Rails Security update

[2.3.7] - 2022-07-13

Removed

• link to ERA A+V #2765

Chores

• remove Ruby 2.6 PR#2878
• bump rubocop and fix new nags PR#2900
• bump ruboocp-rails and fix new nags PR#2899

Security

• bump rails 6.1.6 to 6.1.6.1

Rails Security update

[2.3.6] - 2022-04-28

Security

• bump rails 6.1.5 to 6.1.5.1

Chores

• bump omniauth-saml 2.0.0 to 2.1.0 PR#2767

File Uploads

[2.3.5] - 2022-04-06

• Jupiter II work is continuing to incorporate Digitized materials into Jupiter in the digitalcollections namespace.
  • batch ingest reports PR#2612

Fixed

• File Upload error PR#2798

Sidekiq Upgrade

[2.3.4] - 2022-02-08

Added

• Add readiness healthchecks for Rails and Sidekiq PR#2657

Security

• Bump Sidekiq from 5.2.9 to 6.4.1 #2189
• Bump follow-redirects
• Bump puma
• Bump actionpack

Bug Fix for Thesis DOI

[2.3.3] - 2022-01-19

Fixed

• Bring back illogical date range faceting flash message #2030
• Thesis not being assigned a DOI #2707
• Render markdown when viewing Collections as an administrator #2708

Bug Fix for Communities

[2.3.2] - 2022-01-10

• Jupiter II work is continuing to incorporate Digitized materials into Jupiter in the digitalcollections namespace.
  • newspaper metadata for ACN digitization #2645

Fixed

• nil Class error when viewing Collections drop down on Communities page #2655
• Render markdown when viewing Communities as an administrator #1322

Chores

• Bump rubocop-rails to 2.13.0 and fix cop violations PR#2683
• Bump rdf-vocab to 3.2.0 PR#2696

Bug Fix for StrongMigrations

[2.3.1] - 2021-12-07

• Fix Gemfile so that strong_migrations is used in all environments

Datacite API transition, Rails 6.1 and File size limit

[2.3.0] - 2021-12-01

• The EZID Compatibility API is sunsetting at the end of this year, per https://blog.datacite.org/sunsetting-of-the-ez-api/. datacite-client is a ruby gem that wraps the Datacite API for our use. The main changes are the DOI's no longer have the doi: prefix, the format of metadata attributes, and the event mechanism for publishing/hiding the metadata from the public. Requires datacite_api feature flag and new secrets for our datacite credentials. #2268

• We had a request by a researcher to attach several 2.8 Gb zip files to an existing object. We couldn't fulfill this request because of the way we were storing metadata about the file. Using Integer put an artificial limitation of 2,147,483,647 (2^31-1) on the size of files we could attach. We migrate the byte_size of blobs to use BigInt, 9,223,372,036,854,775,807 (2^63-1), instead.

Chores

• bump sidekiq-unique-jobs from 7.0.12 to 7.1.8 and fix a long missed deprecation
• fixed deprecation warning on tests #2604
• revise uat deploy configuration and watchtower script #1985
• fixes: Docker demo Redis bad URI error #2610
• add strong_migrations to catch unsafe migrations in development #2621
• Upgrade Rails to version 6.1 #2079

Fix Authentication Failure

[2.1.1] - 2021-11-24

Fixed

• bump ruby-saml to resolve compatability issue with rexml #2636

Subdomains, Feature Flags and GoogleDrive Ingest Workflow

[2.2.0] - 2021-10-21

• Refactored Controlled Vocabulary support to allow for new, raw vocabs without i18n translations. The motivation here is that we have a bunch of URIs we want to machine-map to human readable values, and it doesn't make sense to introduce intermediate symbols we'd have to cobble together somehow, plus that would involve polluting the i18n file with hundreds of new entries.

