Versions Compared

Old Version 1

changes.mady.by.user Brian Lowe

Saved on

compared with

New Version Current

changes.mady.by.user Georgy Litvinov

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

Date

23 Feb

Call-in Information

Time: 11:00 am, Eastern Time (New York, GMT-04:00)

...

  1. Brian Lowe
  2. Benjamin Gross 
  3. Nicolas Dickner 
  4. Georgy Litvinov (star)
  5. Benjamin Kampe 
  6. Michel Héon 
  7. Huda Khan 
  8. Alexander (Sacha) Jerabek
  9. Ralph O'Flinn  

Agenda

...

  1. Alpha release
    1. Blockers
      1. Jira
        serverLYRASIS JIRA
        serverIdc815ca92-fd23-34c2-8fe3-956808caf8c5
        keyVIVO-1956
      2. Jira
        serverLYRASIS JIRA
        serverIdc815ca92-fd23-34c2-8fe3-956808caf8c5
        keyVIVO-1966
      3. Jira
        serverLYRASIS JIRA
        serverIdc815ca92-fd23-34c2-8fe3-956808caf8c5
        keyVIVO-1967
      4. Jira
        serverLYRASIS JIRA
        serverIdc815ca92-fd23-34c2-8fe3-956808caf8c5
        keyVIVO-1969
    2. Deployment and developer workflow
      1. Not requiring home directory tar file
        Jira
        serverLYRASIS JIRA
        serverIdc815ca92-fd23-34c2-8fe3-956808caf8c5
        keyVIVO-1968
      2. Offer ability to copy war to Tomcat during installation?
          1. https://github.com/vivo-project/VIVO/pull/224
      3. Maven lifecycle semantics : install versus package?
    3. Jira
      serverLYRASIS JIRA
      serverIdc815ca92-fd23-34c2-8fe3-956808caf8c5
      keyVIVO-1965
      License text: should other changes be considered?
    4. Upgrades from monolingual to multilingual VIVO with ingested data
      1. Are we satisfied with providing documentation vs. tooling?
    5. Nonblockers and documentation tasks
      1. Jira
        serverLYRASIS JIRA
        serverIdc815ca92-fd23-34c2-8fe3-956808caf8c5
        keyVIVO-1762
      2. Jira
        serverLYRASIS JIRA
        serverIdc815ca92-fd23-34c2-8fe3-956808caf8c5
        keyVIVO-1914
  2. New developments
    1. Jira
      serverLYRASIS JIRA
      serverIdc815ca92-fd23-34c2-8fe3-956808caf8c5
      keyVIVO-1962
      Configurable SPARQL queries for comprehensive individual deletion operations
      1. Pull request for testing:
        1. https://github.com/vivo-project/Vitro/pull/213
    2. Deploy VIVO with Solr in one "docker-compose up"
    3. Browser plugins that are able to double-submit editing forms
  3. Post-i18n priorities
    1. VIVO-in-a-box
    2. Ingest / Kafka
    3. Advanced Role Management
    4. Moving Scholars closer to core - next steps

...

  1. Status of In-Review tickets

    Expand

    Jira
    serverDuraSpace JIRA
    jqlQueryfilter=14416
    serverIdc815ca92-fd23-34c2-8fe3-956808caf8c5


Notes 

Link on DuraSpace Wiki:

https://wiki.lyrasis.org/display/VIVO/2021-02-23+-+VIVO+Development+IG

Attendees

  1. Brian Lowe
  2. Benjamin Gross
  3. Nicolas Dickner
  4. Georgy Litvinov
  5. Benjamin Kampe
  6. Michel Héon
  7. Huda Khan
  8. Ralph O’Flinn
  9. Don Elsborg
  10. Alexander J. Jerabek
  11. William Welling

