OAR/DSpace: Incorrect counts for Collections

[hhkitsune]

Description:

The Departmental Collections page is showing incorrect counts for some of the collections. For example, Art & Archaeology is showing "0" but there are "26" items when you click the collection link and go to that collection's page.

Here is a list of the collections that need to be fixed:

• African American Studies (shows 1, but has 29)
• Art & Archaeology (shows 0, but has 26)
• Astrophysical Sciences (shows 466, but has 779)
• Chemical & Biological Engineering (shows 130, but has 287)
• Chemistry (shows 48, but has 443)
• Civil And Environmental Engineering (shows 93, but has 97)
• Comparative Literature (shows 0, but has 19)
• Computer Science (shows 185, but has 1069)
• East Asian Studies (shows 21, but has 22)
• Ecology and Evolutionary Biology (shows 393, but has 420)
• Electrical Engineering (shows 87, but has 320)
• English (shows 0, but has 21)
• Faculty Publications (School of Public and International Affairs) (shows 0, but has 619)
• Geosciences (Shows 0, but has 476)
• German (Shows 0, but has 1)
• History (shows 0, but has 64)
• Journal of Public and International Affairs (shows 0, but has 1)
• Library (shows 0, but has 3)
• Liechtenstein Institute on Self-Determination (shows 0, but has 58)
• Mathematics (shows 296, but has 297)
• Mechanical and Aerospace Engineering (shows 76, but has 289)
• Molecular Biology (shows 318, but has 564)
• Music (shows 0, but has 3)
• Near Eastern Studies (shows 0, but has 23)
• Operations Research and Financial Engineering (shows 144, but has 312)
• Philosophy (shows 1, but has 20)
• Physics (shows 1102, but has 1339)
• Politics (shows 64, but has 69)
• Program in Gender and Sexuality Studies (shows 0, but has 2)
• Psychology (shows 250, but has 251)
• Religion (shows 0, but has 36)
• School of Public and International Affairs (shows 620, but has 682)
• Sociology (shows 33, but has 35)
• University Center for Human Values (shows 0, but has 11)

Note: These are the count differences I found using the UI. However, the numbers should be checked internally to be certain all counts are made correct and no assets are missing from the collection display.

Acceptance Criteria:

• [x] Counts are checked and verified, there aren't missing items from any collection
• [x] The counts will be corrected to show the correct amount for each collection

[hhkitsune]

@jrgriffiniii FYI, this is the problem with OAR/DSpace I had mentioned in the meeting.

[jrgriffiniii]

This should also confirm that the results on the Collection show page are valid (the number on the page, as well as confirming that each member item is listed as they should be).

[jrgriffiniii]

The collection members are rendered using database queries within interfaces such as https://oar.princeton.edu/handle/88435/pr1mw2m (hence, these should be accurate), while the item count for the https://oar.princeton.edu/collection-list interface is populated using other mechanisms.

[jrgriffiniii]

https://github.com/PrincetonUniversityLibrary/princeton-5/tree/oar-TEST-fix-collection-count resolves most of this issue, however, there is still one final bug affecting the collection member item total within the Collection show page.

[jrgriffiniii]

I have been able to isolate https://github.com/PrincetonUniversityLibrary/princeton-5/commit/28d4c7dc4dc8584b3d56f506863e0573501b31fe#diff-b31192a1c4cab83405c6a20f9c6704c3dabe1a627ecb7efc67638497b51808bd as the area of the source code base where the query is failing (some time was consumed attempting to trace this through the logging). I am looking to try and implement a remedy for this now.

[jrgriffiniii]

While trying to resolve this as a bug, it appears that entirely rebuilding the Solr Index using ($ /dspace/bin/dspace index-discovery -b) is also going to be required.

[jrgriffiniii]

This is indeed what is going to be required for fully resolving this issue. I shall finalize this by opening pull requests for the source code changes.

[jrgriffiniii]

The reindexing is being finished for the deployments before this is shared for review with the stakeholders.

[jrgriffiniii]

For the production, staging, and development environments, the reindexing is now finished.