Samvera Tech Call 2019-10-30

How to connect: https://psu.zoom.us/j/613720745 (link will launch Zoom client – if you do not have Zoom, expand the instructions below)

Time: 9:00am PDT / Noon EDT

Moderator: @Jim Coble

Notetaker: @James Griffin

Attendees:

•  

  • @Jim Coble (Duke U)

  • @Julie Allinson (Notch8)

  • @tamsin woo (UCSB)

  • @Chris Colvard (Deactivated) (Indiana University)

  • @Thomas Scherz(Cincinnati)

  • @Collin Brittle (Emory)

  • @Andrew Myers (WGBH)

  • @Trey Pendragon (Princeton)

  • Esty Thomas (U. Michigan)

  • @Gordon Leacock (Univ of Michigan)

  • @Jason Corum (WGBH)

Agenda

1. Roll call by timezone per following order - ensure notetaker is present (moderator)

   1. folks outside North and South America

   2. Eastern timezone

   3. Central timezone

   4. Mountain timezone

   5. Pacific timezone

   6. folks who were missed or who dialed in during roll call

   7. Remind everyone to sign in on agenda.

   8. Welcome all newcomers!

2. Agenda (moderator)

   1. Call for new agenda items (moderator)

   2. Turn off Hyrax Batch ingest by default? (@tamsin woo)

   3. Replace Hyrax::Callbacks with Dry::Events? (@tamsin woo)

      1. Straw PR for discussion: https://github.com/samvera/hyrax/pull/4132

      2. See especially: https://github.com/samvera/hyrax/pull/4132/files#diff-e37293a3703317fdbe7afc82c8354728R4-R45

   4. Core Components WG chair volunteers?

   5. Hyrax Speed Test (i.e. benchmarks we can share and compare) (@Andrew Myers)

   6. Samvera Docs WG needs love (i.e. ideas + cycles) (@Andrew Myers)

3. Moderator & notetaker for next time

   1. Moderator:

   2. Notetaker:

4. After call, this week's notetaker should create the agenda for the next call:

   1. Open template agenda titled "Samvera Tech Call 2019-xx-xx"

   2. Click on ... in the top right corner, and select copy.

   3. Popup will open for location. It should contain: 

      1. Space: Samvera

      2. Parent page: 2019

   4. Select copy. New page should be created.

   5. Modify the title to remove "copy of", update it with the next date, add moderator, notetaker, and any carry-over agenda info. Click Publish.

5. PR Review

   1. Review issues:

   2. PR review coordinator for next time: 

Notes

Hyrax Batch Ingest

There is a pull request (4129), which was merged yesterday; this defaults to disabling batch ingest features for Hyrax 3.0.0

Batch ingest has some conflicts if there is customized metadata

It's been on the roadmap to address improving aspects of this feature (or, perhaps, even replacing it)

As this is a feature flipper default, one can still enable it, it just must be done so explicitly

U. Cincinnati noted that they are satisfied with the feature

Hyrax::Callbacks and Dry::Events

There is a pull request (4142) for which might be in a reviewable state (the test failures may be unrelated)

Since 2015, Hyrax has had its own custom pub/sub callback system

Tom Johnson has been looking to shift this towards a more consistent approach

Currently, callbacks are blocks with a name, with one listener per event topic

The subscriber registry is mutable but not thread-safe (different processes can modify this and introduce conflicts)

Dry::Events provides an alternative implementation for this functionality, with the added benefit of building upon the labors of the dry-rb community

Existing callbacks will need to be deprecated from the registration side

Adopters will need to adjust the dry-events configuration

Does this relate to Valkyrie?

This is mostly orthogonal to Valkyrie, but it was discovered while Tom has been addressing Valkyrie-related work

How are events currently being used?

Events are exclusively called from the Actor stack

Callback names are like "after_object_created"

They are used to trigger jobs for background processing

WGBH supports this change in implementation

Drew Myers noted that Hyrax::Callbacks was not originally written with the pub/sub pattern in mind

To what extent is this a new paradigm?