Notes

  1. 1.12 Alpha release blockers
    1. VIVO-1956 needs some reviews. Review please. Should be fairly easy thing to check. Try some queries. Ralph and Don Elsborg are ready to test. 
    2. VIVO-1966 Subject of the statement is not being reindexed. If you create a Faculty member and add sameAs property to second Faculty member. First one won’t be reindexed. sameAs doesn’t have inverse property. Default behaviour should be to reindex both object and subject. Georgy will prepare a pull request for this issue. 
    3. VIVO-1967 Errors in logs appeared while opening the Person page. Binding method wasn’t implemented. Brian implemented it to make it work properly.
    4.  VIVO-1969 - Removing RDF with uploadRDFForm does not lead to reindex dependent uris that should be found by IndexingUriFinder. The problem is that by the time indexFinders start working, triples they are looking for are already gone. After removing some data it could still exist in search indexes. Can lead some embarrassing situations but only relates to search indexes. It is a pre-existed problem. Maybe add some warning in documentation. For now this problem could create a delay to release 1.12. So maybe fix it later. Needs some more investigation.
  2. Deployment
  3. Slack questions
    4. Releasing 1.12 alpha
    • Michel: will there be a tag?
    • Ralph: there is an alpha tag (rel-1.12.0-alpha-1) ; typically 2 weeks after RC; then, it will be merged in main as rel-1.12.0. Usually won’t create an artifact for the alpha
    • Brian: could people be more inclined to test if there was a .war file to download?
    • Benjamin: best to provide options
    • Ralph: previously, source code + artifacts for the installer, so will keep it that way.
    • VIVO-1956
      • Brian : two PR to test (to make initial RDF files more consistent) ; test should be easy, no functional changes ; queries are provided to help testing
      • Added readme file
  4. Brian: William Welling has added an option to build where you can specify if you have a Catalina home variable pointing to your directory, it will copy the .war to Tomcat : https://github.com/vivo-project/VIVO/pull/224
  5. William: it’s a basic Maven profile that copies the war file for the user.
  6. Brian: not tested yet. Michel has issues.
  7. Michel: UQAM problem is that deployment of VIVO home is now done at the moment of deploying the war file ; which creates a dependency between the code and the installer of VIVO ; so it makes the development of an independent installer impossible ; it also makes impossible to configure runtime.properties before starting Tomcat.
  8. William: it was the case before, but the process now allows to configure runtime.properties, which would allow an overlay install strategy such as UQAM. However, there is a caveat: you’ll need to create a digest file for the home directory ; this needs to be documented. 
    1. Michel tested VIVO-1968. We need to save layers of customizations. Simplest way to retain custom installation is to allow changes to core VIVO war-based installers. Good solution could be to have some switch to skip new install process (changing home directory) entirely. Could be a piece of the puzzle. William also reviewed it. 
    2. Still have an issue to solve for UQUAM. UQUAM smoketests are showing problems with new context.xml In addition to packing tarball there have been some changes to context.xml It put one property in context.xml to bootstrap home directory. New example of context.xml
    1. https://github.com/vivo-project/VIVO/blob/main/
    start.sh
  9. https://github.com/vivo-community/vivo-custom-webapp
  10. https://github.com/vivo-project/VIVO/blob/main/start.sh#L19
    1. installer/webapp/src/main/webResources/META-INF/context.xml . The name of home directory property has changed. Property vivo-home renamed to vitro-home. New installation doesn’t know about old properties. Possible resolution - take old context.xml and test with it. Michel is going to test it with updated context.xml. Error messages strangely has links to Tomcat9. Should new release documentation require Tomcat 9 or still 8.5 enough? Using Tomcat 9 is just a recommendation. No hard requirements for Tomcat 9 for now.
  11. VIVO-1962 Georgy added some fixes to his initial pull request.
  12. Don (in chat) Moving forward I would have a digest file checked into source code control and embedded with the rdf directory in our 3rd tier
  13. Michel: this new procedure has to be documented as soon as possible, as UQAM and UQAM partners current install processes are blocked.
  14. William: it is possible to keep on using UQAM installer, with some configuration.
  15. Michel is unsure whether it would work with overlays ; thinks it’s bound to cause problem with i18n installs
  16. Ralph: might not be as bad as it looks ; we could put up documentation for the alpha release, so we can assess it.
  17. Michel: there could be a quick fix. When Tomcat starts, if VIVO home is already there, it shouldn’t be erased ; it could be in runtime.properties or developper.properties
  18. William: those files are in the VIVO directory, so it’s a chicken and egg problem.
  19. Michel: that’s bad  practice; config files should be there before starting the application.
  20. Brian: the goal was to make it easier for people who have a simple install process. Michel is right that it seems backward for more complex install processes. However, it looks like the digest file William was talking about would play the role of that config file. So the digest file could be used as a switch for a custom installer.
  21. Michel: it’s too early to deploy this in the release ; more reviewing has to be done on that feature.
  22. Brian: is it easy to revert?
  23. Ralph : it has been checked in, there’s code on top of it, would be hard to revert it. There’s 4 weeks of testing and tweaking left before the release. It’s not that much code, it’s still possible to keep both approaches.
  24. Michel: do we have to just touch the digest file?
  25. William: it’s the checksums of the content of the file ; you just have to write it the first time.
  26. Brian, let’s schedule a meeting to discuss this  : Thursday 10 EST.
    • Release process
    • a) blockers
    • b) Deployment and developer workflow
  27. New developments
    • Brian: long needed feature ; PR in review. General idea is you can create a query with specific arguments (MostSpecified type, etc) and it will use DESCRIBE to spot and delete the corresponding triples.
    • Georgy: it’s working, even though some cleaning might be needed
    • Brian: options could be used in order to generate more complex queries
    28. a) VIVO-1962

Draft notes on Google Drive