v4.9.3

Fixed

• Update to graphANNIS 2.2.1, which fixes a bug where the left/right context was
  switched when using a segmentation as context.

v4.9.2

Fixed

• Documents which special characters (like %) in their name could not be opened
  in full text visualizes (like the document browser).
• The raw_text visualizer did not work properly and did not show the whole
  text.
• Avoid issues with possible problematic random number generation in visualizer
  result ID. Since the ID is only locally used for one match view, the impact of
  a flawed random generator are limited, but it is still bad practice to create
  a new random generator each time.
• Multiple texts in parallel corpora where treated as a single text, causing
  several problems. E.g. the KWIC view was showing a single token row with gaps
  instead of several token rows for parallel corpora.
• Fixed issue with qualified annotation names in HTML visualizer, which resulted
  in incomplete results.

v4.9.1

Fixed

• Addressed some non-security issues confusing static code analyzers (#782)
• Bumps commons-email from 1.3.2 to 1.5.
• Update Kotlin test dependency to 1.6 and also Karibu Testing
  (which needs the Kotlin dependency)
• Do not restart possible crashed graphANNIS service in integration tests.

v4.9.0

Added

• Allow to change the displayed order of corpus metadata annotations using the
  corpus_annotation_order configuration in the corpus-config.toml
  file. (#500)
• Automatically restart the bundled web service when it crashed.

Fixed

• Only show warning on MacOS with M1 processor instead if exiting.
• Use several lines for showing the MacOS M1 warning, because there is no
  automatic line brake.

v4.8.0

Fixed

• Upgrade to graphANNIS 2.1.0 to fix near operator which failed to work with
  segmentation constraint and also remove the db.lock file when ANNIS is closed
• Do not skip error messages when started service is aborted. (#761)

Added

• Because of the update to graphANNIS 2.1.0, it is also now possible to easily
  allow read-only access to all corpora without login and configuring an
  authentication provider using the graphANNIS anonymous_access_all_corpora
  configuration parameter.

v4.7.1

Fixed

• Upgrade to graphANNIS 2.0.6 to fix subgraph generation when a segmentation was
  defined as context and the match includes a token that is not covered by a
  segmentation node (there are gaps in the segmentation). Because token where
  missing from the graph, it could appear in ANNIS that there are gaps in the
  data and that the token order is incorrect.

v4.7.0

Added

• Show corpus metadata with corpus hits and not only the document metadata (#768)

v4.6.7

Fixed

• Upgraded to graphANNIS 2.0.5.
  Fix timeout handling for queries with a lot of intermediate results, but less
  than 1000 matches.

v4.6.6

Fixed

• Upgraded to graphANNIS 2.0.4. This fixes wrong number of results when using the non-existant
  operators with node attribute searches without a value (e.g. dipl !_=_ norm).
• Fix error message in corpus browser window when there is a meta annotation with an invalid name
  (like e.g. a space "corpus description" in the Bematac corpus).
• HTML visualizer might output the elements several times (#755).

v4.6.5

Fixed

• Mitigate against security issue CVE-2022-22965 by upgrading Spring boot
  dependency to 2.5.12 See https://spring.io/blog/2022/03/31/spring-framework-rce-early-announcement for more information.