Tom: Listeners are registered on an application-level basis, with callbacks being invoked explicitly

Wants to ensure that any concerns regarding the new paradigm is captured from community contributors

Chris Colvard: Nested indexing also may be registering methods with Hyrax::Callbacks (outside of Hyrax)

Chris will review this elsewhere and confirm that this is the case

Core Components Working Group Chair Volunteers

Call for volunteers to serve as the chair for the next phase of the Working Group

This WG can no longer be a viable Working Group without a chair

With a chair, this WG will run for a full year, with the hope that new charter will be drafted at the end

What is the deadline for a new chair?

Looking to draw the old group to a close in three weeks

Hyrax Speed Test

During Connect, attendees were sharing benchmarks/speed tests within production-like environments to debug infrastructures within each institution

This also provided a better expectation as to what to expect

At the end of these discussions, Drew volunteered to serve as a coder, but wanted to stop shy of forming a formal Working Group

Calling for volunteers with Hyrax installations available in staging environments (a vanilla installation with an empty repository is strongly preferred)

Drew will look to follow this call with an e-mail to the samvera-tech Google Group

Tom: There is another line of performance and observability-related work which is related

Rebuilding Nurax in a containerized environment, with the aim of integrating Honeycomb into this Docker build

It might be best to assess what others are using in their productions environments for analyzing performance troubles for Hyrax in production

Samvera Documentation Working Group

Status of the WG is that there has been little progress recently in working to progress with improving documentation

Drew is looking to call for a renewed effort and reviving the Working Group

The aim should also be to attempt to have this as a standing Working Group (as documentation maintenance should be ongoing)

Some of the proposed documentation improvements may require minimal maintenance

Should the existing Working Group be canceled, with a new WG formed?

It was noted that the charter for the current WG is long expired

Should this be an Interest Group which exists indefinitely, or remain a Working Group with deliverables and a sunset date?

Drew: Started with the preference for a Working Group, but this requires volunteers

Trey: Strong proponent of the requirements of Working Groups, if they are not active, they should be removed from the list of existing WGs

If there aren't the resources, then perhaps the WG should be run later

Recommends rechartering, and accepting that there simply might not be the resources for renewal

Tom: Chartering process does not have to be heavy-weight, but it does have the benefit of forcing a reconsideration with each new renewal

Drew will issue a CfP for the new Working Group

Next Scheduled Call:

Moderator: Drew Myers

Notetaker: Jason Corum

Meeting adjourned at 09:32 PDT/12:32 EDT

Hyrax Pull Request Review

Attendees:

Chris Colvard

Gordon Leacock

Jason Corum

Pull Requests:

3881

Updated with some changes

There is a long history

3947

Testing with Rails 6

Chris does not have the availability to continue working on this

Will close this for now

4067

Date issues in the Safari browser

Front-end issue

Embargo release date formats need to be parsed and formatted consistently using JavaScript

Should we be moving away from `jQuery` for Hyrax JavaScript?

Should we using `var` as opposed to `let` in .es6 files?

Proposed changes affect all browsers as opposed to targeting just Safari (this might not be an issue)

James left a comment regarding the `var`/`let` usage

4115

This was approved, but there are still legitimately failing tests

There is still work to be done to complete this pull request

4045

Methods for attaching file_set to a Work

Lynette needs to still return to this

4046

New query for the metadata adapter

There is still work going on for this

4049

There were changes requested

4055

Still has failing tests

4056

This just has one build failing

Links to 4057

There are in-depth conversations

Best to leave Tom and Lynette finalize the work on this

8743

Has legitimately failing tests

4024

There are two failing tests for the changes introduced here

Blacklight::SearchState might be a service class

Perhaps the tests themselves need to be updated?

Chris provided a comment addressing this

4074

Questions regarding #extract_subjects, but otherwise looks like a good improvement

Chris will review this further

Session concluded at 13:02 EDT