v4.12.3

Fixed

• Grid visualizer now also display spans that do not belong to any node layer
  when the "layer" is not specified in the resolver configuration. Before this
  fix, the layer name was not checked, but the node still needed to be part of a
  layer.
• Updated to graphANNIS 3.5.1 which fixes an issue where a backup folder in the
  database might not be loaded correctly and an issue where impossible precedence
  queries created a "File too large" error.
• Allow to show text visualizations for documents where the token are not
  connected to the textual data source.
• Resolver entries with the "element" type "node" but no layer are now always
  shown.
• Make the audio and video visualizer a little bit more robust when they are not
  pre-loaded.

v4.12.2

Fixed

• Upgraded to graphANNIS 3.3.1 which fixes queries where optional nodes are
  located not at the end but somewhere in between the query.

v4.12.1

Added

• Show the corpus size in tokens or segmentation units. This dependends on new
  functionality
  in graphANNIS 3.3.0 and you might need to re-import or re-optimize an already
  imported corpus to see the corpus size.
• Added "selected only" checkbox to show only selected corpora in the corpus
  list. This helps to get an overview of currently selected corpora without
  needing to scroll through the list and without the annoyance that the selected
  corpora are automatically at the top of the list.

Fixed

• Upgraded to graphANNIS 3.3.0 which fixes several issues with corpus unloading
  and should improve the sitation when ANNIS is under heavy load.
• Updated issue in release script that was using the wrong graphANNIS version.

v4.11.3

Fixed

• Upgraded to graphANNIS 3.2.2 which fixes a regression with showing paged
  results when selecting multiple corpora.

v4.11.2

Fixed

• Use plural form when there is no match in result status: "0 matches".
• Upgraded to graphANNIS 3.2.1 which fixes an issue in the subgraph generation.

v4.11.1

Fixed

• Do not add "annis:doc" labels to sub-corpora when importing relANNIS corpora.
  This will fix queries where you just search for documents, e.g. by annis:doc
  but previously also got the sub-corpora as result. This bug only occurred in
  corpora with sub-corpora and for the document searches.
• Avoid null pointer exceptions when merging corpus configs for search options.
• Updated to graphANNIS 3.2.0 which contains several performance improvements
  for working with large corpora (>10 million token). You might have to
  re-import the corpora to get the full performance improvement, because a new
  graph storage implementation and statistics have been added.
• Omit the document count for results where there are no matches inside a
  document, but only the documents or sub-corpora themselves. A document count
  of 0 would otherwise be very confusing.

v4.11.0

Added

• Added the possibility to log user search interactions of the QueryController
  anonymously. You have to set the logging level for the component
  org.corpus_tools.annis.gui.QueryController to the value DEBUG, e.g. by
  adding the following lines to the
  application.properties
  file:

# Set general log level for ANNIS
logging.level.org.corpus_tools.annis=WARN
# Log user interactions
logging.level.org.corpus_tools.annis.gui.QueryController=DEBUG

Fixed

• Updated to graphANNIS bugfix version 3.1.1, which fixes issues with token
  searches when the result order is randomized or not sorted and the corpus is
  not loaded yet (korpling/graphANNIS#280).

v4.10.6

Fixed

• Updated to graphANNIS 3.1.0 which has improvements in memory handling. This
  should fix some situations where ANNIS becomes unresponsive in a enviroment
  with a lot of different corpora.

v4.10.5

Fixed

• Update to graphANNIS 3.0.0 with various bugfixes.
• Polling the import status in the adminstration view fixed.

v4.10.4

Fixed

• Update to graphANNIS 2.4.8 with various bugfixes.
• When an optional node (for negation without existence) was not at the end of
  the query, find queries could give an empty output.