Skip to content

May 16, 2018

Jump to bottom Edit New page
Seth Shaw edited this page May 16, 2018 · 14 revisions

Time/Place

This meeting is a hybrid teleconference and IRC chat. Anyone is welcome to join. Here is the info:

Attendees

  • Danny Lamb
  • Seth Shaw
  • Rosie LeFaive 🌠
  • Bryan Brown
  • Jared Whiklo
  • Marcus Barnes
  • Paul Pound
  • Mark Jordan (regrets)
  • Andrija Sagic
  • Bryan Brown
  • Yamil Suarez
  • Natkeeran

Agenda

  1. CLAW Update
    1. Welcome our new committer: Seth Shaw
    2. Flysystem exploration
  2. Content Model Overhaul blocking block placement.
    • Standard block placement doesn't work but context placement does
    • We might be able to fix this by restructuring the seven conditions into two that act similar to the existing content types condition:
      1. IsEntityType (IsFile, IsMedia, IsNode, and IsTerm become check-boxes)
      2. HasTerm (MediaHasTerm, NodeHasTerm, and ParentNodeHasTerm become check-boxes OR a HasTerm condition with a parent option boolean checkbox that we use in conjunction with the IsEntityType condition)
    • Do we fix this now, or is it a separate issue/PR once the content model overhaul merges?
  3. ... (feel free to add agenda items)

Minutes

  1. CLAW Update:

    1. Congrats Seth Shaw, and thanks for all your work on content modelling and migrations!
    2. Has a rough Flysystem adapter using Chullo. But it has no way of knowing mimetype - it has to guess. Works with Core Flysystem, but not with Drupal in any way (yet) - need to finish the bridge (which will let you use fedora as stream wrappers in PHP) -- this will change a handfull of assumptions, not as bad as the current PR, but these will need to be sorted out. Lets you address files as fedora://name-of-your-file. --> this will solve the double-storage issue.
    3. We'll probably do a sprint when the release (7.x-1.11) ends.

  2. In the content model overhaul, we added conditions on block visibility, that hide normal drupal blocks. (did a screenshare using Zoom) -- a drupal block still needs to satisfy the new conditions to be displayed, unless you use Context to create a specific workaround. Options:

    1. Add a toggle on each condition to "enable this condition"
    2. Consolidate all these conditions into two conditions: first is a checkbox for what content type it is, and second is a condition for hasTerm?
    3. Do we fix this now, or make a new PR? Jared offered: merge now, fix later, constitution-style. Seconded by Seth Shaw.
    4. Discussing how this will work in the CLAW Playbook - there's a bit of work to do before slicing a version and testing claw playbook, after which it's "in".

DISAMBIGUATION PARTY BREAK:

  • 'context' is a third-party drupal module that does if-this-then-that logic.
  • plugins have a 'context' which is a cache of values - e.g. these block conditions have 'context' so they can evaluate themselves.

  1. Need for more documentation, especially regarding new stuff - documentation sprint/party?

  2. [Seth Shaw] Issue with Migrations not triggering derivatives - figured out what was going on - using Drush to do migrations, all context is done as anon - crayfish and Houdini see anonymous authentication, so while they were doing all the work, they couldn't post new files to Drupal. Easy solution : a switch that was discontinued in Drush 9! (oh no!) BUT: you can do post-action modifications with PHP code. Given a new switch that is provided, the user ID submitted, will use AccountSwitcher, and pass those credentials off to Houdini. In the Apollo proof of concept, you can see the code that adds the drush option. (essentially had to re-implement -u 1). The code: https://github.com/seth-shaw-unlv/claw-migrate-files-poc/blob/master/migrate_apollo/src/Commands/MigrateApolloCommands.php

Closing early, gonna push the PRs through!!

Discussed hidden dependencies: you NEED Something in your mapping that points to schema:dateModified(sp?) and memberOf and for tags, and for media type, to say "mediaOf" and "tags". In theory, you should be able to index any content typo you want so long as it has an RDF mapping. The tag stuff is for contexts within Drupal, but doesn't affect indexing.

You could have a context for a new content type to cause it to ingest, even without tags.

Another subtlety: code and configuration are well separated now. so long as you have the code, our configuration is basically an example (turnkey) to start with. How can we better wrap this, and provide messaging/phrasing around this? Documentation party!

[Natkeeran] The tags we have now are pretty good for the demo purposes, but should we solidify them? We probably don't need a whole vocabulary... but they should probably get looked at by metadata experts. Right now we just have one "tags" field, we probably don't want this - we should separate out the different uses of tags, like permissions, vs display, so you can change them.

[Jared Whiklo] The JSON-LD pull doesn't seem to do anything. Danny: it should render out URIs in the json ld. Whiklo: not seeing the PURL URI. in the drupal, i'm seeing the taxonomy terms. This is probably one of the weirder aspects... this is expected behaviour for now. it's just pumping out the taxo term ID. The taxonomy term itself has an owl:sameAs to an externalURI. So when you query your triplestore, it'll show hasType:[drupalURI] and by inference, hasType:[externalURI] Jared: seems the inferencing isn't working. When you export the JSON it makes sense to have the Drupal URI but when you export just JSON it makes sense to have the external URI.

You may be looking for the islandora-community wiki · new to islandora? · community calendar · interest groups · roadmap

This is an archive. For new Tech Call notes, click here

⚠️ ARCHIVED Islandora Tech Calls

⚠️ ARCHIVED Islandora User Calls

Clone this wiki locally