API Examples:

      ControlledVocabulary.value_from_uri(namespace: :digitization, vocab: :subject, uri: "http://id.loc.gov/authorities/names/n79007225")
       => ["Edmonton (Alta.)", false]
      ControlledVocabulary.uri_from_value(namespace: :digitization, vocab: :subject, value: "Edmonton (Alta.)")
       => "http://id.loc.gov/authorities/names/n79007225"

      uri = "http://id.loc.gov/authorities/names/n79007225"
      ControlledVocabulary.digitization.subject.from_uri(uri)
       => "Edmonton (Alta.)"
      ControlledVocabulary.digitization.subject.from_value("Edmonton (Alta.)")
       => "http://id.loc.gov/authorities/names/n79007225"
      ControlledVocabulary.era.language.english
       => "http://id.loc.gov/vocabulary/iso639-2/eng"

Further discussion of the context can be found at #2119

• Many "description" or "abstract" fields (at the Item level as well as Communities and Collections) contain HTML tags. Because these are text fields, HTML is not rendered in the UI and text looks garbled and it's way less readable than ideal. Markdown should work really well for this since that's already used in many of the tools staff working in repositories are familiar with. Added redcarpet gem which renders markdown in our decorators and strips markdown in our Solr exporters #1322

• Added feature flags to Jupiter. The motivation for this change is so that we can continuously deploy and turn on or off features as needed. Admins can enable features through the admin panel. #1897

• Jupiter II work is underway to incorporate Digitized materials into Jupiter in the digitalcollections namespace. We've begun by modelling, developing the user interface, and tasks for ingest of FolkFest Programs.

  • Peel redirects #1769
  • Make Digitization::Book Depositable
  • Make Digitization::Newspaper Depositable
  • Make Digitization::Image Depositable
  • Make Digitization::Map Depositable
  • Volume and Issue label attribute to Digitization::Book
  • Corrected missing pluralization in Digitization::Book attributes
  • Make Digitization::Book more like other items and other small fixes
  • book metadata for folk fest digitization #2010
  • Add Digitization::Book ingest artifacts to model #2011
  • Add task that will kick off job for batch ingestion of digitization metadata from a csv containing triples #2011

• Improve batch ingest workflow by using Google Drive for staging and a user interface for creating and reviewing batches. Requires new secrets to be configured and batch_ingest feature flag. Further context can be found #1986

  • Add new models (BatchIngest and BatchIngestFile) for improved batch ingest work
  • Add new google drive client service to be able to retrieve files/spreadsheets from Google Drive
  • Add batch ingest controller and views for CRUDing batch ingests
  • Add batch ingest form with google file picker and spreadsheet validation
  • Add batch ingestion job for batch ingesting items into ERA
  • Add various fixes and improvements to batch ingestion work

• Subdomains being used for front doors to the application. era and digitalcollections are the ones in use so far. This requires that the host subdomain match these exactly, and new secrets to be configured. Further context can be found #1707

  • Add 'era' subdomain and foundation for future frontdoors #1786
  • Add 'digitalcollections' subdomain for future front door
  • Refactored item download/view behaviour in routes and views to be reusable in digitization namespace and application wide

• Added highlighting of terms within search results descriptions. Requires fulltext_search feature flag. #1800

• Added category labels for active facet badges. Requires facet_badge_category_name feature flag. #1261

• Changes default behaviour within a facet to 'OR'. Requires or_facets feature flag. #1990

Added

• Namespaces for the Controlled Vocabularies #2118
• Brakeman linting to Github Actions workflow
• Make autocomplete explicit PR#2449
• RdfAnnotation changes will be output to a file to facilitate testing and database setup acts_as_rdfable#12

Removed

• Remove logo_id foreign key on item/thesis which was causing issues with deletions
• Removed rufus-scheduler -- it's an unused dependency PR#2434

Changed

• Moved visibility vocabulary into a jupiter_core namespace
• Move doi_url to Doiable class
• UAT VIRTUAL_HOSTS configuration on just the containers that need it

Fixed

• Fix error when parsing n3 files which include objects with elements as values.
• Bump flipper-ui, flipper-active_record and flipper and remove redundant configuration
• Bump rubocop to 1.15.0 and Style/TrivialAccessors default changed PR#2343
• Bump rubocop to 1.18.2 and fix cop violations PR#2415
• Bump rubocop-minitest to 0.14.0 and note really smelly tests PR#2416
• Fixed a flaky test where the page hasn't finished loading #2129
• Refactored one of our smelliest tests to use fixtures and reduce number of assertions per test #2419