Skip to content

Releases: mu-semtech/mu-search

v0.9.3

12 Jan 09:54
Compare
Choose a tag to compare

changes since v0.9.2

🐛 Fixes

  • fix: typo in regex for kebap-cased property names

v0.9.2

16 Nov 13:36
Compare
Choose a tag to compare

changes since v0.9.1

🐛 Fixes

  • fix: Support kebab-cased property names with modifiers: previously :gte:order-date resulted in ['gte', ['order']] while the expected result is ['gte', ['order-date']]

Changes

  • change: pin base docker image to semtech/mu-jruby-template:2.13.0

v0.9.1

15 Nov 15:37
Compare
Choose a tag to compare

What's Changed since 0.9.1

🐛 Fixes

  • fix: correctly handle deletion of a resource. due to a bug not all deletions in the DB also resulted in removal from the search index

v0.9.0

25 Sep 18:03
Compare
Choose a tag to compare

What's Changed since 0.8.0

🚀 Features

  • better support for composite types: now properly handling delta messages
  • experimental support for having multiple rdf types in one (non composite) index
  • basic highlighting support
  • search indexes defined in eager_indexing_groups can now partially match the user's allowed groups
  • experimental support for multilingual strings
  • improved handling of delta messages

🐛 Fixes

  • correctly respond to SIGTERM, so containers stops quickly
  • the admin endpoint to delete indexes should now work correctly
  • search should no longer be blocked when a new index is created

v0.8.0

26 Oct 12:57
Compare
Choose a tag to compare

What's Changed since 0.7.0

🚀 Features

  • Allow to specify wildcards and per-field boosts in fields parameter
  • Multi match support for fuzzy filter
  • Basic config validation on startup
  • Include URI of nested objects in the document
  • Extracting file content using external Tika service
  • Configurable log levels per scope
  • Improved error logging
  • Add documentation on eager indexing groups and update queues

🐛 Fixes

  • No longer throw an error in the delta handler when composite types are used (composite types are still not handled though)
  • Improved, but far from perfect composite type support
  • Using connection pool for all SPARQL queries
  • Error handling in case a file is not found for indexing
  • Allow all permutations of lt(e),gt(e) in search query params
  • Taking authorization headers into account on index management endpoints
  • Indexing of boolean values
  • Allow dashes and underscores in search property names
  • Ensure same index name independently of order of auth group keys

Contributors

Changelog: https://github.com/mu-semtech/mu-search/blob/master/CHANGELOG.md#v080
Commits: v0.7.0...v0.8.0

v0.6.1

12 Jun 07:32
Compare
Choose a tag to compare

Fixes since 0.6.0:

  • if a search is executed and the user does not have allowed groups yet, mu-search now fetches those first.
  • better error response if no matching indexes could be found (now returns a 400 bad request)