicate The values for that are of range xsd:boolean (one of 'true', 'false', 1, 0) so I think we're ok there, even though our TRUE/FALSE are capitalized

Option B: Change predicate to http://purl.org/ontology/bibo/status in DefaultMetadata and all items with this predicate Change PeerReviewed property values (TRUE/FALSE) to

• http://purl.org/ontology/bibo/status/peerReviewed
• http://purl.org/ontology/bibo/status/nonPeerReviewed

This would allow other Statuses defined by BIBO ('draft', 'forthcoming', 'legal', 'published', 'rejected, 'unpublished' have URIs) and any others we would want to define. The existing Peer Reviewed facet could be configured to stay as it is now, or it could be merged in to a bigger 'Document Status' one that has Peer Reviewed and Non Peer Reviewed along with the other values.

[wickr]

Steve said go with Option A.

[vantuyls]

@wickr please also update the MAP when you finalize this.

[vantuyls]

@simholt @wickr if this is properly reflected on the MAP, please close this issue.

[revgum]

The current thinking is that I run a job to change the predicate in Fedora in conjunction with a code update so the application knows about the property/predicate change, and our “QA” service is aware of the metadata value. The new values should be lowercase true and false.

The solution to this ticket will need to be coordinated and planned to run the Fedora update immediately after the code is pushed to production.

[carakey]

We are changing the predicate to http://dbpedia.org/ontology/isPeerReviewed (as in Option A, with agreement as documented in this ticket in Jan 2018). The new predicate should have xsd:boolean conformant values true or false.

Aspects of this change include:

• [ ] Change predicate for existing items from http://purl.org/ontology/bibo/peerReviewed to http://dbpedia.org/ontology/isPeerReviewed
• [ ] Update code with new predicate
• [ ] Change values for existing items from TRUE/FALSE to true/false
• [ ] Update controlled vocabulary to use lowercase values: https://github.com/osulp/Scholars-Archive/blob/master/config/authorities/peerreviewed.yml
• [ ] Update MAP

[revgum]

Production update process;

• [ ] Tunnel to the SOLR server and query all of the works (~44k) with peerreviewed set : http://localhost:8983/solr/hydra-prod/select?fl=id,%20peerreviewed_tesim&indent=on&q=peerreviewed_sim:*&rows=60000&wt=csv
• [ ] Rake task to iterate through each work; deleting the statement with the old predicate from the graph, setting the peerreviewed to lowercase value, save the work (updates Fedora and index)

[carakey]

QA TL;DR: QA Pass for new instances of the Peer Reviewed statement / field. However QA fail for updates to existing Peer Reviewed statements.

---

QA in Staging based on criteria in 4/30/19 comment above:

• Criteria: Change predicate for existing items from http://purl.org/ontology/bibo/peerReviewed to http://dbpedia.org/ontology/isPeerReviewed

  • Found an existing item v692t643n with Peer Reviewed = Yes
  • Downloaded the NT and examined with Sublime Text. The file has the statement: <https://ir-staging.library.oregonstate.edu/concern/defaults/v692t643n> <http://purl.org/ontology/bibo/peerReviewed> "TRUE" .
  • QA fail. The existing item still has the old bibo predicate.

• Criteria: Change values for existing items from TRUE/FALSE to true/false

  • QA fail. See above; the existing item still has the old allcaps "TRUE" value.

• Criteria: Update code with new predicate

  • Found an existing test work of mine with no Peer Reviewed value pk02c995f
  • Edited metadata to add Peer Reviewed = Yes & saved changes
  • Downloaded the NT and examined with Sublime Text. The file has the statement: <https://ir-staging.library.oregonstate.edu/concern/articles/pk02c995f> <http://dbpedia.org/ontology/isPeerReviewed> "true" .
  • QA pass A new entry in the Peer Reviewed field uses the new dbpedia predicate.

• Criteria: Update controlled vocabulary to use lowercase values: https://github.com/osulp/Scholars-Archive/blob/master/config/authorities/peerreviewed.yml

  • QA pass See above; a new entry in the Peer Reviewed field uses the lowercase "true" value.

• ~Update MAP~ (for @carakey to do when the others are done)

[lamtu1]

Hi @carakey, have you run the rake task that is include in this PR, :update_peer_review_data? This file will update it all the old peerreview to the new predicate and change the value as well.

[carakey]

Hi @lamtu1, running rake tasks is not a thing I do. I don't have back end access and I wouldn't know how to run one if I did.

[lamtu1]

Hi @carakey, I will ask either Brandon and Corey about running on staging to test and see if we could do it.

[carakey]

QA Round 2 -- fail. A statement with the new dbpedia predicate was added, but the old bibo statement was not removed.

---

QA detail:

• Found another existing item cr56n115b with Peer Reviewed = Yes
• Downloaded the NT (here) and examined with Sublime Text. The file has both the old and new statements:

(line 26) <https://ir-staging.library.oregonstate.edu/concern/articles/cr56n115b> <http://purl.org/ontology/bibo/peerReviewed> "TRUE" .
(line 41) 
<https://ir-staging.library.oregonstate.edu/concern/articles/cr56n115b> <http://dbpedia.org/ontology/isPeerReviewed> "true" .

Staging has 3 total files with pre-existing Peer Reviewed values:

• cr56n115b
• v692t643n
• vt150j424

...Plus one that I added in QA round 1:

• pk02c995f

Looking further into this, I checked the NT files for all 4 in this round. All 3 pre-existing ones now have both statements. The post-update one only has the dbpedia statement.

[carakey]

After @lamtu1 reran the script, all 4 staging works in the previous comment have only the dbpedia statement with lowercase "true" values.

QA pass.

BUT double check after the first run on Prod to make sure the old statements are removed.

[carakey]

I changed the predicate in the MAP.